Major NVidia cards comparing PPD/W and PPD/$

Moderators: slegrand, Site Moderators, PandeGroup

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby MtM » Sun Aug 10, 2008 5:32 pm

Xilikon wrote:Nice idea of having a PPD/W chart, this will help people get the best bang for the buck while staying very efficient in terms of power use.


Uncle_Fungus his site already has a huge clunk of the needed info, but even while it's there the owner seems to ignore me :e?:
MtM
 
Posts: 3054
Joined: Fri Jun 27, 2008 2:20 pm
Location: The Netherlands

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby legoman666 » Sun Aug 10, 2008 5:59 pm

MtM wrote:Legoman, do you think you could assemble a same chart for the Ati cards? I know you have an ati ppd spreadsheet but not sure if holds the wattage.

And the same goes for the smp client, I'm looking into some projects on my own but either I will have to rely on your collected data, look around myself ( which I don't think so, to time consuming ) or make a data miner tool for people to run so they don't have to go through allot of hassle to get the info we need.


Yea, that'd be no problem. I didn't do it before now because the performance is so poor and there's a lot less data. But I'll do it tomorrow at work...
Image
legoman666
 
Posts: 296
Joined: Sat Dec 22, 2007 6:26 pm

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby MtM » Sun Aug 10, 2008 6:23 pm

Well for those who run them, atleast the ati cards have been producing scientific results for some time already.

Legoman I would like to help you but you seem very efficient in collecting data and putting them in your spreadsheets! Now that I'm thinking about it.. ygpm ( in a min )
MtM
 
Posts: 3054
Joined: Fri Jun 27, 2008 2:20 pm
Location: The Netherlands

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby uncle_fungus » Sun Aug 10, 2008 11:15 pm

MtM wrote:
Xilikon wrote:Nice idea of having a PPD/W chart, this will help people get the best bang for the buck while staying very efficient in terms of power use.


Uncle_Fungus his site already has a huge clunk of the needed info, but even while it's there the owner seems to ignore me :e?:


No, not ignoring, just been away form the internet for a few days.
User avatar
uncle_fungus
Site Admin
 
Posts: 1682
Joined: Fri Nov 30, 2007 9:37 am
Location: Oxfordshire, UK

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby MtM » Mon Aug 11, 2008 10:27 am

uncle_fungus wrote:
MtM wrote:
Xilikon wrote:Nice idea of having a PPD/W chart, this will help people get the best bang for the buck while staying very efficient in terms of power use.


Uncle_Fungus his site already has a huge clunk of the needed info, but even while it's there the owner seems to ignore me :e?:


No, not ignoring, just been away form the internet for a few days.


Just my luck, the moment he im's me I'm not at home :lol:

Sorry if the previous post sounded negative, it wasn't ment to reflect such an feeling. I was more lost then upset :?:
MtM
 
Posts: 3054
Joined: Fri Jun 27, 2008 2:20 pm
Location: The Netherlands

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby legoman666 » Mon Aug 11, 2008 1:14 pm

Press refresh.

Image
legoman666
 
Posts: 296
Joined: Sat Dec 22, 2007 6:26 pm

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby YashBudini » Tue Aug 12, 2008 1:01 am

Dang I noticed which card has good PPD, but not which one leads in PPW.
No matter, 2 more cards to buy in the future.
Where ever you are, you're there.
YashBudini
 
Posts: 87
Joined: Wed Dec 12, 2007 1:44 am

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby shatteredsilicon » Tue Aug 12, 2008 1:23 pm

So, in conclusion, FAH doesn't scale past 96 pipelines (GS). No wonder the ATI cards don't perform up to their capabilities.
Image
1x Q6600 @ 3.2GHz, 4GB DDR3-1333
1x Phenom X4 9950 @ 2.6GHz, 4GB DDR2-1066
3x GeForce 9800GX2
1x GeForce 8800GT
CentOS 5 x86-64, WINE 1.x with CUDA wrappers
shatteredsilicon
 
Posts: 627
Joined: Tue Jul 08, 2008 2:27 pm

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby cavyman » Tue Aug 12, 2008 1:32 pm

