Folding with multiple GPUs?

[ddrueding]

Thinking about putting together a rig with some spare parts and have some questions about multiple GPUs.

1. Can there be 3 different cards in the three slots? (nVidia 400 and 500 series)

2. Do I need monitors connected to the cards? I was hoping to run it headless.

 

[CougTek]

You certainly don't have to plug monitors to them. Beware of the GeForce 400 series. Many people, including Mark, have had noise issues with them while running the GPU folding client. Some of them are buzzing (think about a coil noise from a defective PSU). I don't think the 500 series have the same problem.

 

[CougTek]

And I don't think, in fact I'm pretty sure, that mixing the card models would pose any problem for the GPU folding client. As long as you don't operate them in SLI, things should be fine.

 

[ddrueding]

Great, thanks. I hope to have that system up by the end of the month.

 

[Santilli]

WHY?
I was under the impression that running a video card at a high percentage is more likely to have it go south then the cpu?

I'm getting good speed running the multiple CPU setup, and, if you are doing that, doesn't it preclude doing folding on the gpu?

How about this for a f@H unit?
http://www.genstor.com/server_intel_GSMP_i508.php

 

[CougTek]

How about this for a f@H unit?
http://www.genstor.com/server_intel_GSMP_i508.php

Because it's not even affordable by SSDdrueding's standards.

And GPU folding doesn't seem to be that hard on the graphic cards. Your dear friend Mark has done most of its contribution from his GeForce cards and he's still going strong. I'm pretty sure that if he killed cards every now and then, he would have found another way to contribute and quit GPU folding.

 

[Santilli]

The advice I was given is either use the SMP mode, or the GPU, but, the SMP is faster, and, when running it, you can't use the GPU mode.

 

[CougTek]

You could not run the GPU client and expect to crunch big units back when there was big units for 8-core systems, but now that they upped the big units to 16-core systems, that limitation no longer applies.

I can assure you that you can run the SMP client together with the GPU client.

 

[Santilli]

OK:
I'm running mine from the cmd line install, and never got that far.

 

[CougTek]

Greg, I sent you a PM a while ago explaining you at lenght how to install and configure the FAH GPU Tracker V2 client. It is much simpler than the command line version. Why do you use the command line version? Are you a teflon kid?

 

[Santilli]

I tried that. Didn't work on my machine.
The CLV did.

I wish Stanford would send me a check for my electric bill. Or at least something for your dementia.
Teflon kid? GFYS

 

[Santilli]

Greg, I sent you a PM a while ago explaining you at lenght how to install and configure the FAH GPU Tracker V2 client. It is much simpler than the command line version. Why do you use the command line version? Are you a teflon kid?

I guess typing or proof reading escapes your skill set?

 

[CougTek]

Having a bad day?

 

[CougTek]

GFYS

Tried, not flexible enough. Sorry.

 

[Handruin]

Tried, not flexible enough. Sorry.

I'm impressed you even tried.

 

[Santilli]

If anyone has the right to be the bully asshole here it's me. I'd be happy to get in the ring with CougTek at 59. "Teflon Kid"? FY

 

[CougTek]

I almost fell treathen. Whatever. For the record, here's my PMs with you back in October and November when I tried to explain, clearly in my opinion, how to configure the very simple FAH GPU Tracker :

#1
Hi Greg,

It would help your points reward if you'd get a passkey for the folding@home client and if you would configure your client with it.
Also, do you use the FAH GPU tracker V2 tool to configure your client? It's simpler than the console client. If you don't know how, I could even login remotely to your computer and configure it for you if you want (assuming you trust me enough, of course).

If it's too much trouble for you, then it's ok. It's already kind of you to contribute. I'm just trying to help you (and our team) to get more points from your effort.
Have a nice evening,

Coug

#2

Have you enabled SMP for the client? If not, then it's no surprise only 13% of your eight-core CPU is used. You can enable it in the setup section of the FAH GPU tracker V2 tool. Choose "configure". In the client settings, opt for "Use multi-core SMP client" instead of the other option above it. Ideally, also check the "enable bigadv" and "Add -verbosity 9" below. Let me know if you now use 100% of your processor after that.
In the same window, in the "Tracker settings" section, be sure to copy the passkey they sent you and fill the team number (10047) and your username, of course.
Coug

