View Full Version : Radeon HD 2900 (R600) reviews thread
Umm, no, because I like this one. And this belief that Crossfire or SLi are only justified at 2500x1600 is flawed. Enable everything(AA, AF, AAA etc.), in any recent, fairly intensive title and do come back to me and tell me if you find the performance pleasing across the board. And this if off-topic, BTW, and I do apologize for it.
Umm, no, because I like this one. And this belief that Crossfire or SLi are only justified at 2500x1600 is flawed. Enable everything(AA, AF, AAA etc.), in any recent, fairly intensive title and do come back to me and tell me if you find the performance pleasing across the board. And this if off-topic, BTW, and I do apologize for it.
I couldn't agree more. I'm so sick of people force-feeding this bullshit idea that highend cards are somehow completely underutilized and wasted at anything below 1600x1200 with high levels of AA/AF. I can happily report that even with a pretty decent CPU (4600+ X2 @ 2.88ghz) a LOT of games I play have trouble staying fast at 1280x1024 with or without AA/AF, and just with most quality settings upped (x1900xt 512). Perhaps it has to do with what people deem acceptable framerates...for me, below 60 is almost unbearable in most games, though there are exceptions (most RPG's are fine right down to 20fps). People seem to suggest over and over that you should only run high resolutions with a highend card. Why? Take GTA:SA for instance, with all the settings maxed and 2xAA, it's borderline playable most of the time for me. 1600x1200 or, god forbid, 2048x1536 is unplayable...but people still keep insisting that highend = highres. Why?
Also sorry for the O/T but there's not a whole lot to talk about. Waiting on some new drivers for R600 before I go buying one or not.
Something(s) new to add. Two Worlds runs like crap overall, it has these weird slowdowns where there should be now slowdowns...but it seems to take advantage of PCF(the shadows are "softer" that they were with the X1900s, same settings, no new update to the game so I`m attributing it to the addition of PCF to the 2900). I also have that signal loss bug with the drivers available on ATIs site:play something for a while and I get signal loss, must ALT-TAB in order restore things to normal. It seems it`s a bug aknowledged by ATi(at least I think I saw a KB article on it).
I`m really craving those 8.38 drivers...darned be that bug with 1440x900.:)
I'd say that a lower res monitor will guarantee your vid card is useful for a lot longer too. So I don't see it as a disadvantage either.
Silent_Buddha
27-May-2007, 22:24
Ouch, thankfully I haven't run into any signal loss issues with the 2900 XT yet.
CFAA so far is a mixed bag. In some games the increase in IQ is quite nice. In other games, it's more of an annoyance.
I really wish ATI would come out with a per game profile system where I could specify what type of AA/AF is to be used for that game, and then after quitting the game, it reverts back to my default custom general use AA/AF settings.
Especially now that CFAA is around and isn't something that I would use universally on all games.
BTW - Morgoth, if you have a free 3.5" drive bay you might consider getting a booster PSU to drive graphics only.
http://www.fsp-retail.com/en/booster_a.asp?pro_catid=6&categoryid=1&productid=7
That one isn't in retail yet, but it will drive any card coming out in the forseeable future without the need to upgrade your main CPU.
Right now I've got the Booster X3 to drive the 2900 XT. I wish I could take my tower to Japan with me. Wouldn't have had to get that.
Would be cheaper than getting a completely new PSU. The booster X3 can be had for around 70 USD in the states.
Regards,
SB
I really wish ATI would come out with a per game profile system where I could specify what type of AA/AF is to be used for that game, and then after quitting the game, it reverts back to my default custom general use AA/AF settings.
Well ATi Tray Tools is absolutely perfect for that...create your profile, specify whatever levels of AA/AF (and any other misc settings you want like Geometry Instancing) and tick the box telling it to restore current settings once application exits. Besides, if ATi tried to make something even remotely as usefull as ATT, it'd just be a 100mb bloated download built on .Net with random numerous errors that frequently require a restart or massive software updates/patches to function correctly :razz: (Sorry, CCC still leaves a bitter taste in my mouth, I could never use it again).
R600 support shouldn't be too far off now for ATT, though I don't think Overclocking will be included anytime soon.
trinibwoy
28-May-2007, 11:54
More bad news for R600
http://we.pcinlife.com/thread-772385-1-1.html
http://img296.imageshack.us/img296/2472/2900xtultra4xaa838rc7ally8.jpg (http://imageshack.us)
http://img513.imageshack.us/img513/489/8800gtx1920x1080ultra4xbq7.jpg (http://imageshack.us)
The sand looks much more detailed in the r600 shot.
BlizzardOne
28-May-2007, 13:32
And what's with the text on the 8800 pic being considerably more vivid..
And why are they comparing with the GTX when the XT is aimed at the GTS..
And are we going back to comparing FRAPS captures..
And.. etc etc:lol:
I take back what I said about CFAA...it may not suck as badly as I thought. I think it depends on the game being tested:it`s horrid in Marvel Ultimate Alliance, it`s quite great in Oblivion(in my experience)...so I guess I`ll need to look into it more. The edge-detect mode will be restricted to Crossfire setups, if I remember correctly?
And about Dirt...dunno, I ran it at 1440x900 with the 8.37.whatever are available from ATi, 4x AA, Ultra settings and it was...playable. Didn`t use fraps to see actual FPS, but it certainly wasn`t borderline slideshowic as 8 fps would suggest...so who the heck knows what`s the source of the slowdown in those benchies.
trinibwoy
28-May-2007, 13:57
Yeah SSXeon got much better results on his rig (albeit at 1680x1050)
http://www.rage3d.com/board/showthread.php?t=33891905
trinibwoy
28-May-2007, 13:59
And why are they comparing with the GTX when the XT is aimed at the GTS..
Believe me, the XT was aimed squarely at the GTX. It just couldnt get there :)
And are we going back to comparing FRAPS captures..
Yeah, my bad. :oops:
BlizzardOne
28-May-2007, 14:13
Believe me, the XT was aimed squarely at the GTX. It just couldnt get there :)
fair point :lol:
I just installed Medieval II:Total War and played on the 2900, don't remember who asked about it though.
I've patched it up to 1.2 but it seems that none of the siege weapons display their animation properly.
Fire and Smoke display (Archers) but siege weapons just "hop" when they should fire their weapons and seem "useless" don't know if it's just me or the 1.2 patch.
The game runs butter smooth at 1600x1200 4AA+2NT/16AF all detail high/highest.
Yeah SSXeon got much better results on his rig (albeit at 1680x1050)
http://www.rage3d.com/board/showthread.php?t=33891905
He's also running default settings not Ultra.
http://www.rage3d.com/board/showpost.php?p=1334960888&postcount=23
Silent_Buddha
29-May-2007, 13:54
I take back what I said about CFAA...it may not suck as badly as I thought. I think it depends on the game being tested:it`s horrid in Marvel Ultimate Alliance, it`s quite great in Oblivion(in my experience)...so I guess I`ll need to look into it more. The edge-detect mode will be restricted to Crossfire setups, if I remember correctly?
I'm not sure where I read it, I think either in the B3D piece or the TR piece, that mentioned ATI might put in Edge Detect in a future driver for single card use on the HD 2900 XT.
I wouldn't mind pairing up 2x or 4x MSAA with Edge Detect and see how it does.
Regards,
SB
Silent_Buddha
29-May-2007, 13:59
Well ATi Tray Tools is absolutely perfect for that...create your profile, specify whatever levels of AA/AF (and any other misc settings you want like Geometry Instancing) and tick the box telling it to restore current settings once application exits. Besides, if ATi tried to make something even remotely as usefull as ATT, it'd just be a 100mb bloated download built on .Net with random numerous errors that frequently require a restart or massive software updates/patches to function correctly :razz: (Sorry, CCC still leaves a bitter taste in my mouth, I could never use it again).
R600 support shouldn't be too far off now for ATT, though I don't think Overclocking will be included anytime soon.
It's been a long time since I've bothered to use ATT. Not too keen on using it again, however if it does support real per game profiles I may have to give it a try again.
And by supporting profiles do you mean that it will auto-detect what game you are running and apply the profile? Or do you have to run the game through ATT's profile for it?
In other words, if I made a profile in ATT. And then just ran the game from the standard start menu shortcut that the game installs. Will it ATT detect this and auto-load the profile?
If I have to run the game through ATT then it's basically the same functionality as the CCC profile manager. Bleh.
Regards,
SB
100% correct, it happily autodetects the game that's running and applies the created profile based on the executable name. I honestly can't fault it...when I very first tried it I encountered enough bugs to drive me bonkers but I stuck with it and it's now an insanely beautifull piece of work. I have all my games setup on profiles for various AA/AF/GI settings and so on, and I never have to touch a thing. Plus it's extremely satisfying being able to have such open, clear and easy access to every little feature you normally wouldn't see, or would have to wade through multiple menus/buttons to enable. Adaptive/Temporal AA, HQAF, it's all there and easy as hell to access...
I'd say give it a shot again, and see how you go. It can't hurt, and you just might find it's become a whole lot better since you last used it :smile:
What I'd like to know is just how much is 8800 bandwidth starved? It has a boatload of fillrate. I wonder how things would look today if 8800 had been a 512-bit design as well.
From what we know, the GTX isn`t very sensitive to memory clock increases...and it doesn`t seem like it`s mem bandwidth starved, it seems to be quite balanced in that respect.
Silent_Buddha
30-May-2007, 04:11
100% correct, it happily autodetects the game that's running and applies the created profile based on the executable name. I honestly can't fault it...when I very first tried it I encountered enough bugs to drive me bonkers but I stuck with it and it's now an insanely beautifull piece of work. I have all my games setup on profiles for various AA/AF/GI settings and so on, and I never have to touch a thing. Plus it's extremely satisfying being able to have such open, clear and easy access to every little feature you normally wouldn't see, or would have to wade through multiple menus/buttons to enable. Adaptive/Temporal AA, HQAF, it's all there and easy as hell to access...
I'd say give it a shot again, and see how you go. It can't hurt, and you just might find it's become a whole lot better since you last used it :smile:
Thanks, I'll have to check it out for the profiles once he has a non-beta version out that works on Vista 64. Not interested in the least about tweaking other settings however. But profiles would be a HUGE boon now that CFAA varies so much in benefit from game to game.
Thanks for the heads up,
SB
My God, the Lost Planet DX10 demo is a painful experience on this card. Still haven`t gotted 1440x900 through the magic resolution shuffling, but even so, the benchmark sits at about...8-10 FPS, at 1366x760, everything max, no AA(it`s broken). Disabling Motion Blur nets me a hufe 1 fps increase to get to 11. And most of everything is quite obviously bugged....this makes you wonder how the heck did the sites benching DX10 with this actually did their benching :-?
Sc4freak
30-May-2007, 11:40
And what's with the text on the 8800 pic being considerably more vivid..
The X2900 shot appears to be much more compressed.
The sand looks much more detailed in the r600 shot.I'm not quite sure what the comments are in this post (http://we.pcinlife.com/viewthread.php?tid=772385&page=1#pid14571353), but it seems like either they got the R600 16x/no AF shots reversed (hence the much more detailed sand), or for some reason they had 16x AF enabled but didn't realize it (which jives with the framerate, which jumps 2fps to 10fps w/o AF). If you look at the G80 GTX 16x AF sshot, the sand looks similarly detailed.
And what's with the text on the 8800 pic being considerably more vivid..If you look at the big pics (check the above link), they bolded the text in the G80 sshots, so it looks better in the thumbnails.
And why are they comparing with the GTX when the XT is aimed at the GTS..Again, that linked post seems (read: I can't read Chinese or translate "?????????[...]") to say that the 320MB G80 GTS gets only 1fps, but who knows. I can imagine memory getting toit (see: Goldmember :)) at 19x10 4xAA w/ what appears to be HDR.
vertex_shader
04-Jun-2007, 10:59
Factory OC'ed 850/2000mhz BBA HD2900XT 512mb gddr3
http://img521.imageshack.us/img521/8579/bba29001mk2.jpg
http://img376.imageshack.us/img376/7121/bba29002le1.jpg
A source is always welcome;
http://cgi.ebay.com/ATI-Exclusive-Limited-Edition-OCed-HD-2900-XT_W0QQitemZ130117523635QQihZ003QQcategoryZ40161QQ ssPageNameZWDVWQQrdZ1QQcmdZViewItem
SugarCoat
04-Jun-2007, 16:46
nothing like buying a "factory OC x2900XT" from someone who's sold BFG OC2 GTX BIOS on ebay. :roll:
Gee Wonder what hes done to the 2900XTs.
I kinda like the plain cooler. :)
I really wish ATI would come out with a per game profile system where I could specify what type of AA/AF is to be used for that game, and then after quitting the game, it reverts back to my default custom general use AA/AF settings.
Well as the CCC already does most of this, it would seem like your beef is restoring your "general use" settings after you quit a game. Several months ago, I simply added a CCC profile for my default desktop I call "1280x1024" (you could call it anything, really) which I select from the ATi icon in the SysTray after quitting a game, and of course this profile resets everything to my default 3d settings. It's an extra mouse-click over what you're asking for, but there it is...;)
I also think that ATi might've thought that you'd want to use profiles for every game you run--in which case there'd be little need to switch back to some kind of default just for desktop graphics. In any event, setting up a desktop profile would solve the problem.
