GPU3 beta available

and on that note the gpu3 server is now out of work.. arg!!!
Click to expand...

Maybe they underestimated how many of us rabid Folders would test the GPU3 beta as soon as it was available. :)
 
well they have WU's again now..

just noticed its a new WU.. 10627.. PPD isnt any different though.. but its worth 611 credits.. PPD is at 6360.3 1:23 TPF on a GTX 260 @ 602/1512/1161..
 
Last edited:
tjmagneto said:
I've been cured of my rabies and am back to GPU2. :rolleyes:
Click to expand...
i got the rabies vaccine in the form of folding teammates testing gpu3 before i did :)

lets let the fermi folders have gpu3
Posted via [H] Mobile Device
 
jebo_4jc said:
i got the rabies vaccine in the form of folding teammates testing gpu3 before i did :)

lets let the fermi folders have gpu3
Posted via [H] Mobile Device
Click to expand...


no they cant have it bad god damnit! its mine!!!!

honestly i like the gpu3 client.. i could care less about the loss in PPD because i can actually use my system while folding..
 
well looks like they have changed the GPU3 server so that it can now send gpu2 WU's to run on the gpu2 core when they run out of gpu3 WU's.. running a 6602WU now on the gpu2 core..
 
This is BS I haven't gotten a SMP WU in like two weeks and no ATI support for GPU 3. I had too many issues with GPU2 and had to remove it.
 
hpglow said:
This is BS I haven't gotten a SMP WU in like two weeks and no ATI support for GPU 3. I had too many issues with GPU2 and had to remove it.
Click to expand...


if you havent gotten an SMP WU in 2 weeks then you have other issues..
 
woohoo bios volt modded my GTX 260.. now doing 648/1566/1116(lowered the ram clock because it doesnt effect the gpu3 client.. but im now doing 6671 PPD(P10626 WU) compared to the 6200 PPD(P10626 WU) i was doing with the clocks at 602/1512/1161.. so from the looks of it these WU's scale pretty well depending on your shader clock..
 
Ensign.Fodder said:
Running new client but have to finish my GPU2 jobs first.

Will see.

Click to expand...


if you arent running a GTX 400 series card i wouldnt recommend using the GPU3 client.. the only reason ive been using it right now is that it doesnt effect my system when im watching movies.. with the GPU2 client i could watch movies but i could not do anything else on my system otherwise the movie would start lagging.. especially if i loaded a website that was using flash media player which had hardware accel support.. with the GPU3 client i can pretty much do what ever i want now on my left monitor while a movie is playing on my right monitor.. i can also leave the program running but paused since it hardly uses any video memory.. with the gpu2 client id have to close it to play a game since it holds the WU in the video memory while the program is paused which can take up almost 500mb of video memory..
 
zero2dash said:
I installed 197.45's last night on both, no problems (so far). PPD (gpu2) seems the same.

re: ATi v nV f@h
Not sure about GFLOPS, dumb number estimation in my head it looks like f@h favors green because nV uses a larger qty of stream processors with a lower gpu core clock vs. red (ATi) using a smaller qty SPs with a higher gpu core clock. (edit on the old cards at least) Folding craves/needs/wants more SPs than more mhz in the gpu area (as far as I can tell). Again though just my dumb math, just wanted to throw my $0.02 cents in the hat on that subject. :)

Let me rephrase/elaborate: (sorry :eek:)
nV - lower gpu core clock, higher # of shaders, higher shader clock
ATi - higher gpu core clock, lower # of shaders, lower shader clock
f@h eats the shaders, depends on # of shaders and shader clock. ATi cards have less of both that's why they fold like $#!t. :D
Click to expand...

Correlation does not imply causation.

I've written extensively as to the cause(s).

The two main factors were ATi hardware's inability to access threads in protected memory. This is used to buffer preliminary results into system memory (this has been rectified with the 5x00 series). You see, ATi's GPUs actually do twice the GFLOPs in F@H GPU2 in many instances yet end up with lower results.

This is because when an ATi GPU encounters an error. It must flush the entire calculation and start over (GPUs error often btw). NV GPUs supported this feature and could access preliminary results thus not needing to restart the calculation. They could simply revert to a calculation result prior to the error.

Another variance was the fact that GPU2 did not add new Features found on the 48x0 series (it was written with the 3870 in mind). Therefore by the time it was released Scattering support was not added for ATi GPUs (causing quite the large performance deficit).