What do you mean by "FAH doesn't scale past 96 pipelines"?
AMD Athlon 64 x2 5400+ BE - OC @ 3.0ghz | MSI K9A2 Platinum w/4GB RAM
Nvidia 9800 GTX+ | Nvidia 8800 GTS (G92) | Nvidia 8800 GT
Corsair 750 watt PSU

*** FOR SALE *** PM me for details.

Image
cavyman
 
Posts: 36
Joined: Wed Jun 25, 2008 9:05 pm
Location: TN, USA

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby shatteredsilicon » Tue Aug 12, 2008 2:20 pm

8800GS has 96 pipelines. Above that on the G80/G92 series, the performance per clock seems to be pretty flat. GTX only folds faster than the GS because it runs at a higher clock speed. There seems to be no performance gain from extra pipelines.
shatteredsilicon
 
Posts: 627
Joined: Tue Jul 08, 2008 2:27 pm

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby MtM » Tue Aug 12, 2008 3:21 pm

I haven't noticed that all actually. It's the shader clock you should be looking at you know that right? Got some data to back this claim up?
MtM
 
Posts: 3054
Joined: Fri Jun 27, 2008 2:20 pm
Location: The Netherlands

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby cavyman » Tue Aug 12, 2008 3:49 pm

MtM wrote:I haven't noticed that all actually. It's the shader clock you should be looking at you know that right?


Agreed. Plus, anyone who worked with the GPU2 beta should have known that the beta work units were well known, "small" protiens. Perhaps now that we're technically out of beta, we will start to see more and more SPs used with bigger work units. I just don't see it being a fair (or safe) assumption that the GPU2 core is somehow constrained to only 96 SPs--at least not based on the rather limited compilations of early data.
cavyman
 
Posts: 36
Joined: Wed Jun 25, 2008 9:05 pm
Location: TN, USA

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby legoman666 » Tue Aug 12, 2008 4:14 pm

comparison of PPD/shader coming right up...
legoman666
 
Posts: 296
Joined: Sat Dec 22, 2007 6:26 pm

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby legoman666 » Tue Aug 12, 2008 4:33 pm

Image

edit: read my later post. updated image.


Interestingly enough, most cards are all around the same number. There are only a few anomalous cards; mostly in the lower end. So from this chart, you can see that for the most part, performance scales linearly with the amount of shaders and their respective clock speeds.

Upon a closer look, it looks like of the 2 data points for the 8400GS, 1 of them has it's shader overclocked to 1400mhz. For the 8600GT, the average shader clock for the data points I have recorded is 1472mhz, up from the 1180mhz stock speed.

Next step is to change my spreadsheet so the core/mem/shader clock speeds are recorded in their own cell so I can calculate their averages. Weeee.
Last edited by legoman666 on Tue Aug 12, 2008 5:54 pm, edited 1 time in total.
legoman666
 
Posts: 296
Joined: Sat Dec 22, 2007 6:26 pm

Re: Major NVidia cards comparing PPD/W and PPD/$

Postby MtM » Tue Aug 12, 2008 4:58 pm

legoman666 wrote:Image

Interestingly enough, most cards are all around the same number. There are only a few anomalous cards; mostly in the lower end. So from this chart, you can see that for the most part, performance scales linearly with the amount of shaders and their respective clock speeds.

Upon a closer look, it looks like of the 2 data points for the 8400GS, 1 of them has it's shader overclocked to 1400mhz. For the 8600GT, the average shader clock for the data points I have recorded is 1472mhz, up from the 1180mhz stock speed.

Next step is to change my spreadsheet so the core/mem/shader clock speeds are recorded in their own cell so I can calculate their averages. Weeee.


I love your posts, I love your spreadsheets even more :!:

I'm almost in love :ewink:
MtM
 
Posts: 3054
Joined: Fri Jun 27, 2008 2:20 pm
Location: The Netherlands

PreviousNext

Return to NVIDIA specific issues

Who is online

Users browsing this forum: No registered users and 1 guest

cron