Hi Greg,
It would help your points reward if you'd get a passkey for the folding@home client and if you would configure your client with it.
DONE
Also, do you use the FAH GPU tracker V2 tool to configure your client? It's simpler than the console client. If you don't know how, I could even login remotely to your computer and configure it for you if you want (assuming you trust me enough, of course).
INSTALLED HOWEVER, IT LISTS SMP "RUNNING PROJECT UNKNOWN POINTS UNKNOWN % COMPLETE 0% AND PPD 0.0
VIEWING STATS GIVES ALL 0%'S AS WELL.
If it's too much trouble for you, than it's ok. It's already kind of you to contribute. I'm just trying to help you (and our team) to get more points from your effort.
Have a nice evening,
Coug

tAKS mANAGER SHOWS AN A4 CORE RUNNING AT 13% cpu AND THE 78 CORE RUNNING AT 3% cpu
CPU FOLDING AT HOME ICON SHOWS 17/1500, AND WORKING

Coug
#3

Are there any websites I need to unblock that the FAH GPU Tracker V2 uses that peer block might be blocking?

Not that I'm aware of. But I don't block the edu list with peerblock. If you added that list to your blocked lists, then maybe you'll block stanford's servers. Otherwise, no, you won't have issue receiving and sending units.

Does using the GPU really help much, compared to the 940?

The i7 940 should give you between 12000 and 17000ppd, depending if you compute big units or not (the Project 6061 you are currently doing isn't a big unit). I don't remember what's your graphic card, but unless it's an Nvidia card with a Fermi GPU, ideally a GTS 450 or higher, you shouldn't bother. Also, if you start the GPU client, forget about receiving big units since you'd no longer have eight available cores to complete them. Each GPU client occupies one CPU core.
The simple answer to your question is "no".

#4

It's kind of strange. Resource monitor shows heavy, but spiking cpu usage on 4 cores/threads of the 8 available. RM also shows it's running 5 threads.
In trying to get the multiple core setup going, I lost 4300 of a 7000 project;-(

GS
--- Opening Log file [November 18 23:55:56 UTC]


# Windows CPU Console Edition #################################################
###############################################################################

Folding@Home Client Version 6.34

http://folding.stanford.edu

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

Launch directory: C:\Users\Santilli\FAH
Executable: C:\Users\Santilli\FAH\fah6.exe


[23:55:56] Configuring Folding@Home...