Those are some of the reasons. Pound for pound ATi GPUs outmuscle NV GPUs (of the same generation) when it comes to pure Computational Performance.
 
ElMoIsEviL said:
Those are some of the reasons. Pound for pound ATi GPUs outmuscle NV GPUs (of the same generation) when it comes to pure Computational Performance.
Click to expand...

While they get smacked silly when it comes to GPGPU features, making all that performance moot.

It's like coupling a V8 to a 3-way gearbox :p
 
ppp353 said:
I get nothing but "shut down unstable machine" with my 8800GTS. Never had any trouble with GPU2 client. I'm running the 197.45's.

Maybe I'll just wait till the next release.
Click to expand...


I change to the 257.15 and all is working fine now with the gpu3 client.
 
Mr. Pedantic said:
Well, it shows in things like Milkyway@home and Collatz. Just not in Folding.
Click to expand...

Molecular Dynamics is probably a lot more involved, asking a lot more from the hardware than those other projects. Having some experience with simulations myself (albeit primarily in the area of hardware designs) I can definitely imagine how certain bottlenecks in a GPU can make it run like a dog compared to on a GPU without these bottlenecks. Heck, a design my company is thinking of porting to CUDA/OpenCL is already pushing the limits of what Fermi can do hardware-wise.
 
Mr. Pedantic said:
Well, it shows in things like Milkyway@home and Collatz. Just not in Folding.
Click to expand...


thats because F@H uses old worthless coding thats designed for cuda while ATI has moved on to far better things that boinc actually takes advantage of.. the difference is that if they do the ATI gpu3 client correctly it will use its own core which is based off openGL and/or directcompute and not that lame ass openMM shit that they use with cuda.. but thats highly unlikely since the 2 people that work on the GPU client are unpayed volunteers.. but then again they wouldnt want the nvidia people crying foul on them for creating a core that actually out performs nvidia now would we.. honestly i just wish they would call up some one that works on MW@H and or collatz and ask them to fix their stupid F@H gpu client..

Elledan said:
Molecular Dynamics is probably a lot more involved, asking a lot more from the hardware than those other projects. Having some experience with simulations myself (albeit primarily in the area of hardware designs) I can definitely imagine how certain bottlenecks in a GPU can make it run like a dog compared to on a GPU without these bottlenecks. Heck, a design my company is thinking of porting to CUDA/OpenCL is already pushing the limits of what Fermi can do hardware-wise.
Click to expand...


in the case of F@H its not about how demanding the WU's are its really about using old inefficient coding thats the problem and the fact that they just dont have the man power to actually design a up to date client.. also a lot of the problem is that they have to continue supporting older cards like the 8 and 9 series which dont have the same advantages that the GT200 and the GT100's have.. i mean hell MW@H didnt even have ATI support til a member of the project decided to go in and change some of the coding in the application client which ported right over to ATI without having to do any crazy modifications to the application.. so the problem isnt the differences between cuda and not using cuda.. its about staying ahead of the game instead of relying on old shit and then only changing it when your forced to change like in the case of the fermi cards.. if it wasnt for the fermi's not working in the GPU2 client i doubt we would of even seen the gpu3 client by the end of this year and its quite obvious when you consider the lack of optimization of the core so far.. i think by the end of the year we will see fermi cards pushing into the 20k PPD mark if the cores done correctly..

as far as the program your company uses.. if its already pushing what the fermi can do why bother porting it to it in the first place? that means you leave your self absolutely no headroom to add onto the program.. which means your stuck for another year and half before you see a refresh coming from Nvidia to add any extra performance that would benefit the program you are using.. and a year and half is being optimistic.. i really dont see anything new coming from Nvidia until early to mid 2012 because there sure as heck wont be a refresh of the GT100 series unless they decided to do a die shrink to 28nm which TSMC isnt ready for..
 
