View Full Version : OMAP 2420 Dev Board and Renesas SHMobile Board
Kristof
10-Apr-2005, 11:57
Some Dev Board pictures from CTIA are now online :
http://www.paraknowya.info/images/pvr/CTIA/TI_Booth_05.JPG
http://www.paraknowya.info/images/pvr/CTIA/Renesas_Booth_02.JPG
More pictures and info available :
http://www.paraknowya.de/
K-
A good update of new shots from CTIA and, finally, the PowerVR Racer movie. Thanks for the coverage from the booths!
PC-Engine
10-Apr-2005, 15:04
Neat! :D
Those screens look so small in comparison to the circuit board...
And the 2nd one has two screens! Look, for once companies actually did copy nintendo's 'revolutionary' ideas!
Driving two screens is pretty useful for development/presentation purposes. DS could've been its own dev station!
That racing demo looks a bit crap IMO. PowerVR should get a software developer to do a demo for them.
And the 2nd one has two screens! Look, for once companies actually did copy nintendo's 'revolutionary' ideas!
For once? :?
Well, Nintendo always claims they won't unveil something because other companies would steal their ideas, but it seems to very rarely happen. Or perhaps I missed all the GameCube and Super Mario Sunshine clones?
If Mario Sunshine was Nintendo's only game and GameCube they're only peice of hardware you might just have a point. If Nintendo had ever actually claimed that they wouldn't show GameCube because they feared a company would copy it then you might have a point. None of that is the case though.
BTW here's a question for you. Did the Playstation pad start off with analog sticks or did Sony get the idea from somewhere else?
PowerVR Racer highlights MBX's scalability. According to the booth signs, it was designed able to run with ample headroom for other tasks and on a fixed-point platform with a ~50-MHz MBX-Lite and no companion VGP.
The definition and color precision from the demo will be great qualities to see on a portable.
Oh right I see, thanks for the info
archie4oz
11-Apr-2005, 04:47
Did the Playstation pad start off with analog sticks or did Sony get the idea from somewhere else?
Well it certainly wasn't Nintendo... :P
The definition and color precision from the demo will be great qualities to see on a portable.
In comparison to what? A current cell phone? PDA?, DS? PSP?
Acert93
11-Apr-2005, 05:05
Did the Playstation pad start off with analog sticks or did Sony get the idea from somewhere else?
Well it certainly wasn't Nintendo... :P
Yep, total coincidence that in the middle of the 1st 3D gen, the same gen that Nintendo introduced analog sticks to the console mass market, that Sony introduced a similar input device. From what I hear Sony had always planned to put analog sticks on the PS, but only remembered their original plan after the N64 was launched :wink:
The Saturn also added an analog control pad in 1996 with the superb NiGHTS 3D Controller. Had analog triggers, too.
archie4oz:
In comparison to what? A current cell phone? PDA?, DS? PSP?
Compared to anything from the sector. Supersampled anti-aliasing and full precision blended color are great even by home console standards.
darkblu
11-Apr-2005, 05:52
impressive demo. could have been a bit more artistically polished, though. like using aniso for the road textures, or at least not using mipmapping : ) seriously, when do i get one of those in a consumer unit with something less proprietary than psp-os, or at least !wince?
Hopefully this year. Img Tec's latest investor update reveals that there are fourteen SoCs licensing the MBX family for the PDA/phone sector between Renesas, Intel, Texas Instruments, Samsung, Philips, and Freescale. Recently, Texas Instruments extended their license to MBX Lite in order to target power/price-sensitive phones, and Intel's preparing by extending their license to cover more of the MBX family's IPs.
Simon F
11-Apr-2005, 10:04
impressive demo. could have been a bit more artistically polished, though. like using aniso for the road textures, or at least not using mipmapping : )
I do see the smiley, but seriously? No MIP mapping? Apart from making it look hideous, the performance would drop like a stone. I have a "development tool" hanging on the office wall that I'd like to apply to 3D coders who don't use MIP mapping...
seriously, when do i get one of those in a consumer unit with something less proprietary than psp-os, or at least !wince?
What about the The Dell Axim?
Maybe that said "not wince".
Simon F
11-Apr-2005, 12:35
Maybe that said "not wince".
