Silent_Buddha
Legend
What will you do when Dx11 games start being benched? Throw those out also, since you'll get the exact same perf increases with Dx11 vs Dx10 as you do with Dx10.1 vs Dx10.
Regards,
SB
AMD: R8xx Speculation
- Thread starter Shtal
- Start date
What will you do when Dx11 games start being benched? Throw those out also, since you'll get the exact same perf increases with Dx11 vs Dx10 as you do with Dx10.1 vs Dx10.
Regards,
SB
Yep, i knew you talked about the GPU end.
I should have been more clear about what i meant.
If i understood correctly, your interpretation about the AMD slide is that:
AMD has difficulties engineering the GPU end and scale the GDDR5 speed beyond 5Gbps (it has technical difficalties in general...)
and my interpretation is for one parameter only: temperature.
I mean:
AMD has difficulties engineering the GPU end and scale the GDDR5 speed beyond 5Gbps regarding only the temperature that AMD wants to maintain for the GPU core.
The old designs had 90°C and maybe AMD decided that it must do something to lower the temperature level.
It is very important for the future progression to design with such policies.
Like i said, i may misunderstood the AMD slide due to my lack of technical background.
I agree about the possibility that the 58XX is bandwidth limited (i posted about it one week before the AMD event and i was dead on that 5870 will use 5Gbps modules.
Like i said back then for me the main reasons that this is happening is because price level and volume (quantity) of the GDDR5 ICs that AMD is targeting for the 5870 SKU.
Qimonda was behind Samsung and Hynix regarding process technology (for example when Samsung was at 60nm, Hynix was at 66nm and Qimonda was at 70nm).
The thing is that Qimonda made a very good business decision to skip GDDR4 and they could compete much more effectively with Samsung and Hynix for GDDR5.
6Gbps ICs have also Samsung and Hynix.
I am not sure but i think the Qimonda ICs had better power consumption characteristics than the Samsung/Hynix ICs (same or nearly same process tech)
For me 5870 needs around 1,5GHz to show its true colors. I don't mean that with 1,5GHz will reach the level of 4870 performance/bandwidth efficiency, but that after 1,5GHz the performance improvements will be meaningless from a performance / memory scaling ratio perspective.(for the vast majority of the games at 4X AA)
Afterall 4870 had +20% clock speed and +80% memory bandwidth than 4850 and only +33% perf. (weighted average at 1920X1200 4XX AA 16X AF)
and the memory controller in the 5870 will be a new one, possible better than 4870's...)
Sorry to interfere but i think i understand why Rangers did that.
If you see his conclusion:
"If anything 5870 typically increased it's lead the higher the resolution/AA/AF settings were turned, so I dont see a bandwidth limiting problem, at least versus the current competition"
So his main goal was to see if 5870 is bandwidth limited or not.
So he had to use only those games that the performance difference was not due to the potential DX10.X codepath implementation.
But the logic, that 5870 doesn't have a bandwidth limiting problem because typically increased it's lead the higher the resolution/AA/AF settings were turned, is wrong imo.
For example take a 16ROPs GTS250 (738MHz core /1100MHz mem) underclock the memory at 900MHz, then take a low power edition of 16ROPs 55nm 9600GT (600MHz core / 900MHz mem)
You will see that the underclocked GTS250 is increasing it's lead the higher the resolution/AA/AF settings were turned, does this mean that the GTS250 is not bandwidth limited with 900MHz memory?
Last edited by a moderator:
Was it the results from the leaked alleged benchmark slides?
FWIW: What would you think, Nvidia'd come up with, if they'd have to chose some benchmarks to show off GTX285 against 5870? Wouldn't they chose the tests and settings they know would hurt their competition the most?
Silent_Buddha
Legend
Yes, but that's still no reason to exclude the Stormrise numbers due to Dx10.1 because if anything the performance increase indicated in that will become even more prevalent with Dx11 now that more cards will supports those features of Dx10.1, including Nvidia Dx11 cards, whenever they get released.
Regards,
SB
There is no reason to exclude DX10.1 games,
if you want to find, an indicative weighted average difference, between 5870 and GTX285, in a general sense.
Like you said, it is possible, with the upcoming titles or with the 2010 titles, the performance increase that Stormrise numbers indicate to become even more prevalent.
The thing is that the main goal that Rangers (according to my understanding) wanted, was not to find an indicative weighted average difference between 5870 and GTX285 in general.