Last edited:
sirmonkey1985 said:
thats because F@H uses old worthless coding thats designed for cuda while ATI has moved on to far better things that boinc actually takes advantage of.. the difference is that if they do the ATI gpu3 client correctly it will use its own core which is based off openGL and/or directcompute and not that lame ass openMM shit that they use with cuda.. but thats highly unlikely since the 2 people that work on the GPU client are unpayed volunteers.. but then again they wouldnt want the nvidia people crying foul on them for creating a core that actually out performs nvidia now would we.. honestly i just wish they would call up some one that works on MW@H and or collatz and ask them to fix their stupid F@H gpu client..
Click to expand...

well ati/amd suppose to helping them now make their gpu3 core. thought if the MW@H or collatz guy where doing it i think we would have a next client in weeks not months.
 
Elledan said:
While they get smacked silly when it comes to GPGPU features, making all that performance moot.

It's like coupling a V8 to a 3-way gearbox :p
Click to expand...

What features specifically?

Could you elaborate and specify which features ATi GPUs lack (aside from those I mentioned) which would make all that performance "moot"?
 
ElMoIsEviL said:
What features specifically?

Could you elaborate and specify which features ATi GPUs lack (aside from those I mentioned) which would make all that performance "moot"?
Click to expand...


lol none.. the only difference is that Nvidia calls everything cuda while ATI doesnt bother putting a brand name on the exact same features cuda has.. :p
 
Just reporting back in, got glados stable, eith the fan at 85% the temps are 60-64*, not bad considering ambient is 80 right now

Not sure on PPD tho, FAHmon puts me at 1,200 PPD but Ive already finished 7 WU's today alone at 600 a piece, so thats off lol
 
Vaulter98c said:
Just reporting back in, got glados stable, eith the fan at 85% the temps are 60-64*, not bad considering ambient is 80 right now

Not sure on PPD tho, FAHmon puts me at 1,200 PPD but Ive already finished 7 WU's today alone at 600 a piece, so thats off lol
Click to expand...



470 does about 14k PPD i believe.. some where in that range..
 
Considering FahMon hasn't been updated in over a year, I wouldn't trust it for anything at this point. HFM.NET is pretty much better in every way anyway.
 
I've always had issues with HFM tho, times are all screwy. Hopefully a nice windows bigadv client will be out by the time I finish training and I can switch everthing over to windows and be on my merry way with 7 on all rigs and no need for VMs, then my times will read right and I won't have screwy network issues.

I might just switch over to 7 on all my rigs next month just to simplify things, maybe then the times will read right?
Posted via [H] Mobile Device
 
Vaulter98c said:
I've always had issues with HFM tho, times are all screwy. Hopefully a nice windows bigadv client will be out by the time I finish training and I can switch everthing over to windows and be on my merry way with 7 on all rigs and no need for VMs, then my times will read right and I won't have screwy network issues.

I might just switch over to 7 on all my rigs next month just to simplify things, maybe then the times will read right?
Posted via [H] Mobile Device
Click to expand...


if im not mistaken theres an option in HFM to tell it to ignore the times..
 
But the times to estimate ppd are what throws me, VM times are always wayoff, this screwing ppd numbers
Posted via [H] Mobile Device
 
Results from an EVGA 480 GTX running GPU3. everything set to stock (core=701,shader=1401,mem=1848).
it took about 1 hour, 3 mins to complete one WU. this is with A3 also running @ 3.8ghz (the A3 TPF dropped from 3min -> 3min,30sec while GPU3 was running).
so at 610p/WU = ~14k ppd

gpu temps basically stayed at 90c, fan speed (auto) hovered around 66-68%. this is all in my basement btw which is somewhat cool. installed in an Antec 902 with intake fans set to low.

the fan noise actually wasn't as loud as i thought it would be. however there was an annoying high-pitch noise coming from somewhere while the GPU client runs.. i'm guessing either the video card or the PSU.


Code: 
--- Opening Log file [May 30 18:43:53 UTC] 


# Windows GPU Systray Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.30r2

                          http://folding.stanford.edu

###############################################################################
###############################################################################