Ah, my parser wasn't in "C" mode. Presumably that only leaves Linux....because @ the OpenVG meetings I've attended people often express some frank opinions which I shall not repeat
Anyway, pardon my ignorance, but what's so bad about WinCE? AFAIAA it's running OGL-ES which seems OK.
impressive demo. could have been a bit more artistically polished, though. like using aniso for the road textures, or at least not using mipmapping : )
I do see the smiley, but seriously? No MIP mapping? Apart from making it look hideous, the performance would drop like a stone. I have a "development tool" hanging on the office wall that I'd like to apply to 3D coders who don't use MIP mapping...
The mip map boundaries are uncomfortably visible. I take it then that Kyro's "fast trilinear" either doesn't work with PVRTC or it takes up too much silicon real estate. I'm a little disappointed TBH. The lack of trilinear filtering on many DC games (e.g. MSR, PSO) was IMHO its most annoying deficiency. It seems MBX will be the same.
darkblu
11-Apr-2005, 15:56
Hopefully this year. Img Tec's latest investor update reveals that there are fourteen SoCs licensing the MBX family for the PDA/phone sector between Renesas, Intel, Texas Instruments, Samsung, Philips, and Freescale. Recently, Texas Instruments extended their license to MBX Lite in order to target power/price-sensitive phones, and Intel's preparing by extending their license to cover more of the MBX family's IPs.
thanks, lazy. let's hope we see something along these lines.
I do see the smiley, but seriously? No MIP mapping? Apart from making it look hideous, the performance would drop like a stone. I have a "development tool" hanging on the office wall that I'd like to apply to 3D coders who don't use MIP mapping...
ok, seriously. aside from the performance consideration, mipmapping (w/o aniso) on roads is a big no-no in my book. actually it's a big no-no on any surfaces meant to be viewed under steep angles most of the time. mips weren't designed for such conditions, and do happen to look atrocious under such for the most of the time. so if i can take the performance hit, i'd go with carefully selected textures resolutions and no mipmaps in those cases.
Anyway, pardon my ignorance, but what's so bad about WinCE? AFAIAA it's running OGL-ES which seems OK.
oh, i'm ok with ogl/se, i'm not ok with the rest of the environment. it gives me the creeps. somehow reminds me of 'mini me' from austin powers, just not funny.
The Saturn also added an analog control pad in 1996 with the superb NiGHTS 3D Controller. Had analog triggers, too.
Heck, I think my Vectrex has an analog stick in its controller but I'll have to fire it up to be 100% sure.
If Mario Sunshine was Nintendo's only game and GameCube they're only peice of hardware you might just have a point. If Nintendo had ever actually claimed that they wouldn't show GameCube because they feared a company would copy it then you might have a point. None of that is the case though.
BTW here's a question for you. Did the Playstation pad start off with analog sticks or did Sony get the idea from somewhere else?
My comment was in jest over something nintendo normally gets a lot of flak over.
There's quite a bit that has been copied from nintendo and times when they would have been legitimately worried about someone stealing ideas....but it hasn't really happened recently, certainly I don't recall any innovative idea in super mario sunshine that would have been copied.(super mario 64 on the other hand, that was a game to keep locked up until near release date)
Driving two screens is pretty useful for development/presentation purposes. DS could've been its own dev station!
Not to mention every cellphone has 2 screens... don't read too much into it.
Simon F
12-Apr-2005, 14:29
The mip map boundaries are uncomfortably visible. I take it then that Kyro's "fast trilinear" either doesn't work with PVRTC or it takes up too much silicon real estate.
Remember Kristof's old adage. "Nothing is free in 3D". Anything you do will cost some silicon area...
I'm a little disappointed TBH. The lack of trilinear filtering on many DC games (e.g. MSR, PSO) was IMHO its most annoying deficiency. It seems MBX will be the same.
To put this in to perspective, you are complaining that MBX/MBX-lite does not support FREE trilinear.
That demo is available for the Dell Axim, which is an MBX-lite. Given how small it is, I think you're being overly picky. I suppose you can always turn on trilinear if you really want to.
I suppose I could console myself with the fact that a 100MHz+ MBX powering a screen at mobile resolutions probably has pixel fill rate to burn on such things as trilinear filtering, maybe with a spot of anisotropic too.
darkblu
12-Apr-2005, 20:40
That demo is available for the Dell Axim, which is an MBX-lite. Given how small it is, I think you're being overly picky. I suppose you can always turn on trilinear if you really want to.
speaking of axim and the demo, do you guys offer ogl/es for linux too? IOW, can i get an axim, put a sane os on it and still have ogl/es?