[00:04:00] - Ask before connecting: No
[00:04:00] - User name: Gregory_Santilli (Team 10047)
[00:04:00] - User ID: 14BFA6471E809C69
[00:04:00] - Machine ID: 1
[00:04:00]
[00:04:00] Work directory not found. Creating...
[00:04:00] Could not open work queue, generating new queue...
[00:04:00] - Preparing to get new work unit...
[00:04:00] Cleaning up work directory
[00:04:00] + Attempting to get work packet
[00:04:00] Passkey found
[00:04:00] - Connecting to assignment server
[00:04:01] - Successful: assigned to (129.74.85.15).
[00:04:01] + News From Folding@Home: Welcome to Folding@Home
[00:04:01] Loaded queue successfully.
[00:04:02] + Closed connections
[00:04:02]
[00:04:02] + Processing work unit
[00:04:02] A4 will attempt to use 0 threads.
[00:04:02] Core required: FahCore_a4.exe
[00:04:02] Core not found.
[00:04:02] - Core is not present or corrupted.
[00:04:02] - Attempting to download new core...
[00:04:02] + Downloading new core: FahCore_a4.exe
[00:04:02] + 10240 bytes downloaded
[00:04:02] + 20480 bytes downloaded
[00:04:02] + 30720 bytes downloaded
[00:04:02] + 40960 bytes downloaded
[00:04:02] + 51200 bytes downloaded
[00:04:02] + 61440 bytes downloaded
[00:04:02] + 71680 bytes downloaded
[00:04:02] + 81920 bytes downloaded
[00:04:02] + 92160 bytes downloaded
[00:04:02] + 102400 bytes downloaded
[00:04:02] + 112640 bytes downloaded
[00:04:02] + 122880 bytes downloaded
[00:04:02] + 133120 bytes downloaded
[00:04:02] + 143360 bytes downloaded
[00:04:02] + 153600 bytes downloaded
[00:04:02] + 163840 bytes downloaded
[00:04:02] + 174080 bytes downloaded
[00:04:02] + 184320 bytes downloaded
[00:04:02] + 194560 bytes downloaded
[00:04:02] + 204800 bytes downloaded
[00:04:02] + 215040 bytes downloaded
[00:04:02] + 225280 bytes downloaded
[00:04:02] + 235520 bytes downloaded
[00:04:02] + 245760 bytes downloaded
[00:04:02] + 256000 bytes downloaded
[00:04:02] + 266240 bytes downloaded
[00:04:02] + 276480 bytes downloaded
[00:04:02] + 286720 bytes downloaded
[00:04:02] + 296960 bytes downloaded
[00:04:02] + 307200 bytes downloaded
[00:04:02] + 317440 bytes downloaded
[00:04:02] + 327680 bytes downloaded
[00:04:02] + 337920 bytes downloaded
[00:04:02] + 348160 bytes downloaded
[00:04:02] + 358400 bytes downloaded
[00:04:02] + 368640 bytes downloaded
[00:04:02] + 378880 bytes downloaded
[00:04:02] + 389120 bytes downloaded
[00:04:02] + 399360 bytes downloaded
[00:04:02] + 409600 bytes downloaded
[00:04:02] + 419840 bytes downloaded
[00:04:02] + 430080 bytes downloaded
[00:04:02] + 440320 bytes downloaded
[00:04:02] + 450560 bytes downloaded
[00:04:02] + 460800 bytes downloaded
[00:04:02] + 471040 bytes downloaded
[00:04:02] + 481280 bytes downloaded
[00:04:02] + 491520 bytes downloaded
[00:04:02] + 501760 bytes downloaded
[00:04:02] + 512000 bytes downloaded
[00:04:02] + 522240 bytes downloaded
[00:04:02] + 532480 bytes downloaded
[00:04:02] + 542720 bytes downloaded
[00:04:02] + 552960 bytes downloaded
[00:04:02] + 563200 bytes downloaded
[00:04:02] + 573440 bytes downloaded
[00:04:02] + 583680 bytes downloaded
[00:04:02] + 593920 bytes downloaded
[00:04:02] + 604160 bytes downloaded
[00:04:02] + 614400 bytes downloaded
[00:04:02] + 624640 bytes downloaded
[00:04:02] + 634880 bytes downloaded
[00:04:02] + 645120 bytes downloaded
[00:04:02] + 655360 bytes downloaded
[00:04:02] + 665600 bytes downloaded
[00:04:03] + 675840 bytes downloaded
[00:04:03] + 686080 bytes downloaded
[00:04:03] + 696320 bytes downloaded
[00:04:03] + 706560 bytes downloaded
[00:04:03] + 716800 bytes downloaded
