Milkyway@Home - GPU & CPU performance stats wanted, any capable h/w, old or new!

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
See for yourself:

Computer
Tasks

And a lot of the later tasks were without a dedicated core.

The only tasks I see are for 159.86 credit WUs (I went back as far as I could).
Is that what you submitted a time for?

[update]
Now I see why, your only crunching MW v1.02 WUs (how come?) which by looking from your & my tasks are exclusively 159.86 credit WUs, so the score is wrong.
I did say I needed times for 213.76 credit WUs only ;), although if this becomes a problem (your the 2nd user to only be crunching 159.86s) I may consider switching but I'd rather not now we've got quite a few results in.

Can't you enable the other apps?
 
Last edited:

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,360
4,066
75
Huh, there are other apps.

It would help to specify that users should enable only the Separation (Modified) app in their Milkyway@Home preferences. But even then, WUs seem to come in different sizes.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,360
4,066
75
for some reason I can't get 213.76? I'll keep trying.
I've got a bunch of 79.93s now, a bunch of 320.63s, but no 213.76s.

Edit: I got one 213.76. It's just one, but it took longer than the 320.63s. :eek:
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
Huh, there are other apps.

It would help to specify that users should enable only the Separation (Modified) app in their Milkyway@Home preferences. But even then, WUs seem to come in different sizes.
No I never meant that, their all on by default which is fine.
I thought you had turned off separation modified, if not then I don't get why your not getting those WUs :confused:.
 

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
Deerslayer
Turned out that ken g6 had used the wrong WU for the benchmark. So maybe your time is ok afterall?
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
:thumbsup:

Btw this article by AnandTech shows that your GPU has a 1/12th DP multiplier. Which going by the wiki gives your GPU DP power 105.3 GLFOPs, your HD 7770 GHz ed 1GB has 80 GFLOPs, a HD 7850 110, HD 4830 147, GTX 560 Ti 448 Core (1/8th multi) 164, to give you some idea of where your 560s performance should lie for MW & other DP apps.
It does seem on the face of it that your 560 is a little down, was it running at nr 100% GPU load?

I've added additional information as an edit.
I also corrected the amount of memory of the video card.
See Post # 13. Link
Ok thx :), but is it a 1GB card or 512MB? You say both ;)
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
Thx for the updates :).

GleeM
Going by the info in my previous post your GTX 560 Ti 448 Core @ 880 MHz has virtually the same DP power as an HD 4850, yet it crunches MW quite a bit faster :). I wonder what gives your card the advantage?
Btw if anything I'd say your 213.76 average was a it faster than your posted, I saw 5-6 @ ~467s.
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
Aka GPU utilisation, GPU-Z will show you that under the sensors tab.
I know AMD drivers show it too, don't know about NVidia.
 
Last edited:

deerslayer

Lifer
Jan 15, 2001
10,153
0
76
No, it's not staying at 100%...bouncing around. In fact, I haven't seen it hit 100% yet.

Same thing with the 7770. What setting do I need to change? I have it set to use GPU while computer is in use. I don't see anything else related to GPU usage.
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
It should stay near 100% not at 100, that's assuming you have a dedicated CPU core for each GPU. What's it hovering around at?

That's all you should need to do AFAIK, unless you were overclocking, but you said you weren't.
 

deerslayer

Lifer
Jan 15, 2001
10,153
0
76
It should stay near 100% not at 100, that's assuming you have a dedicated CPU core for each GPU. What's it hovering around at?

That's all you should need to do AFAIK, unless you were overclocking, but you said you weren't.

Around 80 to 90 percent.
 

biodoc

Diamond Member
Dec 29, 2005
6,285
2,238
136
I took a break from GPUGrid to generate some numbers for the table. One cpu core free for all times. There's no way I can check loads on linux.

780Ti: 395s on linux. 403s on win 8.1 (97% GPU load NO memory controller load and low temps and only 5s on CPU time per WU). Low temps too. Very odd.

570: 520s on linux

660Ti: 816s on linux

Fortunately I'm getting just over a million ppd on GPUGrid with these three cards. Over 500K ppd with the 780Ti. :p
 

Assimilator1

Elite Member
Nov 4, 1999
24,125
508
126
Thanks Biodoc, certainly needed more NVidia cards there.
Awesome output btw!...... on GPUGrid anyway :).

I take it their all on stock clocks?
 
Last edited:

Sunny129

Diamond Member
Nov 14, 2000
4,823
6
81
ok, so i had two dual-GPU machines crunching MW@H recently. one had dual Sapphire 7970s (not the same exact models though), and the other had dual Gigabyte 7970's (matching GPUs). the fact that the Sapphire 7970s didn't have matching model #'s is irrelevant though b/c they both had the same amount of memory (3GB) and were clocked the same. so, without further ado:

Sapphire 7970, core @ 950MHz (factory OCed), VRAM @ 800MHz (underclocked), BOINC v7.2.28, Catalyst 13.9, Win7 x64: ~67s

Gigabyte 7970, core @ 1000MHz (factory OCed), VRAM @ 850MHz (underclocked), BOINC v7.2.28, Catalyst 13.9, Win7 x64: ~69s

...now i know it doesn't make any sense that the average run time for the lower clocked Sapphire 7970 is slightly better than the average run time of the higher clocked Gigabyte 7970, particularly since the supporting hardware is all the same (both use the same CPU, mobo, type and amount of RAM, etc.)...but i suspect that the small sample size of 5 tasks might be the culprit. i used to have a back log of hundreds of valid tasks on each host, but those back logs have dwindled to ~60 valid tasks each since having pulled both hosts off the project 4 days ago...so unfortunately i can't simply add more tasks to the average without getting back on the project again for a few days...and i can't do that either b/c those two hosts have since been completely taken apart. at any rate, i suppose you should post the better of the two times i reported...
 

Icecold

Golden Member
Nov 15, 2004
1,100
1,028
146
Are these supposd to be the Milkyway@Home Separation (Modified Fit) tasks?

If so, my average for 5 is 176.27 seconds on a stock clocked 6950, Windows 7, BOINC 7.2.33.
 

Sunny129

Diamond Member
Nov 14, 2000
4,823
6
81
Are these supposd to be the Milkyway@Home Separation (Modified Fit) tasks?

If so, my average for 5 is 176.27 seconds on a stock clocked 6950, Windows 7, BOINC 7.2.33.
from what i can tell, all 213.76-point tasks are Modified Fit tasks...but not all Modified Fit tasks are 213.76-point tasks. a Modified Fit task can earn 79.93 points, 213.76 points, or 320.63 points...so please make sure that your average includes only Modified Fit tasks that earned 213.76 points (and not any Modified Fit tasks that earned either 79.93 points or 320.63 points).
 

Icecold

Golden Member
Nov 15, 2004
1,100
1,028
146
from what i can tell, all 213.76-point tasks are Modified Fit tasks...but not all Modified Fit tasks are 213.76-point tasks. a Modified Fit task can earn 79.93 points, 213.76 points, or 320.63 points...so please make sure that your average includes only Modified Fit tasks that earned 213.76 points (and not any Modified Fit tasks that earned either 79.93 points or 320.63 points).
Yeah they were the 213.76 point tasks that I averaged. There were way more 320 point tasks that I had to wade through to find the 213.76 ones. I just can't run any tasks on 6950 machine other than the Modified Fit tasks and wanted to make sure I was running the right ones. I'm running some on my GTX 480 now and will update once I have an average.