Driving two screens is pretty useful for development/presentation purposes. DS could've been its own dev station!
Not to mention every cellphone has 2 screens... don't read too much into it.
Wah? Most cell phones I know of don't have 2 screens, not even flip phones. Could you name some phones with 2 screens?
Wah? Most cell phones I know of don't have 2 screens, not even flip phones. Could you name some phones with 2 screens?
Perhaps he means one on the inside (the larger one) and one on the outside (the smaller one) - this is for clamshell designs naturally, NOT candy bar ones.
I suppose I could console myself with the fact that a 100MHz+ MBX powering a screen at mobile resolutions probably has pixel fill rate to burn on such things as trilinear filtering, maybe with a spot of anisotropic too.
at cell phone screen sizes is trilinear very important ? I would think fast bilinear with fsaa would be better than just trilinear . The screens are so small .
The actual size of the mbx is also very small and barely increases the cost of the cpus which its embeded in . This will bring dreamcast lvl graphics to cell phones very cheaply which is hte most important thing . NOt trilinear filtering .
Wah? Most cell phones I know of don't have 2 screens, not even flip phones. Could you name some phones with 2 screens?
Perhaps he means one on the inside (the larger one) and one on the outside (the smaller one) - this is for clamshell designs naturally, NOT candy bar ones.
Bleh, the one on the outside is usually just a clock or something.(my cell phone is a flip phone and doesn't have it btw, but mine is super old)
Bleh, the one on the outside is usually just a clock or something.(my cell phone is a flip phone and doesn't have it btw, but mine is super old)
My cell (kinda old too) has one on the outside and it shows a picture of the person calling (I can assign pictures to it). The PQ is no different than the inner screen other than being smaller.
Kristof
13-Apr-2005, 08:18
Wah? Most cell phones I know of don't have 2 screens, not even flip phones. Could you name some phones with 2 screens?
Perhaps he means one on the inside (the larger one) and one on the outside (the smaller one) - this is for clamshell designs naturally, NOT candy bar ones.
Correct, the SHMobile Dev Board is actually a dev board for a clamshell type phone with one highres display and one lower res display (outside one).
Kristof
14-Apr-2005, 17:11
Samsung DevBoard picture can be seen in the latest Khronos Presentations:
http://www.khronos.org/devu/library/seoul_04_2005/Samsung_-_OpenGL_ES_Hardware_k.ppt
Slide 7
OpenGL ES AP – S3C2460
Silicon Sampling & Board Ready!
3D Graphics Engine powered by MBX R-S Core
OpenGL ES 1.0/1.1 compliant
K-
Ailuros
18-Apr-2005, 13:16
at cell phone screen sizes is trilinear very important ? I would think fast bilinear with fsaa would be better than just trilinear . The screens are so small .
Bilinear + Supersampling gives damn close 2xAF filtering quality; albeit you'd need 4x sample SSAA to match 2xAF in both directions. However 2x vertical Supersampling + bilinear are IMHO good enough for such small devices and a better investment in terms of fill-rate overall, compared to "just" trilinear or even worse "strict" AF.
If die-size in the future allows for fast adaptive anisotropic algorithms, a combination with pure Multisampling would save of course a lot more fill-rate, especially when resolutions tend to scale even more in the future.
As a reminder in terms of texture filtering even on today's high end PC GPUs, I can see only bilinear being free. Else replace with newly found terms like bri- or try-linear. Is there a PDA/mobile core out there that is capable of real single-cycle trilinear?
Iron Tiger
19-Apr-2005, 02:04
Anyway, pardon my ignorance, but what's so bad about WinCE? AFAIAA it's running OGL-ES which seems OK.The same thing that's wrong with WinXP. Buggy and bloated. BTW, I have an Axim X50v with Windows Mobile 2003 Second Edition, so I speak from experience.
That demo is available for the Dell AximPublicly? :shock: If so, got a link?
http://www.powervr.com/Downloads/Demos/index.asp
Iron Tiger
19-Apr-2005, 04:52
http://www.powervr.com/Downloads/Demos/index.aspGracias! :D
Samsung DevBoard picture can be seen in the latest Khronos Presentations:
OpenGL ES 1.0/1.1 compliant
I don't think this is really true. Haven't seen any announcement so far that the MBX would be compliant with OpenGL ES 1.0/1.1.