Cartoon Corpse
04-Jun-2007, 21:14
wow the x1900xtx (mine) only consumes around 120W (i googled for it)
the top line 2900 consumes 240W! i mean i have the headroom on my 680W PSU, but i was thinking we were getting more power and reduced consumption.
what about laptop dx10? (edit...following HD 2600 and 2400 in late june...i read)
which of the reviews is a nice overview, without teaching me how to build one? (i was looking at tomshardware)
Silent_Buddha
04-Jun-2007, 22:21
wow the x1900xtx (mine) only consumes around 120W (i googled for it)
the top line 2900 consumes 240W! i mean i have the headroom on my 680W PSU, but i was thinking we were getting more power and reduced consumption.
what about laptop dx10? (edit...following HD 2600 and 2400 in late june...i read)
which of the reviews is a nice overview, without teaching me how to build one? (i was looking at tomshardware)
Somehow you managed to google the correct power use of the x1900xtx and weren't able to get the correct power use of the HD 2900 XT?
http://www.xbitlabs.com/articles/video/display/r600-architecture_14.html#sect0
Try roughly 160ish. I'm amazed that people continue to parrot out 240W as the power figure when the power connectors can't provide more than 225W max (all that is required to run the card)
Regards,
SB
Silent_Buddha
04-Jun-2007, 22:28
Well as the CCC already does most of this, it would seem like your beef is restoring your "general use" settings after you quit a game. Several months ago, I simply added a CCC profile for my default desktop I call "1280x1024" (you could call it anything, really) which I select from the ATi icon in the SysTray after quitting a game, and of course this profile resets everything to my default 3d settings. It's an extra mouse-click over what you're asking for, but there it is...;)
I also think that ATi might've thought that you'd want to use profiles for every game you run--in which case there'd be little need to switch back to some kind of default just for desktop graphics. In any event, setting up a desktop profile would solve the problem.
Well, considering that for general use in game 4xMSAA or 8xMSAA + 16x AF is fine that's why I'd like it go back to a default value as that is what I'd use in the majority of games.
Certain games I've found to run well and benefit from CFAA. Some games when run with CFAA don't benefit much with high levels of AF so 8x is enough for certain titles. And again with CFAA sometimes transparency AA (adaptive AA) isn't required or only the performance version is needed.
Yes I could launch the games through the profile manager or or though a profile off the CCC tray, however it's much more convenient to just run the game through it's default icon. Especially since I have a game bar on my secondary monitor that I use for the 10 games I currently play, and 10 other games that I play during LAN gaming.
Regards,
SB
Does CFAA work in OGL?My eyes tell me it does, but the performance tells me it doesn`t. If it does, add Chronicles of Riddick to the list of games that do good with CFAA.
vertex_shader
05-Jun-2007, 15:14
http://www.ocworkbench.com/2007/computex/day1/Picture%20076.JPG
http://www.ocworkbench.com/2007/computex/day1/Picture%20077.JPG
Link (http://www.ocworkbench.com)
Why Gecube call the card as "OC edition" when its runing on reference clock? :wink:
Because you`ll be able to OC it on your own and void your warranty, seeing that impressive cooling solution.
Geeforcer
06-Jun-2007, 07:38
The card formerly known as 2900 XTX (R600 + 1GB GDDR4) makes a debut (http://www.extreme-pc.ca/showproduct.asp?menu1id=12&menu2id=99&menu3id=39&productid=371414). (Not a review, rather a link to the site selling the mythical beast).
AlStrong
06-Jun-2007, 15:26
Why Gecube call the card as "OC edition" when its runing on reference clock? :wink:
Because it has Turbo fans. ;)
The card formerly known as 2900 XTX (R600 + 1GB GDDR4) makes a debut (http://www.extreme-pc.ca/showproduct.asp?menu1id=12&menu2id=99&menu3id=39&productid=371414). (Not a review, rather a link to the site selling the mythical beast).
No bias at all! IMO, they'd be better without the single comparison to the 8800GTX SLI. :lol:
edit: I guess it's still pre-order with a date estimate. Wonder if that's just a guess..
INKster
06-Jun-2007, 15:44
The card formerly known as 2900 XTX (R600 + 1GB GDDR4) makes a debut (http://www.extreme-pc.ca/showproduct.asp?menu1id=12&menu2id=99&menu3id=39&productid=371414). (Not a review, rather a link to the site selling the mythical beast).
Interesting as that is the only place where it is found, and with no mention of the actual brand (i've never heard of that one on the cooler sticker at least).
Where are the big-name brands with this one ?
Since the 1GB variant seems to be something that only SIs cared about, these dudes, as SIs have 1GB(or should have) R600s and cash in on the noise surrounding this launch. As well, if you read their stuff carefull, they say that they have some nice price cuts if you get a complete system from them bla bla. I find it to be too expensive, even considering the memory footprint and the slight OC, bearing in mind where the R600 stands in terms of performance, at least in present times.
Interesting as that is the only place where it is found, and with no mention of the actual brand (i've never heard of that one on the cooler sticker at least).
Where are the big-name brands with this one ?
Will the 2900 XTX end up like the famous Voodoo 5 6000?
Colourless, time to come up with another special mip mapping scheme for the XTX! (Or whatever it was... you know).
Will the 2900 XTX end up like the famous Voodoo 5 6000?
Sure, if Nvidia buys AMD.
:lol:
Because you`ll be able to OC it on your own and void your warranty, seeing that impressive cooling solution.
Gecube's coolers are not that great for staying in contact with the core.
I had the AGP version of their 1950 card and it was perfectly fine while it worked. Extremely cool AND quiet and the heatsink setup allowed me to adjust the height of the cooler (for clearance above my s-ata headers).
But 2 and a half weeks in, boom, issues. Over-heating.
The cooler is simply too heavy and the retention mechanism too (for lack of a better word) crap. Unless they have changed the design around somewhat to compensate for the weight and the way that it started to sag, I think a few 2900's might end up burning up.
That being said, while contact was maintained, a fabulous cooler.
Cartoon Corpse
08-Jun-2007, 15:51
Somehow you managed to google the correct power use of the x1900xtx and weren't able to get the correct power use of the HD 2900 XT?
http://www.xbitlabs.com/articles/video/display/r600-architecture_14.html#sect0
Try roughly 160ish. I'm amazed that people continue to parrot out 240W as the power figure when the power connectors can't provide more than 225W max (all that is required to run the card)
Regards,
SB
that's the figure that was in the Tomshardware articalle i was reading before i posted. so i googled for the x1900xtx since it wasn't mentioned.
nice pictures :-)
http://img373.imageshack.us/img373/3697/r600dievc0.jpg
http://img137.imageshack.us/img137/623/r600dieshotsb9.jpg
If anyone in the US wants a 2900XT, Best Buy has a VisionTek 2900XT for $319.99 instantly. You may need to purchase on the website though and not in store. If I hadn't just purchased a 8800 GTS, I'be probably jump on it. ;)
Just thought someone might want to know.
Good catch. That's a pretty good price, considering its age. I suppose it's b/c they're offering all Vtek cards at 20% off, though I didn't think R600XT would be $400 retail in brick and mortar stores (those prices really are insane compared to the internet, and even this sale can't stop BB from offering a X1300XT AGP for $130 vs. an X1950GT AGP at NewEgg for the same price after a $20 rebate). Then again, I've been paying more attn to R600's scores and IQ than its price.
SugarCoat
11-Jun-2007, 06:32
Good catch. That's a pretty good price, considering its age. I suppose it's b/c they're offering all Vtek cards at 20% off, though I didn't think R600XT would be $400 retail in brick and mortar stores (those prices really are insane compared to the internet, and even this sale can't stop BB from offering a X1300XT AGP for $130 vs. an X1950GT AGP at NewEgg for the same price after a $20 rebate). Then again, I've been paying more attn to R600's scores and IQ than its price.
well its like that in all huge retail outlets, they dont exactly pay attention or care about how bargain hunters view them, they just cater to Joe Idiot that wanders in off the street. Everything they sell is that way.
I've been a fan of visiontek since they were still doing nVidia cards, also the only ATi AIB partner, as far as im aware, that offers the Lifetime parts/labor warranty on all their products. For awhile i thought they were going under, just after the R420 release, hopefully they can hang in there.
Extended CrossFire review on Computer Base.de (in German so you might wanna use a translator if you don't speak German): http://www.computerbase.de/artikel/hardware/grafikkarten/2007/test_ati_radeon_hd_2900_xt_crossfire/
Unfortunately tested on 8.37.4.3 (= no EATM). Anyway, is EATM for R600 officially supported in Cat. 7.5?
Silent_Buddha
11-Jun-2007, 23:44
If anyone in the US wants a 2900XT, Best Buy has a VisionTek 2900XT for $319.99 instantly. You may need to purchase on the website though and not in store. If I hadn't just purchased a 8800 GTS, I'be probably jump on it. ;)
Just thought someone might want to know.
Damn, at that price it's almost tempting to get one for crossfire. However, since I've sworn off dual GPU setups until they can get multimonitor working with crossfire/SLI, that's a no-go...
Regards,
SB
One user told me, that his new HD2900XT (built by Sapphire) has not the same cooler, as other cards. Most HD2900XTs are equipped with dual heat-pipe cooler (http://ati.amd.com/designpartners/media/images/43006A_R600_9_med_Frt.jpg), but his new card carries slightly different heat-sink with 3 heat-pipes:
http://xs216.xs.to/xs216/07242/DSC04810.jpg.xs.jpg (http://xs.to/xs.php?h=xs216&d=07242&f=DSC04810.jpg)
(mirror link (http://www1.filehost.to/files/2007-06-12_03/190826_DSC04810.jpg))
(most other XT's from Sapphire use dual heat-pipe design)
Did ATi change the design of XT's reference cooler? That would be quite good (lower temperatures, noise...)
I checked ATi's pictures and almost all of them features dual heat-pipe version, only one (two XTs in CrossFire (http://ati.amd.com/designpartners/media/images/43007A_HD_2900_Crossfire_9.5_R.jpg)) shows the better cooler...
vertex_shader
12-Jun-2007, 21:56
One user told me, that his new HD2900XT (built by Sapphire) has not the same cooler, as other cards. Most HD2900XTs are equipped with dual heat-pipe cooler (http://ati.amd.com/designpartners/media/images/43006A_R600_9_med_Frt.jpg), but his new card carries slightly different heat-sink with 3 heat-pipes:
http://xs216.xs.to/xs216/07242/DSC04810.jpg.xs.jpg (http://xs.to/xs.php?h=xs216&d=07242&f=DSC04810.jpg)
(mirror link (http://www1.filehost.to/files/2007-06-12_03/190826_DSC04810.jpg))
(most other XT's from Sapphire use dual heat-pipe design)
Did ATi change the design of XT's reference cooler? That would be quite good (lower temperatures, noise...)
I checked ATi's pictures and almost all of them features dual heat-pipe version, only one (two XTs in CrossFire (http://ati.amd.com/designpartners/media/images/43007A_HD_2900_Crossfire_9.5_R.jpg)) shows the better cooler...
Yes its different, nice find.
Here is the original naked:
http://www.overclockercafe.com/Reviews/VGA/HIS_HD2900XT/pics/1014.jpg
Deja-vu
http://img148.imageshack.us/img148/4760/hd2900xtbi4.jpg
Looks like sapphire (or amd) made the 3 heatpipe version fit to the reference cooler case :smile:
pakotlar
20-Jun-2007, 23:02
Check this out! New Xbit review with Cat 7.5's. This has AAA enabled on both platforms as a fyi. The x2900xt kicks fuckin ASS, beating the 8800ultra in several cases WITH AA and AF enabled. Its not until several pages in that the x2900xt starts to show its usual crappy + strange performance. The first couple pages make u actually think that the extra memory bandwith and shader power do something!
SugarCoat
21-Jun-2007, 00:11
Check this out! New Xbit review with Cat 7.5's. This has AAA enabled on both platforms as a fyi. The x2900xt kicks fuckin ASS, beating the 8800ultra in several cases WITH AA and AF enabled. Its not until several pages in that the x2900xt starts to show its usual crappy + strange performance. The first couple pages make u actually think that the extra memory bandwith and shader power do something!
Only see an Improvement in Farcry myself and the Battlefield/Call of Juarez at the lower resolutions. Several Cases sounds like 7 games, its infact 4 "cases" over three games out of eighteen where it was faster then the Ultra. Also got kicked to the curb when resolution went up with the exception of farcry again. Not really anything great from that driver in my opinion . Still broken performance in STALKER, FEAR, SupCom etc etc or performing way too close to X1950 performance.
pakotlar
21-Jun-2007, 03:26
Only see an Improvement in Farcry myself and the Battlefield/Call of Juarez at the lower resolutions. Several Cases sounds like 7 games, its infact 4 "cases" over three games out of eighteen where it was faster then the Ultra. Also got kicked to the curb when resolution went up with the exception of farcry again. Not really anything great from that driver in my opinion . Still broken performance in STALKER, FEAR, SupCom etc etc or performing way too close to X1950 performance.