According to my understanding, he was trying to figure out,
if the 5870 is bandwidth limited with 1,2GHz GDDR5, or not.
To do this and have valid results, you have to exclude the differencies in perf. that 10.1 codepaths brings.
If Rangers was trying to figure out an indicative weighted average difference, between 5870 and GTX285, then i agree with you.
I just had the feeling that his main goal was to see if the 5870 is bandwidth limited with 1,2GHz GDDR5, or not.
Silent_Buddha
Legend
Here's the thing though, if Dx11 somehow alleviates some of the need for more bandwidth through using performance enhancing features (Dx10.1 also) then it's relevant when discussing whether a card is or will be bandwidth limited.
Additionally comparing two entirely different architectures to come to a conclusion about whether one is bandwidth limited compared to the other is...well, there's too many variables involved here to come to any reliable conclusion.
Regards,
SB
Additionally comparing two entirely different architectures to come to a conclusion about whether one is bandwidth limited compared to the other is...well, there's too many variables involved here to come to any reliable conclusion.
Regards,
SB
Anyone else hear anything about 3600 shaders? :|
I think that belongs to the Area-51 forum
SiliconAbyss
Newcomer
3600 shaders? Where did you pick that up from? [edit button] Arty cleared that up, from a UFO encounter.
You'll see I listed several reasons beside DX10.1 for throwing it out :smile:
I just didn't feel like a title that was running over 2X as fast on the ATI hardware because of DX 10.1 represented anything meaningful, when trying to compare the cards power on an even playing field. Especially given all the other caveats already listed, low sales, mediocre reviews, and not a game you see in many benchmark suites Plus the fact IIRC the next highest ATI win was like 72%, while Stormrise was 125%, making it an extreme outlier. I would have thrown out any such game that showed a huge bias to Nvidia as well. Anyways I posted both numbers anyway, so if you prefer, just use the 54% one
And yeah the purpose of the comparison was to detemine the general strength of the cards..not see if it was BW limited, that was an ancillary conclusion (and not much of one since I know 1/1000th as much about this stuff as you guys :smile
CJ only compared the the 5870 to the 295 and the 5850 to the 285, so I just had to cobble together the numbers whenever same settings/games across the 5870/285 were available.
I'll be interested if "professional" reviews show the 5870 in a better performance light with higher numbers, because 44% isn't THAT impressive to me.
Silent_Buddha
Legend
Right-o, and I agree that game was utter poo. I feel sorry for my friend that bought it. It could have been an interesting RTS, but was saddled with a REALLY horrible console control interface.
Regards,
SB
Regards,
SB
hope this hasnt been posted
"Radeon 5750 and 5770 are scheduled for launch around the official launch date of Windows 7 [give or take around October 22nd-23rd, depending on what continent you live on]. The prices are set at just 149 US Dollars for 5750, with 5870 being priced at 199.
And here comes the kicker - when AMD introduces ATI Radeon HD 5750 and Radeon HD 5770, all of Radeon HD 4800 series are getting "EOL" [End Of Life] marking, since they will be pitched at same price points [149 bracket has HD4870 and HD5750, 199 bracket has HD4890 and HD5770]. Bear in mind that some e-tailers in the USA might offer a MIR [Mail-In-Rebate], ranging anywhere between 5-15 USD."
"Radeon 5750 and 5770 are scheduled for launch around the official launch date of Windows 7 [give or take around October 22nd-23rd, depending on what continent you live on]. The prices are set at just 149 US Dollars for 5750, with 5870 being priced at 199.
And here comes the kicker - when AMD introduces ATI Radeon HD 5750 and Radeon HD 5770, all of Radeon HD 4800 series are getting "EOL" [End Of Life] marking, since they will be pitched at same price points [149 bracket has HD4870 and HD5750, 199 bracket has HD4890 and HD5770]. Bear in mind that some e-tailers in the USA might offer a MIR [Mail-In-Rebate], ranging anywhere between 5-15 USD."
Didn't yall decide they were gonna be BW limited at 128-bit?
If not those are amazing cards for AMD at so small..
Here:
http://www.brightsideofnews.com/new...ing-revealed2c-end-for-hd4800-draws-near.aspx
Salt truck backing to unload: meeep, meeep, meeep
Even then selling a ~450 mm2 gpu at $250 is gonna hurt. It will have to sell for less than 299. More importantly, what happens to all the rebranded g92's? They don't even have dx10.1. :smile:
Similar threads