PC-Engine
19-Apr-2005, 07:58
Samsung DevBoard picture can be seen in the latest Khronos Presentations:
OpenGL ES 1.0/1.1 compliant
I don't think this is really true. Haven't seen any announcement so far that the MBX would be compliant with OpenGL ES 1.0/1.1.
Kristof works for Imagination Technologies...
Ailuros
19-Apr-2005, 09:09
Samsung DevBoard picture can be seen in the latest Khronos Presentations:
OpenGL ES 1.0/1.1 compliant
I don't think this is really true. Haven't seen any announcement so far that the MBX would be compliant with OpenGL ES 1.0/1.1.
That's a public document quoted up there; in any case what would you suggest the requirements are for 1.1 exactly?
Not sure about 1.1, but in 1.0 they needed proprietary extensions to support the VGP functionalities.
EDIT: removed, because off topic material.
...and everyone in the biz long enough surely must remember Commodore's legendary "paddles" ? :) those were definately analog controllers. :)
Oh absolutely PCs had analog controls for zillions of years. I was just trying to stick to consoles or video game machines.
The graphical quality in that PowerVR Racer vid is far beyond anything I've seen from the PSP. And to think, it's the absolute weakest version of the chip.
I take my hat off to Imagination for producing a chip of such wonder in 2002 thats beating the hell out of competing products in the year 2005.
Absolutely amazing.
I can't wait to see what these guys will deliver for SEGA SAMMY's next Arcade board. It shall be truly monstrous.
You guys ROCK!
Not really bothered about how many obscure systems had some kind of analog controls. The fact is Sony changed their controllers to add the analog stick because of Nintendo putting an analog stick on the N64 pad.
Not really bothered about how many obscure systems had some kind of analog controls.
Vectrex wasn't obscure. The Colecovision might have even had an analog joystick as well (but I'm not 100% sure).
The fact is Sony changed their controllers to add the analog stick because of Nintendo putting an analog stick on the N64 pad.
Uh, ok. Thanks for that bit. Don't get me wrong, I'm quite willing to believe that Sony looked at their nearest competition and copied it (as I'm sure Nintendo has done so). The fact remains, when it comes to analog joysticks in video games/consoles. It's been done before.
Don't get me wrong, I'm quite willing to believe that Sony looked at their nearest competition and copied it
Well that's all I said so...
Well that's all I said so...
Then why did you ask earlier? Honest question. Were you simply trying to lead us into a discussion?
Like I said, it's quite possible heck, in fact I'd say Sony looked at Nintendo. They probably looked elsewhere as well. But that does NOT mean BECAUSE of Nintendo Sony did so as well - unless of course you have a copy of the memo sent down to engineering. ;)
Do you think Nintendo looks elsewhere to copy/refine ideas from?
It was a rhetorical question. I was just giving Fox5 a pretty obvious example (and it is obvious AFAICS) of Nintendo doing something and its competitor following suit. Since he was suggesting that never happens.
BTW yeah I'm sure Nintendo get loads of their ideas from elsewhere. I'm sure their direct competitors even influence their designs at times. To be clear my comment was never supposed to be about Nintendo vs Sony. Sony just happened to be part of that particular example.
BTW what's a Vectrex? I've seriously never heard of it :)
You're probably too young to remember (or it never made it across the pond).
http://www.roachnest.com/vectrex.html
It was a "portable" console that accepted carts (it's portable because the screen was built-in like the first Apple Macintoshs). It was a vector based rasterizer (like Asteroids, the old Battlezone, Red Baron, etc). Because these tend to be monochromatic, you'd place these plastic overlays over the screen to give areas of the screen (static/non-moving) color.
It had 2 joystick ports and each joystick (more like a wide joy pad actually) was analog and there were 4 buttons.
It's funny, some guys have even captured the data from old carts and reburned them onto newer carts - but these newer carts hold like 75 games. Too bad they lack the overlays though, that is what makes the game. :)
PC-Engine
21-Apr-2005, 12:57
Uh, how do you know for a fact the directional controller for the Vectrex was actually analog? Did any of the Vectrex games support analogy control? If yes which one?
archie4oz
21-Apr-2005, 17:46
The fact is Sony changed their controllers to add the analog stick because of Nintendo putting an analog stick on the N64 pad.
I find this absurdly amusing... :) If the Dual Shock had been release because of the N64, why did Sony wait 2 years to do it, rather than release it with the launch of Gran Turismo?