In those first couple pages it comes close or beats the gtx in no less than 7 benchmarks in that review and 5 games in that review. Looks closely. Funny you chose 7 as the number you thought it sounded like. And several means at least 2 fyi :wink: Ill list the games for you and resolutions.
Battlefield 2142: All res, beats gtx at 1600 and 1920
COJ: All res, beats gtx at all res, ultra at 1280.
Far Cry, Pier: Beats gtx at 1600 and 1920, ultra at 1920
Far Cry Research: Beats gtx at all res, ultra at 1920
Far Cry Research HDR: Debateable because at 1920, gtx is 13% faster, but who would really notice the diff. Also, Min fps is higher at ALL resolutions in this benchmark than Ultra. Which would make the gameplay experience better on the 2900xt.
Fear Extraction point: Comes within 3 frames of the GTX at 1920, though Min fps is less than half of the gtx. However, considering that the 1950xtx has higher min framerate, Id bet that will change.
Splinter Cell DA HDR: Comes within 2.5 fps of the GTX at 1600, betas GTX at 120.
X3 Reunion: Comes within 2.4 fps of GTX at 1280, within 4.6 fps of GTX at 1600. Gets killed at 1920 for some reason. though still much faster than GTS.
Command in Coquer 3: Is equal with GTX and Ultra. Min frame beats GTX at 1920. Of course all cards still average 30fps so this is a "who cares"
Company of Heroes: Comes pretty dang close at all resolutions, including within 5.1 fps at 1920 (80.3 vs 75.2).
So it is actually 8 games and 10 benchmarks in that review that the 2900xt does very well against the GTX. It does get killed in a few games, including NWN2, Supreme Commander. However, in the majority of the games in that review it does very well, and although in some of them, like HL2 it only pulls up to the GTS that is not a clear LOSS IMO, considering the price point. There is clearly some room for improvement in the drivers, as there are cases where the 1950XTX beats it, or is even with it.
I'm not a fan of this card because I, like many others, was expecting much better performance than it delivered, but it is encouragin to see driver improvements. If it can keep this up, at 350-400 it would be an extremely tempting buy over the GTX.
As you said, in Stalker, SC performance is definetely broken, and fear could certainly improve, but if you look at it, thats 3 games (2 really since fear performance scales really well on 2900xt to 1920 vs gtx) out of 11. NWN2 is also extremely shitty, but who really cares, that game sucks :D.
In those first couple pages it comes close or beats the gtx in no less than 7 benchmarks in that review and 5 games in that review. Looks closely. Funny you chose 7 as the number you thought it sounded like. And several means at least 2 fyi :wink: Ill list the games for you and resolutions.
Battlefield 2142: All res, beats gtx at 1600 and 1920
COJ: All res, beats gtx at all res, ultra at 1280.
Far Cry, Pier: Beats gtx at 1600 and 1920, ultra at 1920
Far Cry Research: Beats gtx at all res, ultra at 1920
Far Cry Research HDR: Debateable because at 1920, gtx is 13% faster, but who would really notice the diff. Also, Min fps is higher at ALL resolutions in this benchmark than Ultra. Which would make the gameplay experience better on the 2900xt.
Fear Extraction point: Comes within 3 frames of the GTX at 1920, though Min fps is less than half of the gtx. However, considering that the 1950xtx has higher min framerate, Id bet that will change.
Splinter Cell DA HDR: Comes within 2.5 fps of the GTX at 1600, betas GTX at 120.
X3 Reunion: Comes within 2.4 fps of GTX at 1280, within 4.6 fps of GTX at 1600. Gets killed at 1920 for some reason. though still much faster than GTS.
Command in Coquer 3: Is equal with GTX and Ultra. Min frame beats GTX at 1920. Of course all cards still average 30fps so this is a "who cares"
Company of Heroes: Comes pretty dang close at all resolutions, including within 5.1 fps at 1920 (80.3 vs 75.2).
So it is actually 8 games and 10 benchmarks in that review that the 2900xt does very well against the GTX. It does get killed in a few games, including NWN2, Supreme Commander. However, in the majority of the games in that review it does very well, and although in some of them, like HL2 it only pulls up to the GTS that is not a clear LOSS IMO, considering the price point. There is clearly some room for improvement in the drivers, as there are cases where the 1950XTX beats it, or is even with it.
I'm not a fan of this card because I, like many others, was expecting much better performance than it delivered, but it is encouragin to see driver improvements. If it can keep this up, at 350-400 it would be an extremely tempting buy over the GTX.
As you said, in Stalker, SC performance is definetely broken, and fear could certainly improve, but if you look at it, thats 3 games (2 really since fear performance scales really well on 2900xt to 1920 vs gtx) out of 11. NWN2 is also extremely shitty, but who really cares, that game sucks :D.
A review from one of 2 sites that manage to get results 180 degrees from each other. [H] being the other one. As far as I'm concerned, both sites should be removed from the good review list this round as IT is clear they both have done something to configure the results the way they wanted them to be. [H] shows total dominance of the 8800s while XBIT shows the opisite. And with the same exact hardware and drivers. I'm sorry, but both are worthless this time. DH, FS, or even B3D would be better than those 2.
pakotlar
21-Jun-2007, 05:01
A review from one of 2 sites that manage to get results 180 degrees from each other. [H] being the other one. As far as I'm concerned, both sites should be removed from the good review list this round as IT is clear they both have done something to configure the results the way they wanted them to be. [H] shows total dominance of the 8800s while XBIT shows the opisite. And with the same exact hardware and drivers. I'm sorry, but both are worthless this time. DH, FS, or even B3D would be better than those 2.
I personally can't stand H reviews. Their testing methodology reeks. I've always thought Xbit had some of the best reviews around. Care to back that up? I think Techreport ranks up there as well, though they don't have a review with 7.5/8.38 drivers.
I personally can't stand H reviews. Their testing methodology reeks. I've always thought Xbit had some of the best reviews around. Care to back that up? I think Techreport ranks up there as well, though they don't have a review with 7.5/8.38 drivers.
Look at all of XBIT initial scores with the 2900, then look at [H]. You'll see they are 180 of each other even thou they both used the same drivers, cpus, mobos, ram amounts(not sure on type) and video cards. And both sites go against what every other review has said. The XT is a good card, loses some, equal in others and when it wins, it competes with the GTX. So if 1 review shows all score skewed one way and another shows them skewed the other with same hardware, then it has to be faulty benching or they puposely skewed it to make the card look good/bad depending on how they thought the out come should be. ANd yes, I agree XBIT usually is one of the better sites. But for some reason, I find their result very suspect, [H] as well mind you, when every other tom dick and harry site that reviewed the cards got results that placed it anywhere from xt<gts<gtx to xt/gts<gtx to gts<xt<gtx.
SugarCoat
21-Jun-2007, 15:23
In those first couple pages it comes close or beats the gtx in no less than 7 benchmarks in that review and 5 games in that review. D.
aha but you said
beating the 8800ultra in several cases ;)
pakotlar
21-Jun-2007, 16:28
aha but you said
;)
and it did in quite a few ;)
edit: eh ok ok, so my first post was a bit exaggerated. I'm just pretty excited at how well its doing. I thought that AA just totally murdered this card, but I guess it can hold its own in many situations. Hopefully later drivers will boost perf. like they did on the X1800, and work out the kinks in games like Supreme Commander. Peace
MistaPi
09-Jul-2007, 16:40
Sorry for being lazy, but with the reviews sumed up and looking exclusively at AA/AF numbers, is HD 2900XT faster than 8800GTS 640MB?
Sorry for being lazy, but with the reviews sumed up and looking exclusively at AA/AF numbers, is HD 2900XT faster than 8800GTS 640MB?
According to Brent and the gang over at [H], I certainly don't think they'd say so....(using Cat 7.6)
Radeon HD 2900 XT CrossFire is most certainly not a good gaming solution. It just does not deliver the kind of performance advantages ATI needs to compete with what NVIDIA already has in the marketplace. Sapphire has created a very complete retail package that would suit any ATI fan well and Sapphire should not be looked down on for the substandard gaming experience that ATI has delivered this generation. I would say stay away from the Radeon HD 2900 XT in CrossFire at this time; it just does not represent a good gaming value. (http://enthusiast.hardocp.com/article.html?art=MTM1OSw5LCxoZW50aHVzaWFzdA==)
*Note: Be sure to read more than just this to be sure I'm not taking anything out of context....there's plenty of info in that review...
John Reynolds
09-Jul-2007, 17:30
Picked up the new GFW issue and Jason Cross' testing showed the 2900 XT to be 20-30% faster than a GTS. That [H] review and its scores are really something of an anomaly.
Tim Murray
09-Jul-2007, 17:32
Picked up the new GFW issue and Jason Cross' testing showed the 2900 XT to be 20-30% faster than a GTS. That [H] review and its scores are really something of an anomaly.
Hmmm... [H]'s testing is anomalous and shows NVIDIA with unexpectedly large leads. That certainly doesn't happen in every generation of cards...
Hmmm... [H]'s testing is anomalous and shows NVIDIA with unexpectedly large leads. That certainly doesn't happen in every generation of cards...
Having been an editor over at [H] in the past, I can say with confidence that Brent is a class-act editor and I highly doubt he'd skew anything for vendor A or vendor B....Regardless, we can all have our own opinions....moving on...
I'm not 110% sure, but isn't [H] the only review using the latest Cat 7.6 drivers? If so, what's the use in debating the results from an older driver in another review?
I guess we'll all have to wait for 7.7's to made a decision... :sad:
John Reynolds
09-Jul-2007, 18:51
Tim can speak for himself, but I don't think either of us were trying to allude to intentional skewing of test results. Just pointing out that those scores aren't exactly congruent with most other results now available for comparison. And the 7.6s didn't see that much of a sea change, at all, to say those [H] scores stand completely aloof.
Tim Murray
09-Jul-2007, 19:22
Having been an editor over at [H] in the past, I can say with confidence that Brent is a class-act editor and I highly doubt he'd skew anything for vendor A or vendor B....Regardless, we can all have our own opinions....moving on...
I'm not 110% sure, but isn't [H] the only review using the latest Cat 7.6 drivers? If so, what's the use in debating the results from an older driver in another review?
I guess we'll all have to wait for 7.7's to made a decision... :sad:
For whatever reason, there does seem to be a pattern of [H]'s results usually showing larger discrepancies in favor of NVIDIA cards. For example, I'm seeing the 2900 XT do very well in Crossfire in TR's initial R600 review, including some games where [H] shows it getting demolished. If someone decides to ask about this, [H] will reply with "oh we do real-world testing and this is what you would see" or something like that. It's maddening, because on the off chance they have found something nobody knows how to reproduce it or explain it.
I've also wondered about the margin of error in [H]'s approach; I can't help but think that it's astronomical compared to other benchmarking schemes.
Tim can speak for himself, but I don't think either of us were trying to allude to intentional skewing of test results. Just pointing out that those scores aren't exactly congruent with most other results now available for comparison. And the 7.6s didn't see that much of a sea change, at all, to say those [H] scores stand completely aloof.
Understood... :cool:
Understood... :cool:
Well I for one smell something fishy. [H] is the only site that show Nvidia winning by good margins overall while xbit shows the exact optisite using exact same hardware and drivers. So either one or both sites are wrong.
Tim Murray
10-Jul-2007, 02:57
Well I for one smell something fishy. [H] is the only site that show Nvidia winning by good margins overall while xbit shows the exact optisite using exact same hardware and drivers. So either one or both sites are wrong.
Or they have totally separate benchmarking methodologies, and the areas that they test are stressing different components of the cards, etc. I'm a long, long way from attributing this to malice.
Blacklash
10-Jul-2007, 06:25
I'd offer this up to demonstrate how "unfair" and biased [H]ard have been to ATi in the past:
From [H]ard:
"If you want the best bang for your buck, the PowerColor Radeon X1950 PRO Extreme 256 MB is it."
http://www.hardocp.com/article.html?art=MTI2NSw5LCwxNTA=
"ATI has proven they are a leader and not a follower with the X1950 XTX. ATI has released the world’s first consumer 3D graphics card with GDDR4 memory clocked at the highest ever stock speed that chews through games when it comes to high definition gaming. Memory bandwidth looks to one again be the defining factor in 3D performance. With a re-designed heatsink/fan unit, faster memory, and lowered price, the ATI Radeon X1950 XTX and CrossFire Edition are both serious 3D gaming video cards for the [H]ardcore that offer some value over NVIDIA’s more expensive 7950 GX2. ATI’s CrossFire dual GPU gaming platform looks to have just grown up."
http://www.hardocp.com/article.html?art=MTE0NCwxMiwsNzA=
"If you want the fastest performing video cards right now that provide the highest level of image quality in games, it doesn’t get any better than the Radeon X1900 XT and XTX. You get High Quality AF, Adaptive AA, HDR with AA, and very fast shader performance. If you have been waiting to play F.E.A.R. at very high quality settings, your wait is over; the Radeon X1900 XT/XTX is the card for you."
http://www.hardocp.com/article.html?art=OTUzLDEzLCxoZW50aHVzaWFzdA==
The problem is with the product, not the reviewer.
pakotlar
10-Jul-2007, 07:32
I've also wondered about the margin of error in [H]'s approach; I can't help but think that it's astronomical compared to other benchmarking schemes.