[18:43:53] - Ask before connecting: No
[18:43:53] - User name: Ryguyvi (Team 33)
[18:43:53] - User ID: 4760738065391920
[18:43:53] - Machine ID: 2
[18:43:53] 
[18:43:53] Gpu type=3 species=30.
[18:43:53] Could not open work queue, generating new queue...
[18:43:53] Initialization complete
[18:43:53] - Preparing to get new work unit...
[18:43:53] Cleaning up work directory
[18:43:53] + Attempting to get work packet
[18:43:53] Passkey found
[18:43:53] Gpu type=3 species=30.
[18:43:53] - Connecting to assignment server
[18:43:54] - Successful: assigned to (171.67.108.20).
[18:43:54] + News From Folding@Home: Welcome to Folding@Home
[18:43:54] Loaded queue successfully.
[18:43:54] Gpu type=3 species=30.
[18:43:55] + Closed connections
[18:43:55] 
[18:43:55] + Processing work unit
[18:43:55] Core required: FahCore_15.exe
[18:43:55] Core not found.
[18:43:55] - Core is not present or corrupted.
[18:43:55] - Attempting to download new core...
[18:43:55] + Downloading new core: FahCore_15.exe
[18:43:55] + 10240 bytes downloaded
[18:43:55] + 20480 bytes downloaded
[18:43:56] + 30720 bytes downloaded
[18:43:56] + 40960 bytes downloaded
[18:43:56] + 51200 bytes downloaded
[18:43:56] + 61440 bytes downloaded
[18:43:56] + 71680 bytes downloaded
[18:43:56] + 81920 bytes downloaded
[18:43:56] + 92160 bytes downloaded
[18:43:56] + 102400 bytes downloaded
[18:43:56] + 112640 bytes downloaded
[18:43:56] + 122880 bytes downloaded
[18:43:56] + 133120 bytes downloaded
[18:43:56] + 143360 bytes downloaded
[18:43:56] + 153600 bytes downloaded
[18:43:56] + 163840 bytes downloaded
[18:43:56] + 174080 bytes downloaded
[18:43:56] + 184320 bytes downloaded
[18:43:57] + 194560 bytes downloaded
[18:43:57] + 204800 bytes downloaded
[18:43:57] + 215040 bytes downloaded
[18:43:57] + 225280 bytes downloaded
[18:43:57] + 235520 bytes downloaded
[18:43:57] + 245760 bytes downloaded
[18:43:57] + 256000 bytes downloaded
[18:43:57] + 266240 bytes downloaded
[18:43:57] + 276480 bytes downloaded
[18:43:57] + 286720 bytes downloaded
[18:43:57] + 296960 bytes downloaded
[18:43:57] + 307200 bytes downloaded
[18:43:57] + 317440 bytes downloaded
[18:43:57] + 327680 bytes downloaded
[18:43:57] + 337920 bytes downloaded
[18:43:57] + 348160 bytes downloaded
[18:43:57] + 358400 bytes downloaded
[18:43:57] + 368640 bytes downloaded
[18:43:57] + 378880 bytes downloaded
[18:43:57] + 389120 bytes downloaded
[18:43:57] + 399360 bytes downloaded
[18:43:57] + 409600 bytes downloaded
[18:43:57] + 419840 bytes downloaded
[18:43:57] + 430080 bytes downloaded
[18:43:57] + 440320 bytes downloaded
[18:43:57] + 450560 bytes downloaded
[18:43:58] + 460800 bytes downloaded
[18:43:58] + 471040 bytes downloaded
[18:43:58] + 481280 bytes downloaded
[18:43:58] + 491520 bytes downloaded
[18:43:58] + 501760 bytes downloaded
[18:43:58] + 512000 bytes downloaded
[18:43:58] + 522240 bytes downloaded
[18:43:58] + 532480 bytes downloaded
[18:43:58] + 542720 bytes downloaded
[18:43:58] + 552960 bytes downloaded
[18:43:58] + 563200 bytes downloaded
[18:43:58] + 573440 bytes downloaded
[18:43:58] + 583680 bytes downloaded
[18:43:58] + 593920 bytes downloaded
[18:43:58] + 604160 bytes downloaded
[18:43:58] + 614400 bytes downloaded
[18:43:58] + 624640 bytes downloaded
[18:43:58] + 634880 bytes downloaded
[18:43:58] + 645120 bytes downloaded
[18:43:58] + 655360 bytes downloaded
[18:43:58] + 665600 bytes downloaded
[18:43:58] + 675840 bytes downloaded
[18:43:58] + 686080 bytes downloaded
[18:43:58] + 696320 bytes downloaded
[18:43:58] + 706560 bytes downloaded
[18:43:58] + 716800 bytes downloaded
[18:43:58] + 727040 bytes downloaded
[18:43:58] + 737280 bytes downloaded
[18:43:58] + 747520 bytes downloaded
[18:43:58] + 757760 bytes downloaded
[18:43:59] + 768000 bytes downloaded
[18:43:59] + 778240 bytes downloaded
[18:43:59] + 788480 bytes downloaded