Uh, how do you know for a fact the directional controller for the Vectrex was actually analog? Did any of the Vectrex games support analogy control? If yes which one?
http://www.playvectrex.com/shoptalk/vecjoyrepair/vecjoyrepair_f.htm
Talks about poteniometers in the joystick. Those are used for measuring small changes.
http://www.classicgaming.com/vectrex/po_digital-analog.htm
Most Vectrex games do not take advantage of the analog capability of the Vectrex Controller.
A. Most Vectrex games are programmed to give an all-or-nothing joystick response during gameplay...This means that you have to push the analog joystick knob so far to get a response...a digital response to your action...
3. The games which do use the variable capability of the original analog controller are few in number.
A. Analog means that you push the stick a little bit, the Vectrex console responds a little bit. You push alot and the console responds alot. Currently there are only three games that use this capability: Version 2 of Star Hawk, Hyperchase and Moon Lander. The former two are dogs, the third is brilliant!
And even if there wasn't a game that used this ability, the joystick would still be analog. I.e. it doesn't really matter or not if any game used analog controls.
The fact is Sony changed their controllers to add the analog stick because of Nintendo putting an analog stick on the N64 pad.
I find this absurdly amusing... :) If the Dual Shock had been release because of the N64, why did Sony wait 2 years to do it, rather than release it with the launch of Gran Turismo?
Maybe it had to prove itself as a viable control method first, or maybe Sony wasn't sure how to integrate it into their controllers.
or maybe Sony wasn't sure how to integrate it into their controllers.
Well it's not really that hard to reverse engineer a joypad though. You simply unscrew it. ;)
or maybe Sony wasn't sure how to integrate it into their controllers.
Well it's not really that hard to reverse engineer a joypad though. You simply unscrew it. ;)
But the dual shock bares little in common with a n64 controller other than that it has analog controls and rumble. I meant maybe sony wasn't sure if they should do 1 stick or two, placement of joysticks, how the joysticks should be made, would they integrate well into all games, etc. Maybe they even considered completely redesigning the control pad before settling on the dual shock.
I find this absurdly amusing... If the Dual Shock had been release because of the N64, why did Sony wait 2 years to do it, rather than release it with the launch of Gran Turismo?
Don't know why it took Sony quite so long to release Dual Shock in the US. Maybe they held its release back a bit so they could release it with one of their own top analog based games (Gran Turismo). In Japan they released the Dual Shock just over a year after the N64.
Look, Nintendo brought out a pad with a stick and rumble function. One year later Sony change their pad to add sticks and rumble function. Yet you think that had nothing to do with the N64 pad? Now that's absurd.
archie4oz
23-Apr-2005, 05:18
Don't know why it took Sony quite so long to release Dual Shock in the US. Maybe they held its release back a bit so they could release it with one of their own top analog based games (Gran Turismo). In Japan they released the Dual Shock just over a year after the N64.
The "Dual Shock" was released 2 years after the N64 launch (both in Japan and the US). You're confusing the Dual Shock, with the Dual Analog Pad (similar to Dual Shock but larger with different analog contollers).
The Dual Analog Pad was unveiled around 4 months after the Japanese N64 launch, and we actually had them months before the N64 Japanese launch because we were developing support for them into Bushido Blade and Tobal #2...
The Dual Analog Pad was also meant as a smaller, more cost-effective substitute for Sony's Dual Analog Flightsticks (made for games like Mechwarrior 2, Ace Combat 2, etc..). Also the system was designed from the get-go to support rumble feedback as the controller ports were designed to provide the necessary power for it (whereas the N64's was a battery powered add-in (after the DAP was released BTW)), although strangely the US DAP didn't have rumble...
Simon F
25-Apr-2005, 07:38
How the hell did a thread about the Omap 3D development system get hijacked for damned game controller discussion?
How the hell did a thread about the Omap 3D development system get hijacked for damned game controller discussion?
if you look at the topics in handheld 3d, it starting to look like this is in fact handheld games industry forum, which again means whole bunch of guys from consoles forum... and the result can be seen here.
not that I would exactly be angry after so many of my handheld 3d based threads turn out to be just plain bashing few years ago and never receiving any apologies after being proven to be right later on. But still it's sad to see this still happening.
I could have done same thing that kristof used to do when something about Bitboys was shown (trying to hijack the thread with competitors' material.) but I decided not to go that low.