[00:04:03] + 727040 bytes downloaded
[00:04:03] + 737280 bytes downloaded
[00:04:03] + 747520 bytes downloaded
[00:04:03] + 757760 bytes downloaded
[00:04:03] + 768000 bytes downloaded
[00:04:03] + 778240 bytes downloaded
[00:04:03] + 788480 bytes downloaded
[00:04:03] + 798720 bytes downloaded
[00:04:03] + 808960 bytes downloaded
[00:04:03] + 819200 bytes downloaded
[00:04:03] + 829440 bytes downloaded
[00:04:03] + 839680 bytes downloaded
[00:04:03] + 849920 bytes downloaded
[00:04:03] + 860160 bytes downloaded
[00:04:03] + 870400 bytes downloaded
[00:04:03] + 880640 bytes downloaded
[00:04:03] + 890880 bytes downloaded
[00:04:03] + 901120 bytes downloaded
[00:04:03] + 911360 bytes downloaded
[00:04:03] + 921600 bytes downloaded
[00:04:03] + 931840 bytes downloaded
[00:04:03] + 942080 bytes downloaded
[00:04:03] + 952320 bytes downloaded
[00:04:03] + 962560 bytes downloaded
[00:04:03] + 972800 bytes downloaded
[00:04:03] + 983040 bytes downloaded
[00:04:03] + 993280 bytes downloaded
[00:04:03] + 1003520 bytes downloaded
[00:04:03] + 1013760 bytes downloaded
[00:04:03] + 1024000 bytes downloaded
[00:04:03] + 1034240 bytes downloaded
[00:04:03] + 1044480 bytes downloaded
[00:04:03] + 1054720 bytes downloaded
[00:04:03] + 1064960 bytes downloaded
[00:04:03] + 1075200 bytes downloaded
[00:04:03] + 1085440 bytes downloaded
[00:04:03] + 1095680 bytes downloaded
[00:04:03] + 1105920 bytes downloaded
[00:04:03] + 1116160 bytes downloaded
[00:04:03] + 1126400 bytes downloaded
[00:04:03] + 1136640 bytes downloaded
[00:04:03] + 1146880 bytes downloaded
[00:04:03] + 1157120 bytes downloaded
[00:04:03] + 1167360 bytes downloaded
[00:04:03] + 1177600 bytes downloaded
[00:04:03] + 1187840 bytes downloaded
[00:04:03] + 1198080 bytes downloaded
[00:04:03] + 1208320 bytes downloaded
[00:04:03] + 1218560 bytes downloaded
[00:04:03] + 1228800 bytes downloaded
[00:04:03] + 1239040 bytes downloaded
[00:04:03] + 1249280 bytes downloaded
[00:04:03] + 1259520 bytes downloaded
[00:04:03] + 1269760 bytes downloaded
[00:04:03] + 1280000 bytes downloaded
[00:04:03] + 1290240 bytes downloaded
[00:04:03] + 1300480 bytes downloaded
[00:04:03] + 1310720 bytes downloaded
[00:04:03] + 1320960 bytes downloaded
[00:04:03] + 1331200 bytes downloaded
[00:04:03] + 1341440 bytes downloaded
[00:04:03] + 1351680 bytes downloaded
[00:04:03] + 1361920 bytes downloaded
[00:04:03] + 1372160 bytes downloaded
[00:04:03] + 1382400 bytes downloaded
[00:04:03] + 1392640 bytes downloaded
[00:04:03] + 1402880 bytes downloaded
[00:04:03] + 1413120 bytes downloaded
[00:04:03] + 1423360 bytes downloaded
[00:04:03] + 1433600 bytes downloaded
[00:04:03] + 1443840 bytes downloaded
[00:04:03] + 1454080 bytes downloaded
[00:04:03] + 1464320 bytes downloaded
[00:04:03] + 1474560 bytes downloaded
[00:04:03] + 1484800 bytes downloaded
[00:04:03] + 1495040 bytes downloaded
[00:04:03] + 1505280 bytes downloaded
[00:04:03] + 1515520 bytes downloaded
[00:04:03] + 1525760 bytes downloaded
[00:04:03] + 1536000 bytes downloaded
[00:04:03] + 1546240 bytes downloaded
[00:04:03] + 1556480 bytes downloaded
[00:04:03] + 1566720 bytes downloaded
[00:04:03] + 1576960 bytes downloaded
[00:04:03] + 1587200 bytes downloaded
[00:04:03] + 1597440 bytes downloaded
[00:04:03] + 1607680 bytes downloaded
[00:04:03] + 1617920 bytes downloaded
[00:04:03] + 1628160 bytes downloaded
[00:04:03] + 1638400 bytes downloaded
[00:04:03] + 1648640 bytes downloaded
[00:04:03] + 1658880 bytes downloaded
[00:04:03] + 1669120 bytes downloaded
[00:04:03] + 1679360 bytes downloaded