[18:43:59] + 798720 bytes downloaded
[18:43:59] + 808960 bytes downloaded
[18:43:59] + 819200 bytes downloaded
[18:43:59] + 829440 bytes downloaded
[18:43:59] + 839680 bytes downloaded
[18:43:59] + 849920 bytes downloaded
[18:43:59] + 860160 bytes downloaded
[18:43:59] + 870400 bytes downloaded
[18:43:59] + 880640 bytes downloaded
[18:43:59] + 890880 bytes downloaded
[18:43:59] + 901120 bytes downloaded
[18:43:59] + 911360 bytes downloaded
[18:43:59] + 921600 bytes downloaded
[18:43:59] + 931840 bytes downloaded
[18:43:59] + 942080 bytes downloaded
[18:43:59] + 952320 bytes downloaded
[18:43:59] + 962560 bytes downloaded
[18:43:59] + 972800 bytes downloaded
[18:43:59] + 983040 bytes downloaded
[18:43:59] + 993280 bytes downloaded
[18:43:59] + 1003520 bytes downloaded
[18:43:59] + 1013760 bytes downloaded
[18:43:59] + 1024000 bytes downloaded
[18:43:59] + 1034240 bytes downloaded
[18:43:59] + 1044480 bytes downloaded
[18:43:59] + 1054720 bytes downloaded
[18:43:59] + 1064960 bytes downloaded
[18:43:59] + 1075200 bytes downloaded
[18:43:59] + 1085440 bytes downloaded
[18:43:59] + 1095680 bytes downloaded
[18:43:59] + 1105920 bytes downloaded
[18:43:59] + 1116160 bytes downloaded
[18:43:59] + 1126400 bytes downloaded
[18:43:59] + 1136640 bytes downloaded
[18:43:59] + 1146880 bytes downloaded
[18:44:00] + 1157120 bytes downloaded
[18:44:00] + 1167360 bytes downloaded
[18:44:00] + 1177600 bytes downloaded
[18:44:00] + 1187840 bytes downloaded
[18:44:00] + 1198080 bytes downloaded
[18:44:00] + 1208320 bytes downloaded
[18:44:00] + 1218560 bytes downloaded
[18:44:00] + 1228800 bytes downloaded
[18:44:00] + 1239040 bytes downloaded
[18:44:00] + 1249280 bytes downloaded
[18:44:00] + 1259520 bytes downloaded
[18:44:00] + 1269760 bytes downloaded
[18:44:00] + 1280000 bytes downloaded
[18:44:00] + 1290240 bytes downloaded
[18:44:00] + 1300480 bytes downloaded
[18:44:00] + 1310720 bytes downloaded
[18:44:00] + 1320960 bytes downloaded
[18:44:00] + 1331200 bytes downloaded
[18:44:00] + 1341440 bytes downloaded
[18:44:00] + 1351680 bytes downloaded
[18:44:00] + 1361920 bytes downloaded
[18:44:00] + 1367545 bytes downloaded
[18:44:00] Verifying core Core_15.fah...
[18:44:00] Signature is VALID
[18:44:00] 
[18:44:00] Trying to unzip core FahCore_15.exe
[18:44:00] Decompressed FahCore_15.exe (4136960 bytes) successfully
[18:44:05] + Core successfully engaged
[18:44:10] 
[18:44:10] + Processing work unit
[18:44:10] Core required: FahCore_15.exe
[18:44:10] Core found.
[18:44:10] Working on queue slot 01 [May 30 18:44:10 UTC]
[18:44:10] + Working ...
[18:44:11] 
[18:44:11] *------------------------------*
[18:44:11] Folding@Home GPU Core -- Beta
[18:44:11] Version 2.09 (Thu May 20 11:58:42 PDT 2010)
[18:44:11] 
[18:44:11] Build host: SimbiosNvdWin7
[18:44:11] Board Type: Nvidia
[18:44:11] Core      : 
[18:44:11] Preparing to commence simulation
[18:44:11] - Looking at optimizations...
[18:44:11] DeleteFrameFiles: successfully deleted file=work/wudata_01.ckp
[18:44:11] - Created dyn
[18:44:11] - Files status OK
[18:44:11] sizeof(CORE_PACKET_HDR) = 512 file=<>
[18:44:11] - Expanded 41720 -> 163067 (decompressed 390.8 percent)
[18:44:11] Called DecompressByteArray: compressed_data_size=41720 data_size=163067, decompressed_data_size=163067 diff=0
[18:44:11] - Digital signature verified
[18:44:11] 
[18:44:11] Project: 10626 (Run 172, Clone 36, Gen 0)
[18:44:11] 
[18:44:11] Assembly optimizations on if available.
[18:44:11] Entering M.D.
[18:44:17] Tpr hash work/wudata_01.tpr:  2812901548 1770194687 343485951 3123841833 2835386662
[18:44:17] Working on 582 p2750_N68H_AM03
[18:44:17] Client config found, loading data.
[18:44:17] Starting GUI Server
[18:44:54] Completed 1%
[18:45:32] Completed 2%
[18:46:08] Completed 3%
[18:46:45] Completed 4%
[18:47:22] Completed 5%
[18:47:59] Completed 6%
[18:48:37] Completed 7%