This is certainly the largest problem with their testing methods. Even if the margin of error wasn't huge, is there any way to quantify it?
This is certainly the largest problem with their testing methods. Even if the margin of error wasn't huge, is there any way to quantify it?
Unless we went through the tests ourselves, or have data about the tests, no real way. In game results as everyone knows can vary widely. And looking at it from the outside, way too many areas where things can shift.
The whole 'playable settings' result just doesn't sit well with me. It's no biggie though, I never even read [T] reviews, there are far better sources for controlled, explainable results.
I'd offer this up to demonstrate how "unfair" and biased [H]ard have been to ATi in the past:
From [H]ard:
"If you want the best bang for your buck, the PowerColor Radeon X1950 PRO Extreme 256 MB is it."
http://www.hardocp.com/article.html?art=MTI2NSw5LCwxNTA=
"ATI has proven they are a leader and not a follower with the X1950 XTX. ATI has released the world’s first consumer 3D graphics card with GDDR4 memory clocked at the highest ever stock speed that chews through games when it comes to high definition gaming. Memory bandwidth looks to one again be the defining factor in 3D performance. With a re-designed heatsink/fan unit, faster memory, and lowered price, the ATI Radeon X1950 XTX and CrossFire Edition are both serious 3D gaming video cards for the [H]ardcore that offer some value over NVIDIA’s more expensive 7950 GX2. ATI’s CrossFire dual GPU gaming platform looks to have just grown up."
http://www.hardocp.com/article.html?art=MTE0NCwxMiwsNzA=
"If you want the fastest performing video cards right now that provide the highest level of image quality in games, it doesn’t get any better than the Radeon X1900 XT and XTX. You get High Quality AF, Adaptive AA, HDR with AA, and very fast shader performance. If you have been waiting to play F.E.A.R. at very high quality settings, your wait is over; the Radeon X1900 XT/XTX is the card for you."
http://www.hardocp.com/article.html?art=OTUzLDEzLCxoZW50aHVzaWFzdA==
The problem is with the product, not the reviewer.
Yay, 3 separate examples of pretty much the same product. Yes you've provided thorough proof :roll:
My only complaint with their reviews is we have to rely on what Brent or another editor deams "playable". That may change by a decent amount to your terms or my terms of what is playable. But that I guess is the same thing about not having a way to reproduce their testing...ect. Still I think their reviews add value so long as its taken into account with all the other information on a product....
Its one review thats done differently from all the rest, its better if reviews are done differently otherwise you end up with the same old same old. Now the margin of error, if someone was to do that for any of these benchmarks (time demo tests too), a review wouldn't be done in one week it would take one month. By then the review is useless, new drivers what not.
Actually the only other site that does actual game play tests that I can think of is B....somthing can't think of the name wierd.
Anyways I wouldn't be suprised if there were areas of games where the 2900 falls, looking at the synthetic results we see areas where the r600 is weak, is it a far stretch to think parts of some levels would push the r600 harder because of the relative weakness?
http://www.firingsquad.com/hardware/diamond_radeon_2900_xt_1gb/page18.asp
some of these also mimick [H]
Actually the only other site that does actual game play tests that I can think of is B....somthing can't think of the name wierd.
Bit-tech.net?
yeah thats the one.:smile:
Its one review thats done differently from all the rest, its better if reviews are done differently otherwise you end up with the same old same old.
The 'same old' tells us what we want to know... The same old is essentially a baseline that provide a great way to visualise improvements with new parts. The same old just works. One reviewers' assumption of what's playable in a game really doesn't mean squat when you've got no way of quantifying what you're seeing. "Oh cool, 55fps in 'Dense Tree Area Outside Town' ", how do I compare this kind of result to that of an opposing part, without possibly reading a review on said part on the same site as this initial review? I can't...I can't in anyway justify taking said results seriously when there is absolutely no way of determining their validity. Now i'm all for abstract / unusual / (realistic?) testing methods, but in all honesty, I / We can't make use of them. There's no control in such cases. :???:
One mans 'playable' is another mans 'who's playing this damn slidesh- oh hell no".
you can't have best of both worlds though, you can't get time demos do everything you do in game, time demos use triggers to set many things off, where in game that won't happen. So you end up loosing things for those controls. Thats why its best if you have both types of tests. Its always better to have more data then less data, or more of the same data is what I'm trying to say. I agree its impossible to assess [H]'s methodology without being there and seeing the tests run but we have to look at credibility and believablity for that.
poopypoo
11-Jul-2007, 03:38
I agree with jb. H uses an innovative method and that's interesting. They pioneered min & max framerates reporting and that's great. However, when they draw conclusions, they seem to draw some unusual conclusions -- ones that often lead me to wonder what their motivations are. They've several times taken products with performance within 5% of each other and deemed one "playable" and one not, they've often tested cards at significantly lower IQ settings from one another despite several other sites reporting respectable numbers with higher settings, and the actual conclusions pages themselves have a several times been in extreme exaggeration of their testing results. For the most part I've stopped reading their reviews, but, other than these flaws, sure, it's great that they try something different. Occasionally something there will get linked here that I'll read. My main concern is actually with their big brother attitude over there -- how many other hardware review sites proclaim their site and methodologies as perfect, and number one, and take time out of every review to point out how their unique (and unreliable) testing methodologies are so superior?
pakotlar
12-Jul-2007, 08:17
A review from one of 2 sites that manage to get results 180 degrees from each other. [H] being the other one. As far as I'm concerned, both sites should be removed from the good review list this round as IT is clear they both have done something to configure the results the way they wanted them to be. [H] shows total dominance of the 8800s while XBIT shows the opisite. And with the same exact hardware and drivers. I'm sorry, but both are worthless this time. DH, FS, or even B3D would be better than those 2.
Absolutely not. Have you ever considered that the benchmark scenarios that Xbit chooses could be different from those of the websites that show severely different results? Ever consider that LOADS differ in different parts of the game, and some scenario are more efficient to render on one card versus another.
An example:
Right now we pretty much universally agree that Oblivion runs way better on R580 than G70/71. That's a fair statement. However there were a few publications at the time that were completely legitimate, but showed results where the nvidia part either met or beat r580. Yes, even in foliage. Those publications didn't lie, but got different results. Got it?
Arnold Beckenbauer
12-Jul-2007, 11:42
One user told me, that his new HD2900XT (built by Sapphire) has not the same cooler, as other cards. Most HD2900XTs are equipped with dual heat-pipe cooler (http://ati.amd.com/designpartners/media/images/43006A_R600_9_med_Frt.jpg), but his new card carries slightly different heat-sink with 3 heat-pipes:
http://xs216.xs.to/xs216/07242/DSC04810.jpg.xs.jpg (http://xs.to/xs.php?h=xs216&d=07242&f=DSC04810.jpg)
(mirror link (http://www1.filehost.to/files/2007-06-12_03/190826_DSC04810.jpg))
(most other XT's from Sapphire use dual heat-pipe design)
Did ATi change the design of XT's reference cooler? That would be quite good (lower temperatures, noise...)
I checked ATi's pictures and almost all of them features dual heat-pipe version, only one (two XTs in CrossFire (http://ati.amd.com/designpartners/media/images/43007A_HD_2900_Crossfire_9.5_R.jpg)) shows the better cooler...
http://www.club3d.nl/index.php/products/graphics/item/270
Click on the high res image: Club3D HD2900XT seems to have 3 heat pipes.
Absolutely not. Have you ever considered that the benchmark scenarios that Xbit chooses could be different from those of the websites that show severely different results? Ever consider that LOADS differ in different parts of the game, and some scenario are more efficient to render on one card versus another.
An example:
Right now we pretty much universally agree that Oblivion runs way better on R580 than G70/71. That's a fair statement. However there were a few publications at the time that were completely legitimate, but showed results where the nvidia part either met or beat r580. Yes, even in foliage. Those publications didn't lie, but got different results. Got it?
You are talking about 1 game, I'm referring to a slew of games used by both site, with same hardware and drivers that show complete opposit results. If it was just 1 or 2 games, I could agree with you. But every single game tested by both sites? Come on man, even you should be able to see that.
demalion
12-Jul-2007, 16:55
I'm wondering how much of these past concerns (http://forum.beyond3d.com/showpost.php?p=196018&postcount=113) are related to the discussed discrepancies and current review methodologies? The original context was cheating, but the gist of it applies to accurate game performance too.
Red Herring
12-Jul-2007, 19:14
http://www.club3d.nl/index.php/products/graphics/item/270
Click on the high res image: Club3D HD2900XT seems to have 3 heat pipes.
I picked up a Powercolour 2900XT from OcUK and it also has the 3 heat pipes in the cooler.
pakotlar
12-Jul-2007, 19:59
You are talking about 1 game, I'm referring to a slew of games used by both site, with same hardware and drivers that show complete opposit results. If it was just 1 or 2 games, I could agree with you. But every single game tested by both sites? Come on man, even you should be able to see that.
It's not just one game. Different websites sometimes use different bencmark routines that run through different portions of the same game. It's not as easy as "this website sucks, this ones good" approach, but it does make a whole lot more sense. And it is definetely not every game that is skewed on Xbit's results, look again. Also, check the description of testing procedures, I believe those differ as well (AAA on i believe).
http://www.club3d.nl/index.php/products/graphics/item/270
Click on the high res image: Club3D HD2900XT seems to have 3 heat pipes.
The early ones had two. I have a Sapphire that I got around launch, and it has two heatpipes. The Palit I got two days ago has three.
The early ones had two. I have a Sapphire that I got around launch, and it has two heatpipes. The Palit I got two days ago has three.
The 1Gb GDDR4 Sapphire I just bought has 2.
The 1Gb GDDR4 Sapphire I just bought has 2.
Then only God knows what`s the story with the heatpipes:D
I'm really surprised by the lack of R600 reviews lately across the net......Is there any update on when Rys is going to have his R600 follow-up ready so we can have some new reading material? :?:
edit: Although not an explicit R600 review, I did come across this article benchmarking (http://www.tweaktown.com/articles/1145/ati_catalyst_7_7_performance_analysis/index.html)the latest Cat 7.7's
I think a lot of sites are probably waiting a few driver revisions in the hope that ATI will pull some driver magic. From the looks of the last couple of versions, doesn't seem to be forthcoming much.
RobertR1
23-Jul-2007, 21:31
I think a lot of sites are probably waiting a few driver revisions in the hope that ATI will pull some driver magic. From the looks of the last couple of versions, doesn't seem to be forthcoming much.
By the time Ati figures out the drivers (if they actually do!) there will be revision out and nvidia will have upped the game with the next series.
Really, Ati will need a whole new card to get back in the game properly. The inconsitancy in performance from game to game is amazing.
Hey guys,
Any rough ETA on the remainder of B3D's R600 review? :cry:
Tim Murray
30-Jul-2007, 20:08
Hey guys,
Any rough ETA on the remainder of B3D's R600 review? :cry:
We've mixed it up a bit. It's coming. It's not what you're expecting. (It's good. I am writing software that's along the same lines, but the article will probably come first...)
We've mixed it up a bit. It's coming. It's not what you're expecting. (It's good. I am writing software that's along the same lines, but the article will probably come first...)
What a tease! lol......Can you give us any more details to whet our appetites? :grin:
Die_Hard_AMD
14-Aug-2007, 19:14
According to E.Demers R6xx"is a lot of work, by over 300 eng., over 4 years...":???:
http://graphics.stanford.edu/cs448-07-spring/FrontPage?action=AttachFile&do=get&target=Eric_Demers_R6XX.pdf
According to E.Demers R6xx"is a lot of work, by over 300 eng., over 4 years...":???:
http://graphics.stanford.edu/cs448-07-spring/FrontPage?action=AttachFile&do=get&target=Eric_Demers_R6XX.pdf
What's confusing about that statement?
chavvdarrr
23-Aug-2007, 07:37
http://www.computerbase.de/news/software/spiele/actionspiele/2007/august/benchmarks_bilder_bioshock/
...
http://www.computerbase.de/news/software/spiele/actionspiele/2007/august/benchmarks_bilder_bioshock/
...
A test at a single resolution and single AF setting? With 1XAA :?: Not really very useful, but thanks for the link.
I don't get much the german writing, but probably they forced AF through the driver, which is not very optimal -- they should have done it by editing the .ini file, instead.
As for AA, at least under Vista & DX10, it can't be enabled/forced.
DirectX 10 Games vs. Contemporary Graphics Accelerators
http://www.xbitlabs.com/articles/video/display/directx10-games.html
HD2900XT, XT-1024, GF8800GTX, GTS-640, GTS-320. Interesting results...
vertex_shader
26-Aug-2007, 10:27
DirectX 10 Games vs. Contemporary Graphics Accelerators
http://www.xbitlabs.com/articles/video/display/directx10-games.html
HD2900XT, XT-1024, GF8800GTX, GTS-640, GTS-320. Interesting results...