[00:04:03] + 1689600 bytes downloaded
[00:04:03] + 1699840 bytes downloaded
[00:04:03] + 1710080 bytes downloaded
[00:04:03] + 1720320 bytes downloaded
[00:04:03] + 1730560 bytes downloaded
[00:04:03] + 1740800 bytes downloaded
[00:04:03] + 1751040 bytes downloaded
[00:04:03] + 1761280 bytes downloaded
[00:04:03] + 1771520 bytes downloaded
[00:04:03] + 1781760 bytes downloaded
[00:04:03] + 1792000 bytes downloaded
[00:04:03] + 1802240 bytes downloaded
[00:04:03] + 1812480 bytes downloaded
[00:04:03] + 1822720 bytes downloaded
[00:04:03] + 1832960 bytes downloaded
[00:04:03] + 1843200 bytes downloaded
[00:04:03] + 1853440 bytes downloaded
[00:04:03] + 1863680 bytes downloaded
[00:04:03] + 1873920 bytes downloaded
[00:04:03] + 1884160 bytes downloaded
[00:04:03] + 1894400 bytes downloaded
[00:04:04] + 1904640 bytes downloaded
[00:04:04] + 1914880 bytes downloaded
[00:04:04] + 1925120 bytes downloaded
[00:04:04] + 1935360 bytes downloaded
[00:04:04] + 1945600 bytes downloaded
[00:04:04] + 1955840 bytes downloaded
[00:04:04] + 1966080 bytes downloaded
[00:04:04] + 1976320 bytes downloaded
[00:04:04] + 1986560 bytes downloaded
[00:04:04] + 1996800 bytes downloaded
[00:04:04] + 2007040 bytes downloaded
[00:04:04] + 2017280 bytes downloaded
[00:04:04] + 2027520 bytes downloaded
[00:04:04] + 2037760 bytes downloaded
[00:04:04] + 2048000 bytes downloaded
[00:04:04] + 2058240 bytes downloaded
[00:04:04] + 2068480 bytes downloaded
[00:04:04] + 2078720 bytes downloaded
[00:04:04] + 2088960 bytes downloaded
[00:04:04] + 2099200 bytes downloaded
[00:04:04] + 2109440 bytes downloaded
[00:04:04] + 2119680 bytes downloaded
[00:04:04] + 2129920 bytes downloaded
[00:04:04] + 2140160 bytes downloaded
[00:04:04] + 2150400 bytes downloaded
[00:04:04] + 2160640 bytes downloaded
[00:04:04] + 2170880 bytes downloaded
[00:04:04] + 2181120 bytes downloaded
[00:04:04] + 2191360 bytes downloaded
[00:04:04] + 2201600 bytes downloaded
[00:04:04] + 2211840 bytes downloaded
[00:04:04] + 2222080 bytes downloaded
[00:04:04] + 2232320 bytes downloaded
[00:04:04] + 2242560 bytes downloaded
[00:04:04] + 2252800 bytes downloaded
[00:04:04] + 2263040 bytes downloaded
[00:04:04] + 2273280 bytes downloaded
[00:04:04] + 2283520 bytes downloaded
[00:04:04] + 2293760 bytes downloaded
[00:04:04] + 2304000 bytes downloaded
[00:04:04] + 2314240 bytes downloaded
[00:04:04] + 2324480 bytes downloaded
[00:04:04] + 2334720 bytes downloaded
[00:04:04] + 2344960 bytes downloaded
[00:04:04] + 2355200 bytes downloaded
[00:04:04] + 2365440 bytes downloaded
[00:04:04] + 2375680 bytes downloaded
[00:04:04] + 2385920 bytes downloaded
[00:04:04] + 2396160 bytes downloaded
[00:04:04] + 2406400 bytes downloaded
[00:04:04] + 2416640 bytes downloaded
[00:04:04] + 2426880 bytes downloaded
[00:04:04] + 2437120 bytes downloaded
[00:04:04] + 2447360 bytes downloaded
[00:04:04] + 2457600 bytes downloaded
[00:04:04] + 2467840 bytes downloaded
[00:04:04] + 2478080 bytes downloaded
[00:04:04] + 2488320 bytes downloaded
[00:04:04] + 2498560 bytes downloaded
[00:04:04] + 2508800 bytes downloaded
[00:04:04] + 2519040 bytes downloaded
[00:04:04] + 2529280 bytes downloaded
[00:04:04] + 2539520 bytes downloaded
[00:04:04] + 2549760 bytes downloaded
[00:04:04] + 2560000 bytes downloaded
[00:04:04] + 2570240 bytes downloaded
[00:04:04] + 2580480 bytes downloaded
[00:04:04] + 2590720 bytes downloaded
[00:04:04] + 2600960 bytes downloaded
[00:04:04] + 2611200 bytes downloaded
[00:04:04] + 2621440 bytes downloaded
[00:04:04] + 2631680 bytes downloaded