[18:49:14] Completed 8%
[18:49:51] Completed 9%
[18:50:27] Completed 10%
[18:51:04] Completed 11%
[18:51:41] Completed 12%
[18:52:18] Completed 13%
[18:52:55] Completed 14%
[18:53:32] Completed 15%
[18:54:09] Completed 16%
[18:54:46] Completed 17%
[18:55:23] Completed 18%
[18:56:00] Completed 19%
[18:56:37] Completed 20%
[18:57:13] Completed 21%
[18:57:50] Completed 22%
[18:58:28] Completed 23%
[18:59:05] Completed 24%
[18:59:42] Completed 25%
[19:00:19] Completed 26%
[19:00:56] Completed 27%
[19:01:33] Completed 28%
[19:02:09] Completed 29%
[19:02:46] Completed 30%
[19:03:23] Completed 31%
[19:04:00] Completed 32%
[19:04:37] Completed 33%
[19:05:14] Completed 34%
[19:05:50] Completed 35%
[19:06:27] Completed 36%
[19:07:04] Completed 37%
[19:07:41] Completed 38%
[19:08:18] Completed 39%
[19:08:55] Completed 40%
[19:09:32] Completed 41%
[19:10:09] Completed 42%
[19:10:46] Completed 43%
[19:11:23] Completed 44%
[19:12:00] Completed 45%
[19:12:37] Completed 46%
[19:13:14] Completed 47%
[19:13:51] Completed 48%
[19:14:28] Completed 49%
[19:15:05] Completed 50%
[19:15:41] Completed 51%
[19:16:18] Completed 52%
[19:16:55] Completed 53%
[19:17:32] Completed 54%
[19:18:09] Completed 55%
[19:18:47] Completed 56%
[19:19:23] Completed 57%
[19:20:00] Completed 58%
[19:20:37] Completed 59%
[19:21:14] Completed 60%
[19:21:51] Completed 61%
[19:22:28] Completed 62%
[19:23:05] Completed 63%
[19:23:42] Completed 64%
[19:24:19] Completed 65%
[19:24:56] Completed 66%
[19:25:33] Completed 67%
[19:26:10] Completed 68%
[19:26:47] Completed 69%
[19:27:23] Completed 70%
[19:28:00] Completed 71%
[19:28:38] Completed 72%
[19:29:14] Completed 73%
[19:29:51] Completed 74%
[19:30:28] Completed 75%
[19:31:05] Completed 76%
[19:31:42] Completed 77%
[19:32:19] Completed 78%
[19:32:55] Completed 79%
[19:33:32] Completed 80%
[19:34:09] Completed 81%
[19:34:46] Completed 82%
[19:35:23] Completed 83%
[19:36:00] Completed 84%
[19:36:36] Completed 85%
[19:37:13] Completed 86%
[19:37:50] Completed 87%
[19:38:27] Completed 88%
[19:39:04] Completed 89%
[19:39:41] Completed 90%
[19:40:18] Completed 91%
[19:40:55] Completed 92%
[19:41:32] Completed 93%
[19:42:08] Completed 94%
[19:42:45] Completed 95%
[19:43:22] Completed 96%
[19:43:59] Completed 97%
[19:44:36] Completed 98%
[19:45:13] Completed 99%
[19:45:50] Completed 100%
[19:45:50] Finished fah_main
[19:45:50] 
[19:45:50] Successful run
[19:45:50] DynamicWrapper: Finished Work Unit: sleep=10000
[19:46:00] Reserved 11340 bytes for xtc file; Cosm status=0
[19:46:00] Allocated 11340 bytes for xtc file
[19:46:00] - Reading up to 11340 from "work/wudata_01.xtc": Read 11340
[19:46:00] Read 11340 bytes from xtc file; available packet space=786419124
[19:46:00] xtc file hash check passed.
[19:46:00] Reserved 3621840 3621840 786419124 bytes for arc file=<work/wudata_01.trr> Cosm status=0
[19:46:00] Allocated 3621840 bytes for arc file
[19:46:00] - Reading up to 3621840 from "work/wudata_01.trr": Read 3621840
[19:46:00] Read 3621840 bytes from arc file; available packet space=782797284
[19:46:00] trr file hash check passed.
[19:46:00] Allocated 544 bytes for edr file
[19:46:00] Read bedfile
[19:46:00] edr file hash check passed.
[19:46:00] Allocated 492609 bytes for logfile
[19:46:00] Read logfile
[19:46:00] GuardedRun: success in DynamicWrapper
[19:46:00] GuardedRun: done
[19:46:00] Run: GuardedRun completed.
[19:46:01] + Opened results file
[19:46:01] - Writing 4126845 bytes of core data to disk...
[19:46:02] Done: 4126333 -> 3097821 (compressed to 75.0 percent)
[19:46:02]   ... Done.
[19:46:02] DeleteFrameFiles: successfully deleted file=work/wudata_01.ckp
[19:46:02] Shutting down core 
[19:46:02] 
[19:46:02] Folding@home Core Shutdown: FINISHED_UNIT
[19:46:05] CoreStatus = 64 (100)
[19:46:05] Sending work to server
[19:46:05] Project: 10626 (Run 172, Clone 36, Gen 0)