From this test we can say hd2600xt ~1.5-2 times faster in "DX10" games COJ, COH than the 8600gts, but when we look deeper COJ: 11 vs 5 fps, 13.1 vs 6.8 fps, 13.8 vs 7 fps, COH 13 vs 5 fps, 21.2 vs 15.4 fps, the only playable frames 51.5 vs 30.8 in mode3 (model quality set at high, All settings low or off).
This test not looks bad from hd series side, too bad not much playable avarage framerates in the test, but maybe enough for some hint whats coming later when real DX10 games arrive :wink:
Twinkie
26-Aug-2007, 11:33
Yet looking at the nVIDIA drivers used (158.24) im not really convinced from their review that ATi cards are better at DX10. Not to mention ATs article contrasts the xbitlabs one. (Uses 162.18)
http://www.anandtech.com/video/showdoc.aspx?i=3029
Yet looking at the nVIDIA drivers used (158.24) im not really convinced from their review that ATi cards are better at DX10. Not to mention ATs article contrasts the xbitlabs one. (Uses 162.18)
http://www.anandtech.com/video/showdoc.aspx?i=3029
But ATI drivers used where also some beta with some time old.
They didn´t use the Cat. 7.8 that bring lots of performance to HD 2600 series.
That review is now obsolete. Drivers are too old to be meaningful.
Periander
27-Aug-2007, 22:30
That review is now obsolete. Drivers are too old to be meaningful.
Here is a R600 vs. G80 review both recent and relevant.
http://www.firingsquad.com/hardware/bioshock_directx10_performance/
Here is a R600 vs. G80 review both recent and relevant.
http://www.firingsquad.com/hardware/bioshock_directx10_performance/
I disagree. There are several other reviews that have very different numbers in Bioshock. And many people on forums are reporting different numbers for Bioshock (especially on the HD2900XT) which are much better than the FS numbers....
http://www.firingsquad.com/hardware/bioshock_directx10_performance/
About the IQ comparison, the smoke soft edges (a.k.a. Z-feather) is hardly a DX10 "exclusive" feature, as it can be done in any SM2 hardware. So pity...
I disagree. There are several other reviews that have very different numbers in Bioshock. And many people on forums are reporting different numbers for Bioshock (especially on the HD2900XT) which are much better than the FS numbers....
Could you point me into these reviews? I'm going to get a new card soon so Bioshock-numbers are really relevent to me :)
Interesting results:
http://www.gamespot.com/features/6177688/p-6.html
HD2600XT near X1950PRO (probably between X1950GT and X1950PRO), HD2900XT pretty fast under WinXP, but slow under Vista (DX9 and DX10 results are comparable) - bug? Playable on last-gen mainstream - X1650XT 30% faster than 7600GT and X1950PRO 40% faster than 7900GS. Best DX10 performance: GF8800GTX
Rangers
28-Aug-2007, 14:31
So wait a minute, are we seeing the beginning of HD 2900's shaders heavy design paying off?
I mean the Gamespot Bioshock benches put it well ahead of 8800 GTX, never mind GTS, under win XP.
possibly...?
The samplers were designed to be 64b samplers from nearly the beginning, and matching that to BW and keeping the 4:1 ratio on ALU:Tex was the design choice made. In the latest games, where ALU:TEX ratios hit 15~20, this really shines.
http://www.beyond3d.com/content/interviews/39/3
John Reynolds
28-Aug-2007, 15:52
That high quality 2048x1536 difference is interesting, considering the additional 256MB of onboard memory the GTX has over the XT. The 2900s architecture is either really shining or the game has a low texture footprint and the 2900 arch. is still somewhat shiney.
Rangers
28-Aug-2007, 15:56
Well I just remembered the Gamespot Bioshock benches were OXAA, a situation where HD2900 had exceled even in the past.
Crysis demo benchmarks will be interesting though. If any game should stretch the shaders...
That drop in Vista is really nasty, I hope they can do something about that by 7.9 :sad:
vertex_shader
28-Aug-2007, 17:07
I disagree. There are several other reviews that have very different numbers in Bioshock. And many people on forums are reporting different numbers for Bioshock (especially on the HD2900XT) which are much better than the FS numbers....
Where is the other reviews?
vertex_shader
28-Aug-2007, 17:07
That drop in Vista is really nasty, I hope they can do something about that by 7.9 :sad:
Maybe in 7.11 :wink:
Where is the other reviews?
http://www.computerbase.de/news/software/spiele/actionspiele/2007/august/benchmarks_bilder_bioshock/
http://www.digitalreport.net/content/view/52/1/
http://www.gamespot.com/features/6177688/p-6.html
vertex_shader
28-Aug-2007, 19:24
http://www.computerbase.de/news/software/spiele/actionspiele/2007/august/benchmarks_bilder_bioshock/
http://www.digitalreport.net/content/view/52/1/
http://www.gamespot.com/features/6177688/p-6.html
Thanks.
In dx9 have much better performance than 8800gts cards, but except digitalreport DX10 performance test in Bioshock dieing in DX10 what hyped for as where r600 will shine, the whole r600 story day by day more ridiculous in my eyes, i can imagine what come when Crysis demo released end of next month :smile:
vertex_shader: It seems, that the problem isn't DX10, but Vista (driver). Even DX9 under Vista performs poorly according to gamespot. In this light, R600 reminds me R100...
vertex_shader
28-Aug-2007, 19:39
vertex_shader: It seems, that the problem isn't DX10, but Vista (driver). Even DX9 under Vista performs poorly according to gamespot. In this light, R600 reminds me R100...
Than whats the reason AMD Vista drivers hyped as much better than NV?
Depends on particular game.
Than whats the reason AMD Vista drivers hyped as much better than NV?
The performance loses but stability and compatibility is 10 stars.
The NVIDIA drivers used on that review (the latest betas) have curruption in many many many games.
So globaly comparing those 2 drivers the ATI driver are 100x better.
Nvidia launched the best driver for Bioshock, but currupted the other games. That´s the worst thing in the world that they can do.
I hate corruption in games.
Thanks.
In dx9 have much better performance than 8800gts cards, but except digitalreport DX10 performance test in Bioshock dieing in DX10 what hyped for as where r600 will shine, the whole r600 story day by day more ridiculous in my eyes, i can imagine what come when Crysis demo released end of next month :smile:
I think the problem is that you personally have too high expectations for the R600. It just shows in every single post that you make lately. Sorry to say this, but it's like I'm hearing the same record over and over again.... Everytime there's some news about AMD/ATI, there's a predictably sarcastic comment coming from your side.
The facts are.... A GPU that costs €330 (over here in Holland) is equaling or surpassing a €450 card in Bioshock, depending on which review you take or if you listen to the internetforums where people with HD2900XTs are reporting much higher numbers than the Firingsquad review has.
And if you would've done some more research you would've known that the UnrealEngine3 is originally a DX9 engine with DX10 added later on on for performance improvements. Epic even admitted to not having really made use of real DX10 features like geometryshading.
PCGH: Are there any plans at Epic to upgrade the engine for DX 10? Have you already made experience with Microsoft's new API?
Tim Sweeney: Yes, we'll ship Unreal Tournament 3 with full DirectX 10 support. Support for multisampling is the most visible benefit. We're also able to use video memory more efficiently on Windows (http://www.pcgameshardware.de/?article_id=602522#) Vista under DirectX 10, enabling a given machine to use higher-detail texture settings than are possible in Windows Vista under DirectX 9. Most of Unreal Engine 3's effects are bound by fill-rate rather than by higher-level features like hardware geometry processing, so the real impact of DirectX 10 is incrementally better performance rather than entirely new features.We all know that the G80 has more fill-rate to spare so if anything it surprises me to see the cheaper R600 doing so well against it.
vertex_shader
28-Aug-2007, 21:51
I think the problem is that you personally have too high expectations for the R600.
I don't think so its a high expectation a ~380$ card beat the ~290$ card in "dx10" without AA.
It just shows in every single post that you make lately. Sorry to say this, but it's like I'm hearing the same record over and over again.... Everytime there's some news about AMD/ATI, there's a predictably sarcastic comment coming from your side.
I not hide my dissapointment like many others do and not write anything, still nothing happend what promised and some things hyped for nothing.
The facts are.... A GPU that costs €330 (over here in Holland) is equaling or surpassing a €450 card in Bioshock, depending on which review you take or if you listen to the internetforums where people with HD2900XTs are reporting much higher numbers than the Firingsquad review has.
4 Bioshock test linked, check out all, 3 of 4 conclusion for me is 380$ hd2900xt much slower in "DX10" than 290$ 8800gts 320mb.
And if you would've done some more research you would've known that the UnrealEngine3 is originally a DX9 engine with DX10 added later on on for performance improvements. Epic even admitted to not having really made use of real DX10 features like geometryshading.
I know, but real DX10 games won't coming until 2008Q4, so this is pointless reason again, and don't forget how many game coming with U3 engine.
We all know that the G80 has more fill-rate to spare so if anything it surprises me to see the cheaper R600 doing so well against it.
Ok, than you say that a hd2900xt now better buy than a 8800gts640mb?
Edit: I search for EU prices, cheapest hd2900xt (Sapphire) i find in stock 320.04euro (http://www.mix-computer.de/html/product/detail.html?artNo=JBXS1R&artNo=JBXS1R&), cheapest 640mb 8800gts (Leadtek) in stock 299.99 euro (http://www.norskit.com/norskit/default.asp?PageNo=DEFAULT&DeepLink=PA10000D%2CHardware%2CGRAKA%2DNVIC%2DGF88 00110035125%2C) , cheapest 320mb 8800gts (Palit) 241.35euro (http://www.mix-computer.de/html/product/detail.html?artNo=JAXHGA&artNo=JAXHGA&)
I don't think so its a high expectation a ~380$ card beat the ~290$ card in "dx10" without AA.
I don´t understand you :razz:
You see that HD 2900 in DX_9 Win XP makes 20 more fps then in DX_9 in Vista.
So i realy don´t understand why you say that in relation to DX_10, when you know that DX_10 only work on Vista and in Vista the card has a bug or something.
I expect you say something:
Well in Win XP HD 2900 roks and kick 8800 GTX ass.
In Win Vista the card as a bug, so I will wait for next cat release.
It´s Simple....
Dave Baumann
29-Aug-2007, 02:01
You see that HD 2900 in DX_9 Win XP makes 20 more fps then in DX_9 in Vista.
There is some funkiness with how the game is enabling DX9/10 - the results here seem to be that its not actually using DX9 mode, but still DX10. See here (http://www.tweakguides.com/Bioshock_6.html)
We've been trying to replicate Brandon's results from FS and, using his demo, we get the following at 1600x1200...
HD 2900 XT:
WinXP DX9: 89.036
Vista DX9: 88.405
Vista DX10: 51.199
ShaidarHaran
29-Aug-2007, 03:00
There is some funkiness with how the game is enabling DX9/10 - the results here seem to be that its not actually using DX9 mode, but still DX10. See here (http://www.tweakguides.com/Bioshock_6.html)
We've been trying to replicate Brandon's results from FS and, using his demo, we get the following at 1600x1200...
HD 2900 XT:
WinXP DX9: 89.036
Vista DX9: 88.405
Vista DX10: 51.199
If it's not using DX9 in Vista why do your numbers not reflect this?
Dave Baumann
29-Aug-2007, 03:09
The Gamespot article isn't doing DX9, or at least thats what their performances indicate. It is possible to force DX9 by jumping throough some hoops (as the linked article points out). Are numbers are showing actual, forced DX9 performances under Vista.
Firing Squad uses 16xAF. But the in-game default appears to be 4xAF, according to TweakGuide.
Jawed
ShaidarHaran
29-Aug-2007, 04:13
The Gamespot article isn't doing DX9, or at least thats what their performances indicate. It is possible to force DX9 by jumping throough some hoops (as the linked article points out). Are numbers are showing actual, forced DX9 performances under Vista.
Ah, gotcha.
Firing Squad uses 16xAF. But the in-game default appears to be 4xAF, according to TweakGuide.
Jawed
AF doesn't have that large of an impact on performance anymore, does it? Heck, I run 16x HQ AF on my X1650 XT with no drop in fps compared to no AF in the only game I play anymore (MSFS 2004). Granted it's just one title but given the draw distance I would imagine Flight Sim would be just about one of the best AF perf tests out there.
AF shouldn't have a large hit on modern GPUs. However the 2900 series is TMU-limited so AF takes a larger hit than we're used to seeing on performance parts.
Perhaps it depends if FS used the in-game AF or forced it through CP. Forcing AF through the CP tends to have a higher performance hit than doing so with the game's own control (generalising to other games, not Bioshock).
Does Bioshock provide user-controlled AF? I don't remember seeing it in the demo's graphics options - it seems to be limited to an ini file change. Also, if the developers decided upon "fixed" 4xAF, is there an IQ benefit to 16xAF?
So, anyway, I'm just trying to present reasons for the performance FS reports.