[00:04:04] + 2641920 bytes downloaded
[00:04:04] + 2652160 bytes downloaded
[00:04:04] + 2662400 bytes downloaded
[00:04:04] + 2672640 bytes downloaded
[00:04:04] + 2682880 bytes downloaded
[00:04:04] + 2693120 bytes downloaded
[00:04:04] + 2703360 bytes downloaded
[00:04:04] + 2713600 bytes downloaded
[00:04:04] + 2723840 bytes downloaded
[00:04:04] + 2734080 bytes downloaded
[00:04:04] + 2744320 bytes downloaded
[00:04:04] + 2754560 bytes downloaded
[00:04:04] + 2764800 bytes downloaded
[00:04:04] + 2775040 bytes downloaded
[00:04:04] + 2785280 bytes downloaded
[00:04:04] + 2795520 bytes downloaded
[00:04:04] + 2805760 bytes downloaded
[00:04:04] + 2816000 bytes downloaded
[00:04:04] + 2826240 bytes downloaded
[00:04:04] + 2836480 bytes downloaded
[00:04:04] + 2846720 bytes downloaded
[00:04:04] + 2856960 bytes downloaded
[00:04:04] + 2867200 bytes downloaded
[00:04:04] + 2877440 bytes downloaded
[00:04:04] + 2887680 bytes downloaded
[00:04:04] + 2897920 bytes downloaded
[00:04:04] + 2908160 bytes downloaded
[00:04:04] + 2918400 bytes downloaded
[00:04:04] + 2928640 bytes downloaded
[00:04:04] + 2938880 bytes downloaded
[00:04:04] + 2949120 bytes downloaded
[00:04:04] + 2959360 bytes downloaded
[00:04:04] + 2969600 bytes downloaded
[00:04:04] + 2979840 bytes downloaded
[00:04:04] + 2990080 bytes downloaded
[00:04:04] + 3000320 bytes downloaded
[00:04:04] + 3010560 bytes downloaded
[00:04:04] + 3020800 bytes downloaded
[00:04:05] + 3028899 bytes downloaded
[00:04:05] Verifying core Core_a4.fah...
[00:04:05] Signature is VALID
[00:04:05]
[00:04:05] Trying to unzip core FahCore_a4.exe
[00:04:05] Decompressed FahCore_a4.exe (10057216 bytes) successfully
[00:04:10] + Core successfully engaged
[00:04:18]
[00:04:19] + Processing work unit
[00:04:19] A4 will attempt to use 0 threads.
[00:04:19] Core required: FahCore_a4.exe
[00:04:19] Core found.
[00:04:19] Working on queue slot 01 [November 19 00:04:19 UTC]
[00:04:19] + Working ...
[00:04:19]
[00:04:19] *------------------------------*
[00:04:19] Folding@Home Gromacs GB Core
[00:04:19] Version 2.27 (Dec. 15, 2010)
[00:04:19]
[00:04:19] Preparing to commence simulation
[00:04:19] - Looking at optimizations...
[00:04:19] - Created dyn
[00:04:19] - Files status OK
[00:04:19] - Expanded 50447 -> 199068 (decompressed 394.6 percent)
[00:04:19] Called DecompressByteArray: compressed_data_size=50447 data_size=199068, decompressed_data_size=199068 diff=0
[00:04:19] - Digital signature verified
[00:04:19]
[00:04:19] Project: 7024 (Run 1, Clone 32, Gen 7
[00:04:19]
[00:04:19] Assembly optimizations on if available.
[00:04:19] Entering M.D.
[00:04:25] Mapping NT from 1 to 1
[00:04:25] Completed 0 out of 10000000 steps (0%)
[00:14:57] Completed 100000 out of 10000000 steps (1%)
[00:25:29] Completed 200000 out of 10000000 steps (2%)
[00:36:02] Completed 300000 out of 10000000 steps (3%)
[00:46:35] Completed 400000 out of 10000000 steps (4%)
[00:57:10] Completed 500000 out of 10000000 steps (5%)
[01:07:45] Completed 600000 out of 10000000 steps (6%)
[01:18:19] Completed 700000 out of 10000000 steps (7%)
[01:28:54] Completed 800000 out of 10000000 steps (8%)
[01:39:24] Completed 900000 out of 10000000 steps (9%)
[01:49:57] Completed 1000000 out of 10000000 steps (10%)
[02:00:33] Completed 1100000 out of 10000000 steps (11%)
[02:11:06] Completed 1200000 out of 10000000 steps (12%)
[02:21:39] Completed 1300000 out of 10000000 steps (13%)
[02:32:14] Completed 1400000 out of 10000000 steps (14%)
[02:42:48] Completed 1500000 out of 10000000 steps (15%)