[19:46:05] + Attempting to send results [May 30 19:46:05 UTC]
[19:46:05] Gpu type=3 species=30.
[19:46:25] + Results successfully sent
[19:46:25] Thank you for your contribution to Folding@Home.
[19:46:25] + Number of Units Completed: 4

[19:46:30] - Preparing to get new work unit...
[19:46:30] Cleaning up work directory
[19:46:30] + Attempting to get work packet
[19:46:30] Passkey found
[19:46:30] Gpu type=3 species=30.
[19:46:30] - Connecting to assignment server
[19:46:30] - Successful: assigned to (171.67.108.20).
[19:46:30] + News From Folding@Home: Welcome to Folding@Home
[19:46:30] Loaded queue successfully.
[19:46:30] Gpu type=3 species=30.
[19:46:31] + Closed connections
[19:46:31] 
[19:46:31] + Processing work unit
[19:46:31] Core required: FahCore_15.exe
[19:46:31] Core found.
[19:46:31] Working on queue slot 02 [May 30 19:46:31 UTC]
[19:46:31] + Working ...
[19:46:31] 
[19:46:31] *------------------------------*
[19:46:31] Folding@Home GPU Core -- Beta
[19:46:31] Version 2.09 (Thu May 20 11:58:42 PDT 2010)
[19:46:31] 
[19:46:31] Build host: SimbiosNvdWin7
[19:46:31] Board Type: Nvidia
[19:46:31] Core      : 
[19:46:31] Preparing to commence simulation
[19:46:31] - Looking at optimizations...
[19:46:31] DeleteFrameFiles: successfully deleted file=work/wudata_02.ckp
[19:46:31] - Created dyn
[19:46:31] - Files status OK
[19:46:31] sizeof(CORE_PACKET_HDR) = 512 file=<>
[19:46:31] - Expanded 43799 -> 163067 (decompressed 372.3 percent)
[19:46:31] Called DecompressByteArray: compressed_data_size=43799 data_size=163067, decompressed_data_size=163067 diff=0
[19:46:31] - Digital signature verified
[19:46:31] 
[19:46:31] Project: 10626 (Run 127, Clone 0, Gen 1)
[19:46:31] 
[19:46:31] Assembly optimizations on if available.
[19:46:31] Entering M.D.
[19:46:37] Tpr hash work/wudata_02.tpr:  4102179661 1384014876 549200760 147204006 2898162853
[19:46:37] Working on 582 p2750_N68H_AM03
[19:46:37] Client config found, loading data.
[19:46:38] Starting GUI Server
[19:47:15] Completed 1%
 