Jawed
ShaidarHaran
29-Aug-2007, 04:20
Perhaps it depends if FS used the in-game AF or forced it through CP. Forcing AF through the CP tends to have a higher performance hit than doing so with the game's own control.
Does Bioshock provide user-controlled AF? I don't remember seeing it in the demo's graphics options - it seems to be limited to an ini file change. Also, if they've decided upon 4xAF, is there an IQ benefit to 16xAF?
So, anyway, I'm just trying to present reasons for the performance FS reports.
Jawed
No option for in-game AF in MSFS 2004, so forced via CP.
But you were talking about FiringSquad, not Flight Sim :p lol, alphabet soup leads to lovely confusion
Dave Baumann
29-Aug-2007, 04:33
So, anyway, I'm just trying to present reasons for the performance FS reports.
Thats not the issue. Far be it for me to speak for Brandon, but I believe he's now seeing a performance trend a bit more inline with what I displayed earlier.
vertex_shader
29-Aug-2007, 09:39
I don´t understand you :razz:
You see that HD 2900 in DX_9 Win XP makes 20 more fps then in DX_9 in Vista.
So i realy don´t understand why you say that in relation to DX_10, when you know that DX_10 only work on Vista and in Vista the card has a bug or something.
Bug? hotfix released for the game, how can be this is a Vista bug when its not affect NV cards performance in the same way?
I expect you say something:
Well in Win XP HD 2900 roks and kick 8800 GTX ass.
In Win Vista the card as a bug, so I will wait for next cat release.
I never said anything like this, we are talking about now for the latest BB game and the hd2900xt performance without AA against other cards (8800gts 320/640mb) in DX10.
This is a "DX10" game in the same way as coh,coj,lost planet but the difference is this game have playable performance with highest setting even at higher resolutions.
The Gamespot article isn't doing DX9, or at least thats what their performances indicate. It is possible to force DX9 by jumping throough some hoops (as the linked article points out). Are numbers are showing actual, forced DX9 performances under Vista.
I think this is merely a case of reviewer ignorance. It's not like it's that difficult to right click the game's shortcut and select "Play in DX9 mode" from there. Limiting oneself to the shiny button in the game's menu is....unimpressive.
That being said, is there any reason for the large performance drop you're experiencing with DX10?The ammount of added work is not that huge, and nV seems to be able to keep de delta between DXs fairly reduced. Is this another case of driver rawness?
nicolasb
29-Aug-2007, 10:40
HD 2900 XT:
WinXP DX9: 89.036
Vista DX9: 88.405
Vista DX10: 51.199So the conclusion we should draw is not "R600 really sucks under Vista" but "R600 really sucks under DX10"? Is that roughly what you're saying, Dave? :devious:
In all seriousness, while it pains me to have to agree with vertex_shader about anything, he does kind of have a point here: if you look at performance comparisons between 8800GTX and 2900XT (not just for Bioshock but in general), the gap seems to be wider under DX10 than it is under DX9; this rather flies in the face of the attempt by ATI fans (and indeed by ATI itself) to make out that R600 is a "forward-looking design" which will show its true strengths with DX10 games.
So the conclusion we should draw is not "R600 really sucks under Vista" but "R600 really sucks under DX10"? Is that roughly what you're saying, Dave? :devious:
In all seriousness, while it pains me to have to agree with vertex_shader about anything, he does of kind of have a point here: if you look at performance comparisons between 8800GTX and 2900XT (not just for Bioshock but in general), the gap seems to be wider under DX10 than it is under DX9; this rather flies in the face of the attempt by ATI fans (and indeed by ATI itself) to make out that R600 is a "forward-looking design" which will show its true strengths with DX10 games.
But the true strenghts will show in the TRUE DX10 games-available in a couple of years. And anyway, this game is TWIMTBP, so it automagically is evil and not valid for comparisons, yes?J/K:D
if you look at performance comparisons between 8800GTX and 2900XT (not just for Bioshock but in general), the gap seems to be wider under DX10 than it is under DX9; this rather flies in the face of the attempt by ATI fans (and indeed by ATI itself) to make out that R600 is a "forward-looking design" which will show its true strengths with DX10 games.
Todays "DX10" games are mostly based on SM3.0 engines, which weren't build with DX10 on mind. They hardly use any exclusive DX10 features (most of their "DX10" efects are better textures, longer view distance, more vegetation, some silly effects, which are possible even on SM2.0 hardware etc.), so why do you think that these games indicates which design is more forward looking?
vertex_shader
29-Aug-2007, 12:06
Todays "DX10" games are mostly based on SM3.0 engines, which weren't build with DX10 on mind. They hardly use any exclusive DX10 features (most of their "DX10" efects are better textures, longer view distance, more vegetation, some silly effects, which are possible even on SM2.0 hardware etc.), so why do you think that these games indicates which design is more forward looking?
This games only what we can compare now, i except real dx10 games maybe in 2008Q4, but than R600 will be only a part of the GPU history, users who buy hd2900xt and not want change the card for 2 years can hope for better days , longterm r600 can be a better deal (i except ALU ratio start rising against TEX), but i don't think so too many user in this price range buy cards for 2 years.
It's clear, that DX10 will, for foreseeable future, be used only as a cherry top addition to the main engine.
And this is not only because the DX10 critical mass is so shallow (and I mean DX10 userbase = Vista + SM4 hardware combo), but if we look from developer PoV, a fully loaded DX10 title/engine profile will just torn down any successful attempt for easy backward compatibility. Just imagine tons of geometry amplification, huge number of state changes/calls, miles long shader programs w/ export (vitalization anyone), colour component differential processing and so on...
I've asked this quite often, but no one seems to explain to lil' old me how is it that if you're not quite stellar doing DX9 stuff, you can become a god in DX10, simply by the API shift. The GS argument doesn't hold that much water.
The paradigm shift between DX9 and DX10 isn't that huge, so why does the above line of reasoning even exist?If it should be read like:future games that care only about sheer math-that's wishful thinking ATM, we don't know wheter or not the gaming-industry will go the way ATi would like it to(I kindof doubt that, but I digress).
Todays "DX10" games are mostly based on SM3.0 engines, which weren't build with DX10 on mind. They hardly use any exclusive DX10 features (most of their "DX10" efects are better textures, longer view distance, more vegetation, some silly effects, which are possible even on SM2.0 hardware etc.), so why do you think that these games indicates which design is more forward looking?
That's not entirely true. For example Company of Heroes uses the GS to render point-light cube-shadowmaps in single pass and Bioshock uses the shader accessible zbuffer for soft particles.
Even FarCry used some SM3.0 features, but do we consider it as typical SM3.0 game from todays standpoint?
ShaidarHaran
29-Aug-2007, 22:33
I've asked this quite often, but no one seems to explain to lil' old me how is it that if you're not quite stellar doing DX9 stuff, you can become a god in DX10, simply by the API shift. The GS argument doesn't hold that much water.
The paradigm shift between DX9 and DX10 isn't that huge, so why does the above line of reasoning even exist?If it should be read like:future games that care only about sheer math-that's wishful thinking ATM, we don't know wheter or not the gaming-industry will go the way ATi would like it to(I kindof doubt that, but I digress).
What do you mean the GS argument doesn't hold water? We've got one synthetic benchmark that uses GS extensively (rightmark 2.0) and it does show R600 hardware performing quite admirably (out-performing GF 8800 Ultra in some cases). There certainly aren't any "real-world" examples of extensive GS-usage out there to go by, so until some come along I think it's too early to say this argument doesn't hold any water.
The paradigm shift between DX9 and DX10 isn't that huge, so why does the above line of reasoning even exist?If it should be read like:future games that care only about sheer math-that's wishful thinking ATM, we don't know wheter or not the gaming-industry will go the way ATi would like it to(I kindof doubt that, but I digress).
Judging by the performance difference between DX9 and DX10 at the moment with little difference in quality I'd say the shift is rather large.
What do you mean the GS argument doesn't hold water? We've got one synthetic benchmark that uses GS extensively (rightmark 2.0) and it does show R600 hardware performing quite admirably (out-performing GF 8800 Ultra in some cases). There certainly aren't any "real-world" examples of extensive GS-usage out there to go by, so until some come along I think it's too early to say this argument doesn't hold any water.
Why am I not surprised you're the one pointing this out?If you believe that the a synthetic-test(quite poorly set-up in terms of actually isolating and taking advantage of the GS, BTW, at least the Cosmos one is borderline useless) where a single feature in isolation is tested, is relevant for how software that uses a myriad of features in interaction, will behave?
Let me put it differently(I've done it before, another thing I didn't get a proper answer for). If in a scene the GS is accountable for let's say, 10% rendering time, and on the 2900 it does that rendering 100% faster than competing products, does that mean it'll be faster overall, if it is slower at the things accounting for the other 90% of rendering time?(which your precious RightMark seems to point to...but I guess it's other benchmarks suddenly become invalid beyond the GS tests, yes?:) ).
By the time the GS will become a real binding element-that means, when amplification(which is one of the points where theoretically the R600 should spank the entire G8x line due to its way of handling the generated data) enters significant use, and GS using algorithms are used without second thought-current parts will be obsolete. So please, tell me again, how the GS argumentation is relevant for the current-gen R600?Beyond having 5 fps rather than 3(ZOMG, almost twice as fast...better slideshow) on the G80 in Crysis 2 released in 2010?Considering that will actually be the case.
3dcgi, dunno...consider titles that have actually tried to keep the things they do in both DX9 and DX10 paths similar(like World in Conflict, Bioshock and Lost Planet without the DX10 added-features enabled). Performance is equal/better, on the G80(let's assume ATi's driver still suck ass, and in a year's time they'll get around to fixing them and the current situation with significant drops is not due to some architectural quirk). The thing is that the situation was a bit of a conundrum:if devs were to implement DX10 paths solely for improving performance, with no visual bonus, everyone now screaming:"OMG poor performance!MS is the suxxor" would've cried about the fact that there are no visual improvements. It's just the way it is, sadly, so some chose to add things in. Which make comparisons rather useless.
Take COJ for example:it's a complete overhaul-new textures, new materials, new way of handling particles etc. That hardly makes for valid comparisons, as new algorithms in DX10 are...well, new, there's no prior experience with em', and generally first iterations are less than optimal. So combine that with the fact that there's more work to be done through the DX10 path and the driver's current status(I'm tempted to say from both IHVs, but nV's strong devrel masks this well, having the ability to work closely with devs and combat bugs in their infancy), and the performance delta experience is not due to the fact that you're doing things wholly differently through DX10, and thus it's a big murky unknown in terms of performance. IMHO.
Judging by the performance difference between DX9 and DX10 at the moment with little difference in quality I'd say the shift is rather large.
It's not. Direct3D 10 is faster or equally fast than Direct3D 9 (Vista!) in games that you can setup to render the same with both APIs.
You still have your vertexbuffers, rendertargets, textures and shaders like in D3D9. You just can do much more with them in a drawcall.
Tim Murray
30-Aug-2007, 17:44
Beyond3D R600 IQ investigation piece is here. It's a little different than what you were expecting. (http://www.beyond3d.com/content/reviews/47)
Comment thread is here. (http://forum.beyond3d.com/showthread.php?p=1057591#post1057591)
It's not. Direct3D 10 is faster or equally fast than Direct3D 9 (Vista!) in games that you can setup to render the same with both APIs.
You still have your vertexbuffers, rendertargets, textures and shaders like in D3D9. You just can do much more with them in a drawcall.
Guess I haven't seen those benchmarks.
I'm still of the opinion that DX10 is a large shift. My previous comment about the performance delta between DX9 and DX10 doesn't really address the real differences, which were changes to the hardware that were necessary to support DX10. Both DX9 and DX10 run on the same hardware so benchmarks can't show this.
Features like unlimited length shaders necessitated having instruction caches in the shader core. I doubt previous generations of GPUs had CPU like instruction caches. The geometry shader and other features were also fundamental to the rise in transistor count that we saw with the DX10 generation of chips.
Anyway, this is getting off topic and not a particularly important issue so I'll leave it at that.
ShaidarHaran
31-Aug-2007, 05:35
Why am I not surprised you're the one pointing this out?If you believe that the a synthetic-test(quite poorly set-up in terms of actually isolating and taking advantage of the GS, BTW, at least the Cosmos one is borderline useless) where a single feature in isolation is tested, is relevant for how software that uses a myriad of features in interaction, will behave?
Let me put it differently(I've done it before, another thing I didn't get a proper answer for). If in a scene the GS is accountable for let's say, 10% rendering time, and on the 2900 it does that rendering 100% faster than competing products, does that mean it'll be faster overall, if it is slower at the things accounting for the other 90% of rendering time?(which your precious RightMark seems to point to...but I guess it's other benchmarks suddenly become invalid beyond the GS tests, yes?:) ).
By the time the GS will become a real binding element-that means, when amplification(which is one of the points where theoretically the R600 should spank the entire G8x line due to its way of handling the generated data) enters significant use, and GS using algorithms are used without second thought-current parts will be obsolete. So please, tell me again, how the GS argumentation is relevant for the current-gen R600?Beyond having 5 fps rather than 3(ZOMG, almost twice as fast...better slideshow) on the G80 in Crysis 2 released in 2010?Considering that will actually be the case.