I'll remove my remarks on old stuff right away.
Ailuros
27-Apr-2005, 12:10
I don't see any particular reason to remove anything from your past feedback Nappe; most of us are somewhat guilty of getting a thread off topic.
In the majority of cases, apart from a bunch of lucky guesses and some weird research methodoly it's damn hard to get answers to most of my PDA/mobile questions and that irrelevant if it's you specifically or anyone else. Perfectly understandable, since way too much valuable information is classified.
If I'd ask what this one is about, do you think I'll get an answer?
http://users.otenet.gr/~ailuros/samsung.jpg
*runs away to avoid lynching* :oops:
Ailuros
27-Apr-2005, 23:02
That's what I'd like to know myself; it's in Samsung's future roadmap (check the recent pdf at Khronos' website).
Ailuros: I decided to remove that because it really wasn't topic related and it really didn't give any additional information as off-topic either.
anyways, interesting picture. If I have any clue left from the days I was really interested 3D HW stuff, it looks like capable doing some SM 3 related stuff like vertex texturing... There's TMU / texture engine for vertex unit as well as the "per frag. unit" looks like special unit that has access to main Bus. So, it could be used for some kind of specialized AA perhaps...
Ailuros
28-Apr-2005, 10:35
It's the primitive engine that caught my eye.
Could it be MBX 2?
http://www.beyond3d.com/forum/viewtopic.php?t=22093
It's the primitive engine that caught my eye.
well, it's after vertex shader, so most likely it is just primitive setup.
(High Order Surfaces and Basic Geometry Primitives would need to be tesselated before vertex shader, right? )
So, most likely it is Primitive Processor as same way as Pyramid 3D had one: it just constructs triangles for rasterizer.
Could it be MBX 2?
http://www.beyond3d.com/forum/viewtopic.php?t=22093
umh... if it is, I am a bit disapointed. besides, I doubt Samsung would have right to distribute charts about licensed unreleased technology.
Ailuros
28-Apr-2005, 13:12
It's the primitive engine that caught my eye.
well, it's after vertex shader, so most likely it is just primitive setup.
(High Order Surfaces and Basic Geometry Primitives would need to be tesselated before vertex shader, right? )
So, most likely it is Primitive Processor as same way as Pyramid 3D had one: it just constructs triangles for rasterizer.
Most likely yes; if it should be programmable though (which I consider unlikely at this stage) we'd be talking about more rapid advancements in the PDA/mobile than I could have imagined.
I doubt Samsung would have right to distribute charts about licensed unreleased technology.
In a best case scenario they could have some sort of commitment at this stage, which would make the chances of releasing any specs this early even smaller. AFAIK didn't license MBX Lite directly from IMG either.
The charts aren't that much telling either if you think about it. One of them states PowerVR MBX and the "FIMG" in question (for the future) just states "programmable" and not who'd design that thing. Has Samsung a graphics department or not?
umh... if it is, I am a bit disapointed.
As I said the best you can get out of those graphs is an architecture with programmable PS/VS. What exactly would you expect for the second generation of PDA/mobile chips anyway?
2004 (MBX Lite) = 1M Tris/sec
2005 (programmable) = 5M Tris/sec
....with the estimated triangle rate scaling up to 40M Tris beyond 2007 (for that mysterious "FIMG" thing).
Sounds more of a prediction than anything and albeit 40M Tris/sec sound way too optimistic IMO in such short notice, I don't see anything underwhelming, rather the contrary. If someone's calculating 1 vertice=1 triangle then of course it's a totally different story.
It's the primitive engine that caught my eye.
well, it's after vertex shader, so most likely it is just primitive setup.
(High Order Surfaces and Basic Geometry Primitives would need to be tesselated before vertex shader, right? )
So, most likely it is Primitive Processor as same way as Pyramid 3D had one: it just constructs triangles for rasterizer.
Most likely yes; if it should be programmable though (which I consider unlikely at this stage) we'd be talking about more rapid advancements in the PDA/mobile than I could have imagined.
If the image has correct connections, the primitive engine does not have access to main bus. It could indicate primitive part not being programmable or it being quite limited.
I doubt Samsung would have right to distribute charts about licensed unreleased technology.
In a best case scenario they could have some sort of commitment at this stage, which would make the chances of releasing any specs this early even smaller. AFAIK didn't license MBX Lite directly from IMG either.