470 at 1420 ish shaders gives me 13.9 k ppd, good to see it matches the much more expensive 480 so closely

I'll try later for a better OC but while I'm gone it's going stock

might look into getting a water cooled setup next year
Posted via [H] Mobile Device
 
High pitch wine? Pause the client, if it stops that instant than it's the coils
Posted via [H] Mobile Device

EDIT
point update, 1485 on a 470 gets me 14,800 ppd. Fans at 85%, about the same noise as my 295's made on 100%, temps are a good 30-40 less tho, I'm running 63-64 now, much better than the 100+ I had on STOCK 295's, and I'm sure the power draw is close to the same for each card
 
Last edited:
good results on those 470s
Posted via [H] Mobile Device
 
Wow, with a price drop the GTX 470 could be a folding beast. It seems you can really crank up the fans and keep temps very low. I was thinking with the fans at 100% you would be somewhere in the high 70's under folding load but low 60's are awesome. What is your ambient temp Vaulter?
 
Ambient is 81F, open case, no direct fan on the card. At stock, fan at 100%, temps are 59-62, at 1485 and fan at 85% I'm getting about 66 now

fans at 85 may seem loud to you but run a quad Sli setup and you'll get used to it. Also, ambient was 20 lower and cards were 35 higher with 295's, same point values
Posted via [H] Mobile Device
 
Vaulter98c said:
High pitch wine? Pause the client, if it stops that instant than it's the coils
Posted via [H] Mobile Device

EDIT
point update, 1485 on a 470 gets me 14,800 ppd. Fans at 85%, about the same noise as my 295's made on 100%, temps are a good 30-40 less tho, I'm running 63-64 now, much better than the 100+ I had on STOCK 295's, and I'm sure the power draw is close to the same for each card
Click to expand...

yeah. the instant i pause the GPU client, the high-pitch whine instantly stops. unpause and it's back on right away. do all video cards do this or is there a way to eliminate it?
 
"Coils."

I thought capacitors caused that. No, there is nothing I know that you can do about that except reduce the load on the video card.

Some video cards do that, others don't. A couple of mine whine like that, but I've got tinnitus so it's not usually very noticeable to me.
 
georgexi said:
yeah. the instant i pause the GPU client, the high-pitch whine instantly stops. unpause and it's back on right away. do all video cards do this or is there a way to eliminate it?
Click to expand...


some do some dont.. theres no real way to stop it exept RMA the card.. some will accept the RMA because of the coil whine some wont..


Vaulter98c said:
High pitch wine? Pause the client, if it stops that instant than it's the coils
Posted via [H] Mobile Device

EDIT
point update, 1485 on a 470 gets me 14,800 ppd. Fans at 85%, about the same noise as my 295's made on 100%, temps are a good 30-40 less tho, I'm running 63-64 now, much better than the 100+ I had on STOCK 295's, and I'm sure the power draw is close to the same for each card
Click to expand...

yeah during the night i run the GPU2 client.. during the day i run the GPU3 client.. just because of the temp differences.. its finally starting to warm up here in california so ill probably be stopping the GPU2 client completely and running GPU3 for now on because of the huge temp difference..
 
Linden said:
"Coils."

I thought capacitors caused that. No, there is nothing I know that you can do about that except reduce the load on the video card.

Some video cards do that, others don't. A couple of mine whine like that, but I've got tinnitus so it's not usually very noticeable to me.
Click to expand...

Capacitors are solid-state. If they make any noise it's usually very brief and catastrophic :)

Coil whine is often caused by feedback from the videocard VRMs. In the past I have heard (of) cases with mainboard VRMs causing a similar kind of feedback which causes the coils inside the PSU to vibrate at an audible frequency.

Only thing you can do about it other than gluing/heatshrink-wrapping the coils is change PSU/GPU, although I had one case in which a system rebuild fixed the issue as well.
 
Always wanted to do the coil glue mod as 7 of my old GPUs used to do that, but I just ended up selling them lol
Posted via [H] Mobile Device
 
You must log in or register to reply here.
Back
Top