3dcgi, dunno...consider titles that have actually tried to keep the things they do in both DX9 and DX10 paths similar(like World in Conflict, Bioshock and Lost Planet without the DX10 added-features enabled). Performance is equal/better, on the G80(let's assume ATi's driver still suck ass, and in a year's time they'll get around to fixing them and the current situation with significant drops is not due to some architectural quirk). The thing is that the situation was a bit of a conundrum:if devs were to implement DX10 paths solely for improving performance, with no visual bonus, everyone now screaming:"OMG poor performance!MS is the suxxor" would've cried about the fact that there are no visual improvements. It's just the way it is, sadly, so some chose to add things in. Which make comparisons rather useless.
Take COJ for example:it's a complete overhaul-new textures, new materials, new way of handling particles etc. That hardly makes for valid comparisons, as new algorithms in DX10 are...well, new, there's no prior experience with em', and generally first iterations are less than optimal. So combine that with the fact that there's more work to be done through the DX10 path and the driver's current status(I'm tempted to say from both IHVs, but nV's strong devrel masks this well, having the ability to work closely with devs and combat bugs in their infancy), and the performance delta experience is not due to the fact that you're doing things wholly differently through DX10, and thus it's a big murky unknown in terms of performance. IMHO.
Way to completely miss the point. That being it's too early to come to a conclusion about GS performance.
I wasn't saying "ZOMG, R600 will pwn j00 w/1337 GS powerzzzzz1111!!11one", as you seem to think I was.
Way to completely miss the point. That being it's too early to come to a conclusion about GS performance.
I wasn't saying "ZOMG, R600 will pwn j00 w/1337 GS powerzzzzz1111!!11one", as you seem to think I was.
If you read carefully, you might've noted that the argumentation was more along the lines that no matter how the GS will be employed, it can not be a "saving grace", nor is it likely to be the prime performance determinant for most titles coming in a reasonable time-frame. Both nV and ATi may well be at gen3 DX10 HW when the GS becomes heavily used and amplification, for example, is generally implemented in games.
I still think that ATi's drivers are bugged somehow, because the significant drop when shifting to DX10 can't be explained otherwise...but it's worrysome that they don't seem to have made large strides in improving this. However, I do not think that my twin 2900s will beat twin GTXs or Ultras at any point in the future, except for some corner-case scenarios-this does not make me any less pleased with them, as I feel they're good cards, in spite of everything.
wide_screen
31-Aug-2007, 13:44
Bit-tech Bioshock gameplay/graphics/performance test (http://www.bit-tech.net/gaming/2007/08/30/bioshock_gameplay_graphics_and_performance)
DX9 hd2900xt beat the 8800ultra with 16x af and maximum ingame setting in 1600x1200/1680x1050 and in pair in 1920x1200.
DX10 8800gts320mb oc beat the hd2900xt.
Looks like a rollercoaster, driver problem? architectural? or other?
Rangers
31-Aug-2007, 14:43
So the reviews trend is now clear, with new Firing Squad updates as well, R600 series is faster in DX9 Bioshock, G80 is faster in DX10.
First off, I think it's a good accomplishment for AMD to even be faster at DX9, given HD2900's price slot. Heck the HD2900 512 even beats the $800 8800GTX Ultra in bit tech's DX9 Bioshock benches. But next, I have to ask why DX10 is so slow relatively on AMD. Drivers? Or could there be some architectural reason?
Prolly Z fillrate - D3D10 graphics options in a variety of games seem to use "nice particle effects", which as far as I can tell uses additional Z rendering (or is that just testing?) to produce the effect. Also, shadow maps tend to increase in size, which also increases Z fillrate.
Though shadow map filtering is mostly a texture fetch bottleneck and D3D10 graphics options normally include nicer shadow map filtering.
Jawed
BTW: The 8800 GTX Ultra has dropped in price down to the $580-$600 (http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=2000380000+1067928976&Tpk=8800+ultra) mark. It's still incredibly expensive, but it's no where near the insanity of $800 and hasn't been for months now.
trinibwoy
01-Sep-2007, 13:14
Prolly Z fillrate - D3D10 graphics options in a variety of games seem to use "nice particle effects", which as far as I can tell uses additional Z rendering (or is that just testing?) to produce the effect. Also, shadow maps tend to increase in size, which also increases Z fillrate.
Yeah but why would Nvidia's numbers go up in DX10 in that case?
Yeah but why would Nvidia's numbers go up in DX10 in that case?
They did? I hadn't noticed. Perhaps it's some of that D3D10 reduced-overhead stuff kicking in.
Jawed
Dave Baumann
01-Sep-2007, 16:47
Feedback I got from anyone that was running the actual DX9 mode on NVIDIA was "its very unstable" (thats why Brandon didn't include them in the midrange). It looks to me like NVIDIA only worked on the DX10 path for these beta drivers and I wouldn't say their DX9 to DX10 is indicative of too much.
ChrisRay
01-Sep-2007, 18:01
Feedback I got from anyone that was running the actual DX9 mode on NVIDIA was "its very unstable" (thats why Brandon didn't include them in the midrange). It looks to me like NVIDIA only worked on the DX10 path for these beta drivers and I wouldn't say their DX9 to DX10 is indicative of too much.
Bull. Install the updates from microsoft and you are fine.
http://support.microsoft.com/kb/940105
http://support.microsoft.com/kb/936710
http://support.microsoft.com/kb/938194
http://support.microsoft.com/kb/938979
DX 9.0 16xAA/16xAF support from my system under Vista Ultimate 64 bit with said patches applied. No problems whatsoever.
http://img381.imageshack.us/img381/9518/dx916xsingleuc7.th.jpg (http://img381.imageshack.us/my.php?image=dx916xsingleuc7.jpg)
SLI
http://img106.imageshack.us/img106/6531/dx916xscalingjz2.th.jpg (http://img106.imageshack.us/my.php?image=dx916xscalingjz2.jpg)
If you dont have these hot fixes installed. You will have problems. Paticularly with memory management.
Feedback I got from anyone that was running the actual DX9 mode on NVIDIA was "its very unstable" (thats why Brandon didn't include them in the midrange). It looks to me like NVIDIA only worked on the DX10 path for these beta drivers and I wouldn't say their DX9 to DX10 is indicative of too much.
Well, I beat Bioshock under XP with the Beta drivers and had zero crashes or any quirks at all. I even forced 4X AA and 16X AF.
Under Vista, the game is unstable if I try to force AA under DX10 mode. Things don't render quite right (almost unnoticeable) and it crashes within a minute. But that's with me forcing things not meant to be used.
Dave Baumann
01-Sep-2007, 20:26
Bull. Install the updates from microsoft and you are fine.
Its not bull that thats the feedback I got.
http://www.firingsquad.com/hardware/bioshock_mainstream_gpu_performance/page6.asp
Unfortunately, the GeForce testbed wouldn’t cooperate with our pure DX9 testing. Regardless of what we tried, the game continually locked up during our FRAPS walkthroughs with the GeForce 8 cards when we forced the DX9 path. We couldn’t complete a single run without the game locking up. As soon as we went back to the DX10 path, or the hybrid DX9/DX10 path, stability came back 100%, so we’re not sure if we just encountered an issue with our particular testbed or if the problem lies elsewhere.
Evidently Brandon has it running now though and the DX9 and DX10 numbers are still very similar.
Dave Baumann
01-Sep-2007, 20:39
Under Vista, the game is unstable if I try to force AA under DX10 mode. Things don't render quite right (almost unnoticeable) and it crashes within a minute. But that's with me forcing things not meant to be used.
AFAIK DX10 AA doesn't work with this.
ChrisRay
01-Sep-2007, 22:05
Its not bull that thats the feedback I got.
http://www.firingsquad.com/hardware/bioshock_mainstream_gpu_performance/page6.asp
Evidently Brandon has it running now though and the DX9 and DX10 numbers are still very similar.
No. He may very well have had problems. If you dont have Vista up to date with its latest hotfixes. its entirely possible. Your assumption that Nvidia did not work on DX 9.0 is bull.
Dave Baumann
01-Sep-2007, 22:16
Whether or not they did "no" work is is by the by, I still don't think that the DX9 vs DX10 performance of NVIDIA is indicative to too much - for one the game is doing different things between the two paths (AFIAK its using deferred shadowing, which is why we see the detail differences between the to render paths on the shadows, and probably why AA doesn't work) and HD 2900 XT has pretty similar performance to the 8800 Ultra under DX9.
Whether or not they did "no" work is is by the by, I still don't think that the DX9 vs DX10 performance of NVIDIA is indicative to too much - for one the game is doing different things between the two paths (AFIAK its using deferred shadowing, which is why we see the detail differences between the to render paths on the shadows, and probably why AA doesn't work) and HD 2900 XT has pretty similar performance to the 8800 Ultra under DX9.
You know that what you're saying can be read as:nV has a lot of optimizing left to do for the DX9 path, yes?And that it can be interpreted as a call to moderation for all of the zealots who are heralding the 2900s as the second coming for really complex games, because it does well in Bioshock on the DX9 path(I'm not exxagerating this, skimming some threads on [H] and R3D is borderline painful)?And if you're saying that, I'm beyond baffled, given the fact that you're an ATi employee:D
Dave Baumann
01-Sep-2007, 23:00
Well, I'm perfectly happy if people are coming away with this thinking that HD 2900 XT is as fast as 8800 Ultra! 8-)
But, by the same token, note I've not said that we're tapped out in performance in this title either. ;)
ChrisRay
01-Sep-2007, 23:29
Whether or not they did "no" work is is by the by, I still don't think that the DX9 vs DX10 performance of NVIDIA is indicative to too much - for one the game is doing different things between the two paths (AFIAK its using deferred shadowing, which is why we see the detail differences between the to render paths on the shadows, and probably why AA doesn't work) and HD 2900 XT has pretty similar performance to the 8800 Ultra under DX9.
So in other words. You dont know. And you are presuming that the game isn't optimised based on some crashes FS and perhaps some other editors may have had. Going as far as saying Nvidias DX 9.0 performance isnt indicative of anything. Yet most people who have installed the relevant microsoft hot fixes are having no issues. Also there are other difference between the render paths than just shadows. For instance how the depth buffers accessed in DX10 verses DX 9.0. To help blending of soft particles as they intersect geometry.
Besides that. I dont see how you can formulate a conclusion based upon such incomplete evidence.
1) Nvidia has already applied its AA hotfixes for HDR + AA with this title. Available with the latest Nvidia drivers for both Windows XP and Windows Vista
2) Nvidia has to support Windows XP. How could you possibly conclude that they have ignored DirectX 9.0?
3) The game runs fine Under Vista in DX 9.0 mode in SLI, Single GPU, and AA situations with the relevant hot fixes. It also runs fine under XP.
4) There are are 163.44 drivers available for Windows XP and Vista. Your not honestly going to tell me that Nvidia released XP drivers optimised for DirectX10 are you?
You know that what you're saying can be read as:nV has a lot of optimizing left to do for the DX9 path, yes?And that it can be interpreted as a call to moderation for all of the zealots who are heralding the 2900s as the second coming for really complex games, because it does well in Bioshock on the DX9 path(I'm not exxagerating this, skimming some threads on [H] and R3D is borderline painful)?And if you're saying that, I'm beyond baffled, given the fact that you're an ATi employee
Regardless of the competitive arguments and the whole "ATI Verses Nvidia" rhetoric going on here. Why is anyone surprised about how the G80/R600 compare to each other with AA disabled?
nicolasb
01-Sep-2007, 23:39
I still don't think that the DX9 vs DX10 performance of NVIDIA is indicative to too muchBut DX9 vs DX10 performance of R600 is indicative, you mean...?
trinibwoy
01-Sep-2007, 23:40
Why is anyone surprised about how the G80/R600 compare to each other with AA disabled?
Simple. R600 usually performs below par. So when it performs as it should it's a ZOMG surprise and accolades all around. Amusing to watch as always though :)
Regardless of the competitive arguments and the whole "ATI Verses Nvidia" rhetoric going on here. Why is anyone surprised about how the G80/R600 compare to each other with AA disabled?
I don't think I worded it properly. I read what Dave said to mean that currently nV has lacking DX9 performance because they didn't optimize that path all that much, and that the current performance outlook in XP, with the 2900 being on par with the Ultra is flawed due to that, which given his current employer would've been odd:D.
I do think that it's a safe bet that optimising on both sides of the fence is still in its infancy for the new titles just now hitting the market, even with nV's strong devrel, so probably performance will keep on going up for a while in these with both the R600 and the G80.
AFAIK DX10 AA doesn't work with this.
Ooops. I should mention I'm using a GF 8800. I was just trying to say that the game itself doesn't seem unstable on my hardware under XP DX9. I forgot that this was a HD 2900 thread. :lol:
And yeah, I get no AA if I try to force it in DX10 mode. Because I prefer AA over some extra particles or water ripples, I just went right to XP and forced AA.
I doubt previous generations of GPUs had CPU like instruction caches.