The charts aren't that much telling either if you think about it. One of them states PowerVR MBX and the "FIMG" in question (for the future) just states "programmable" and not who'd design that thing. Has Samsung a graphics department or not?
umh... if it is, I am a bit disapointed.
As I said the best you can get out of those graphs is an architecture with programmable PS/VS. What exactly would you expect for the second generation of PDA/mobile chips anyway?
well, I thought that mobile stuff could be going faster towards unified architechtures and more raw floating point speed. This is because in mobile you actually have more stuff (that is not visible as floating point operations to user) that needs floating point power. (and right now, flops speed seems to be a one of the biggest problems.)
Also, transistor budgets really would like to idea using new floating point power in more flexible way. (using extra power in compressing video stream could be one example.)
but of course, this is just how I see it and as it is just a opinion, so I might be wrong as well as right.
2004 (MBX Lite) = 1M Tris/sec
2005 (programmable) = 5M Tris/sec
....with the estimated triangle rate scaling up to 40M Tris beyond 2007 (for that mysterious "FIMG" thing).
Sounds more of a prediction than anything and albeit 40M Tris/sec sound way too optimistic IMO in such short notice, I don't see anything underwhelming, rather the contrary. If someone's calculating 1 vertice=1 triangle then of course it's a totally different story.
as said, I see mobile gpus doing lot's of things that they don't do in PC side. there's whole bunch of specialized needs in mobile and they greatly vary depending on target devices. That's why I really don't put much weight on Vertice / Triangle / pixel fill rates. Of course, those are one part but most of the real money is made with other kind of features.
Ailuros
29-Apr-2005, 05:43
If the image has correct connections, the primitive engine does not have access to main bus. It could indicate primitive part not being programmable or it being quite limited.
As I said I doubt it would be programmable, but those kind of "prediction" diagrams aren't very indicative about anything. I'm sure you've seen it already but here's a block diagram from MBX:
http://www.powervr.com/images/BlockDiagrams/MBX.gif
well, I thought that mobile stuff could be going faster towards unified architechtures and more raw floating point speed. This is because in mobile you actually have more stuff (that is not visible as floating point operations to user) that needs floating point power. (and right now, flops speed seems to be a one of the biggest problems.)
Depends on the requirements of OGL-ES2.0 I guess and if a unified architecture saves hardware space after all or not; it then would come down to design-decisions. If a vendor isn't bound by API requirements, it's rather irrelevant how it reaches a specific target.
I'm not at all saying or implying that you don't have a point; I'd rather tend to agree, but since transistor budget and die space is a major headache for such small devices opportunities and API definitions should be logically tied by that.
Also, transistor budgets really would like to idea using new floating point power in more flexible way. (using extra power in compressing video stream could be one example.)
but of course, this is just how I see it and as it is just a opinion, so I might be wrong as well as right.
W/o having seen from a layman's POV a unified architecture comparing to a non-unified architecture, I'd result to unsafe predictions.
as said, I see mobile gpus doing lot's of things that they don't do in PC side. there's whole bunch of specialized needs in mobile and they greatly vary depending on target devices. That's why I really don't put much weight on Vertice / Triangle / pixel fill rates. Of course, those are one part but most of the real money is made with other kind of features.
Uhmmm not entirely true. Today's and future GPUs are more and more moving into the GPGPU direction, for multimedia related functions as one example. Mobile GPUs are the ones and the according APIs are the ones that currently are trying to catch up with PC desktop GPUs as fast as possible and not the other way around. Albeit I'm only guessing here mobile GPUs might have an advantage here in SoCs and especially UMA.
Triangle rate or FLOPs are in my book just an alternative way of interpreting throughput. MBX Lite (since I doubt it would ever be clocked much higher than 80MHz) is low end for the first generation anyway. Clock a MBX at 200MHz and 3M (real) Tris/sec or alternatively 800 MFLOPs/sec and you'll end up with adequate FLOP power for the first generation. TI is according to rumours working already on a 90nm die shrink and if I take the predictions in the SAMSUNG graphs as reliable it would be sensible to expect second generation GPUs to start out at both 130 and 90nm. 5M Tris/sec is merely low end I guess; I wouldn't be surprised if 3x times the throughput for high end chips would be possible.
Low end chips are and will continue to be IMO targeting mostly mobile phones. I doubt SAMSUNG is targeting any other devices in the future either.
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.