FWIW, NV30 had a shader instruction cache. You may not think of it as "CPU like", but it's a cache backed by framebuffer memory nonetheless.
ChrisRay
02-Sep-2007, 03:45
Ooops. I should mention I'm using a GF 8800. I was just trying to say that the game itself doesn't seem unstable on my hardware under XP DX9. I forgot that this was a HD 2900 thread. :lol:
And yeah, I get no AA if I try to force it in DX10 mode. Because I prefer AA over some extra particles or water ripples, I just went right to XP and forced AA.
You dont have to use XP to force AA. Just use your games explorer. Right Click Bioshock and play DX 9.0 mode. And you can use AA under Vista in DX 9.0 mode. Just remember to download the patches I pointed too in the earlier part of this thread. Could be important from a stability point of view.
Chris
CarstenS
04-Sep-2007, 07:28
There are certain artifacts though. Probably one of the reasons, take2 didn't want the support calls to ring on their phones. :)
I wonder if those artifacts might vanish with the use of a SM2.0b card and the shader mod... Hmm.
http://www.elitebastards.com/cms/index.php?option=com_content&task=view&id=455&Itemid=27&limit=1&limitstart=10
It's also difficult to get a real handle on the Radeon HD 2900 XT's performance and potential from our driver comparison today - In some titles, Catalyst 7.8 shows some decent performance gains over the launch driver for the board, but in others (particularly two of our DirectX 10 titles) performance has actually dropped somewhat. It appears that it may still be a while before we get an entirely stable picture of what the Radeon HD 2900 XT can do.
Jawed
ShaidarHaran
04-Sep-2007, 23:44
http://www.elitebastards.com/cms/index.php?option=com_content&task=view&id=455&Itemid=27&limit=1&limitstart=10
Jawed
Wow. Unbelievable. AMD GPG needs to GITG. I understand it's hard work creating a graphics driver in the first place, let alone optimizing for new games and their new API code paths, but come on!
vertex_shader
12-Sep-2007, 21:09
http://www.theinq.com/images/articles/FireStream_R600.jpg
Link (http://www.theinq.com/default.aspx?article=42297)
R600 based steam computing card with passive cooling, maybe not 80nm based? or GPU runs at 125 celsius with this cooler? :smile:
vertex_shader
12-Sep-2007, 22:24
MOH Airborne retail performance review (http://www.legionhardware.com/document.php?id=683&p=0)
http://aycu22.webshots.com/image/27861/2002304436605108002_rs.jpg
http://aycu11.webshots.com/image/28450/2002347037614327376_rs.jpg
Same engine as Bioshock using.
HD2900 sure looks good when it doesn't have to deal with AA.
vertex_shader
12-Sep-2007, 22:38
HD2900 sure looks good when it doesn't have to deal with AA.
UE3 engine not support AA in dx9, and no AA in dx10 yet.
pjbliverpool
12-Sep-2007, 23:14
Its nice to see a game that gains performance in DX10 for a change. And suprising to see that gain on the X2900XT!
embargiel
13-Sep-2007, 00:39
My apology for going slightly off-topic, but is the DX10 mode for Airborne supposed to be any better looking than the DX9? I clearly don't see it being better looking than the DX9 mode from the screenshot comparisons.
ShaidarHaran
13-Sep-2007, 03:35
MOH Airborne retail performance review (http://www.legionhardware.com/document.php?id=683&p=0)
Same engine as Bioshock using.
Nice find. Glad to see a title that the HD 2900 can handle in DX10 mode. Very pleased to see it out-performing the DX9 code path and beating the 8800 GTX while it's at it.
vertex_shader
13-Sep-2007, 11:43
Other side of the coin in Enemy Territory: Quake Wars
http://www.firingsquad.com/hardware/enemy_territory_quake_wars_performance/images/qw1600.gif
4xAA, 16xAF
http://www.firingsquad.com/hardware/enemy_territory_quake_wars_performance/images/etqw1600.gif
Link (http://www.firingsquad.com/hardware/enemy_territory_quake_wars_performance/)
pjbliverpool
13-Sep-2007, 12:14
Other side of the coin in Enemy Territory: Quake Wars
Link (http://www.firingsquad.com/hardware/enemy_territory_quake_wars_performance/)
Hopefully they will fix that with drivers by the time the game is released. It does point to what IMO is the 2900XT's biggest flaw though - inconsistency.
It seems in practically every benchmarking review I come across these days the 2900XT has problems with either performance, rendering, crossfire, DX10 etc..... For every one game it runs flawlessly there seem to be 3 were it has issues.
They are probably just early driver issues which are solved within a few weeks but the problem is that all the benchmarks reviews are done when a game is brand new and thus I never get to see the fixes (if they even exist) and im left with a bad taste in my mouth when I think of the 2900XT.
I really hope R700 is some kind of crazy monster from hell that destroys G90 in every way. The industry really needs it!
MOH Airborne retail performance review
The DX9 path is under WinXP whereas the DX10 path is under Vista.
ETQW uses Doom3 engine so I wonder if this is some OpenGL weakness on the part of 2900. NV has always been really strong with that engine.
The DX9 path is under WinXP whereas the DX10 path is under Vista.
Did EA officialy state this?I've seen about 0 mentions WRT DX10 and this title, and I would've guessed they'd make a little more noise about it. Granted, I didn't follow MOH:A all that closely.
The Legion Hardware review that v_s linked mentions DX9 running in XP and DX10 running in Vista 64.
Is ATI still considered deficient in OGL? I thought that rep was behind them, given the huge strides they'd made recently. Or are they still sub-par at launch, with significant performance increases expected with later drivers?
The Legion Hardware review that v_s linked mentions DX9 running in XP and DX10 running in Vista 64.
Is ATI still considered deficient in OGL? I thought that rep was behind them, given the huge strides they'd made recently. Or are they still sub-par at launch, with significant performance increases expected with later drivers?
I think they have issues with ET:QW in particular, more than a general issue with OGL. The game and the 2900 haven't really gotten along nicely since the first public betas were released, when it used to crash with FSAA enabled, and they particle smoothing feature was greyed-out.
Dave Baumann
13-Sep-2007, 19:33
ETQW is being affected by a texture upload issue on the current drivers, if you force the app to pre-load the textures [image_useBackgroundLoads 0] then then performances increase, although I don't know to what degree on HD 2900 XT.
vertex_shader
13-Sep-2007, 22:07
ETQW is being affected by a texture upload issue on the current drivers, if you force the app to pre-load the textures [image_useBackgroundLoads 0] then then performances increase, although I don't know to what degree on HD 2900 XT.
So the hotfix more a coldfix :sad:
Next driver ~4weeks away, game coming out in september 28th, what will do AMD release a hotfix2 for ET:QW? or users need to wait +~2weeks after game released to the proper driver?
fivefeet8
14-Sep-2007, 07:19
MOH Airborne retail performance review (http://www.legionhardware.com/document.php?id=683&p=0)
Same engine as Bioshock using.
I have the game and there's no option to enable Dx10 in the game. Running Vista32bit with 8800gtx here. The game setup application also shows only Pixel/Vertex shader 3 as available in the "system" tab.
Unless they can say how they enabled it, I think they may be confused and thought that running the game in Vista automatically makes it run in Dx10 which isn't the case.
Did EA officialy state this?I've seen about 0 mentions WRT DX10 and this title, and I would've guessed they'd make a little more noise about it. Granted, I didn't follow MOH:A all that closely.
No idea. I just RTFA, where they give the specs of the systems used for testing.
No idea. I just RTFA, where they give the specs of the systems used for testing.
I think they're pulling stuff out of their arse, basing it on how Bioshock behaves and the fact that both games use UE3. Unless they've been in contact with the Dev and that's how they got their info. Which is a tad bit doubtful.
vertex_shader
18-Sep-2007, 11:35
Crysis beta2 on 2900XT with C2Q 3G ALL HIGH (http://en.hardspell.com/doc/showcont.asp?news_id=1440)
Single frame with fraps benchmarking :smile:
Damn well for the red team. And it's still in beta stage. ;)
Am I the only one who's a tad bit underwhelmed by the Crysis beta stuff?The pics I've seen are rather meh, not some great outstanding bar-raising thing. OTOH, I understand that some stuff is disabled for the beta, so maybe significant improvements are in store?
AlStrong
18-Sep-2007, 15:51
It was a bit boring to look at. I think we'd all be more impressed with single-player footage though. ;)
Well, I guess you have to experience the game in first hand to appreciate it in proper fashion. ;)
Well, I guess you have to experience the game in first hand to appreciate it in proper fashion. ;)
You're one of them experiencers, eh?Gah, us poor pissants have to wait till the 25th to join your elitist bunch:D
No, I just have rather bold imagination. :D
I watched some people playing the game at CES and was rather impressed with it back then. Obviously if you've let yourself be enticed too much by bullshots, you are going to end up disappointed a bit. Every game has flaws.
Crysis beta2 on 2900XT with C2Q 3G ALL HIGH (http://en.hardspell.com/doc/showcont.asp?news_id=1440)
Single frame with fraps benchmarking :smile:
That is definitely not high settings, that since I am beta testing Crysis to! *Seals lips*
ShaidarHaran
26-Sep-2007, 00:10
That is definitely not high settings, that since I am beta testing Crysis to! *Seals lips*
So the first screenshot showing all the settings on "high" was fake? And please, don't come out and say "I'm beta-testing Crysis" and then expect everyone to just take your word for it and allow you to end the convo at that. Besides, usually non-public beta-testing requires one to sign an NDA which includes a clause about not discussing it publicly...
So the first screenshot showing all the settings on "high" was fake? And please, don't come out and say "I'm beta-testing Crysis" and then expect everyone to just take your word for it and allow you to end the convo at that. Besides, usually non-public beta-testing requires one to sign an NDA which includes a clause about not discussing it publicly...
Well as you said im under NDA so I cant back it up with ss. And to show a ss with menu settings doesn't mean that the ingame ss are the same settings. But neverthless I'll make a nice Crysis beta screenshot album for you when the NDA is lifted. :wink:
Silent_Buddha
28-Sep-2007, 13:16
I watched some people playing the game at CES and was rather impressed with it back then. Obviously if you've let yourself be enticed too much by bullshots, you are going to end up disappointed a bit. Every game has flaws.
Well, considering I've managed to avoid looking at any screens or video related to Crysis, hopefully I'll be suitably impressed when the game launches. :)
Although to be honest, I'm more looking forward to Hellgate: London.
Ah, good to be back in the US, now to head to a store and pick up Bioshock finally. :D
Regards,
SB
vertex_shader
08-Oct-2007, 22:11
HD2900XT TriFire (http://www.theinquirer.net/gb/inquirer/news/2007/10/07/sapphire-shows-triple-r600)
http://images.vnu.net/gb/inquirer/news/2007/10/07/sapphire-shows-triple-r600/sapphiretricard.jpg
http://images.vnu.net/gb/inquirer/news/2007/10/07/sapphire-shows-triple-r600/sapphiresixconnector.jpg
Considering how sucky the 2-card variant is ATM, no thanks:D
Silent_Buddha
08-Oct-2007, 23:35
Er, holy crap this quote from the article might make me reconsider Crossfire.
ATI showed us beta drivers that enabled Crossfire running on three monitors.
IF ATI manage to enable multi-monitor while in Crossfire mode that's one step to drawing me back into possibly trying out multi-GPU setups again.
Now, if they also come up with a more elegant multi-GPU rendering solution than AFR (/barf) then I'll jump back on the bandwagon to at least try it out.
Regards,
SB
Considering how sucky the 2-card variant is ATM, no thanks:D
But you can replace your central heating unit with that ...
But you can replace your central heating unit with that ...
You'd think that, but asides from obnoxious noise under load, it doesn't produce anything in excessive quantities(heat included) and thus no:D. Perhaps when I'll build the water-loop for em, I can strategically place the rad....cheers for the suggestion:P:)
Mucho monitor: a no-SLi zone! :D
ShaidarHaran
09-Oct-2007, 01:16
Er, holy crap this quote from the article might make me reconsider Crossfire.
IF ATI manage to enable multi-monitor while in Crossfire mode that's one step to drawing me back into possibly trying out multi-GPU setups again.
Now, if they also come up with a more elegant multi-GPU rendering solution than AFR (/barf) then I'll jump back on the bandwagon to at least try it out.
Regards,
SB
Same here. Multi-GPU multi-mon is the bee's knees, especially for Flight Sim :D
LunchBox
22-Oct-2007, 04:35
so on the trifire, the two does the rendering and the other does physics?
like this set up??
http://ati.amd.com/technology/crossfire/physics/index.html
ShaidarHaran
22-Oct-2007, 04:50
so on the trifire, the two does the rendering and the other does physics?
like this set up??
http://ati.amd.com/technology/crossfire/physics/index.html
No. GPU Physics is little more than a concept with a few tech demos as proof-of-concept used to back the ridiculous marketing B.S. that these GPUs "support physics acceleration". Too bad there's no API nor game that support GPU physics, so it's rather pointless ATM.
Dang. And i've got a perfectly good x1900 just sitting here doing nothing.
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.