CougTek> Why don't you just try to use the FAH GPU Tracker V2 like I told you to do? It's much simpler to configure.

That was months before your January complain. When I keep explaining to someone how to do a simple task and the person just doesn't follow simple instructions and then complain it doesn't work, asking if that person is/was a teflon kid is perfectly justified. I don't care wether you use the command line version or the FAH GPU tracker, or none at all. Just don't wine it doesn't work when I tell you how to make it work.

As for your complain about your electricity bill, no one forced you to run the client. I really doubt running the client on a single machine will increase your electricity bill by a 100$ a month (your other heating source probably contributed much more to that), it was your decision to use your computer that way. You're supposed to be a grown man. Stop looking at others to put the blame of your own decisions.

One last thing : I know Québec and California are not on the same time zone, but I don't think there a two months time difference between our places on Earth. Don't you think it's a bit late to go mad about my comment? It kinda had a "best before" moment and you're way past due.

 

[Santilli]

Thank you for posting that.

"treathen" Something you type when you have indulged in the "wine".

"Wine" something you drink. "Whine" what you are doing in the above post.

I know you are trying to help. However F@H software has a history of working on some configurations and not others.

FAH GPU tracker may not have worked for a variety of reasons. I can deal with that. Why would you be surprised it doesn't work on every computer?

From the research I did it was pretty much older generation software and others have had similar problems(FAH forums).

Your ad hom has no excuse and is not justified.

It's becoming standard procedure around here if you don't like what someone says use an ad hom.

I thought I'd take a page from your book and give it back.

Yes I am patient. I put up with this kind of stuff for a long time. I think you would be appreciating my thick skin and tolerance instead of whining about my timing.

I usually don't indulge in these sort of childish tit for tat bullshit in this forum. Nothing constructive comes from it. I have restrained myself from making comments
about your childish behavior. I suggest you mature a little bit and develop a bit of restraint with your ad homs.

 

[Santilli]

Power wise running at 100%:

Video Card 289 Watts

i7 940: idle 158 watts, peak 258 processor runs at 100% when doing FAH.

System tests have idle range for around 250 watts, with peak draw around 450 watts.

With the PG&E staggered billing system going over freezing can be expensive.

It does appear the key factors in our power bill have been the dishwasher, old server, and the antique central heating system often not being turned off by my roommate, in reverse order.

This month our test is LED lighting and if that reduces the bill.

 

[Santilli]

37" monitor 70W. LCD Vizio uses less energy then the LCD (94W), plus 20W for second monitor.

 

[LunarMist]

Are they calibrated to the same level? Some displays are more effecient, but some cheat on evenness of illumination (delta E).

 

[Santilli]

I don't know. I looked up the ratings on their website. I was thinking that with LED's I would use less energy then an LCD. I was wrong.

I suspect you are correct. The LED screen is edge lit. I guess that is more even and takes more power. It might also be some of the other components that go into the TV
and have nothing to do with the screen.

I'm sorry. I'm taking this off topic. The point is the considerable draw this machine makes running F@H at 100% processor would be enough to effect the electric bill.

 

[ddrueding]

Having an issue where every time a new WU is started my machine nearly freezes/studders for a couple minutes. To the point where Pandora starts and stops every couple seconds and the mouse is largely unresponsive. With 4 GPUs this is more than once an hour. If I can get the FAHControl up, I can see that one of the GPUs shows 0.0% to about 0.6%.

Any tips?

 

[P5-133XL]

I do not know why it only lags at the start of a new WU but if you can ID which slot runs the primary display, then pause it and see if that helps. Not folding on the primary display generally gets rid of all lag. If that helps then we can configure the client to permanently prevent that slot from running for three GPU slots is far better than none and the goal is for it not to affect your usage of the machine.

 

[ddrueding]

Interestingly enough, any slot lags the machine at the start of a WU. Pausing the primary does seem to fix the issue regardless of which slot is starting a new WU. I'm not going to stop folding, or even reduce production by 25%, but eventually fixing it would be nice.

 

[P5-133XL]

The lag problem is not going to go away any time soon. The primary issue is a lack of a priority system for GPU's and that is a large-scale change that will require big changes in HW, drivers, and Windows. The Windows video subsystem was never designed with the concept of multiple applications competing for GPU resources.

A folding screensaver is in development and that should help for many. For many, using onboard video as the primary GPU solves their problem. Neither is a perfect solution.

 

[ddrueding]

500k in 24-hours....not bad

 

[Handruin]

Wow, nice push! You caused the graph to blip.

My machines were offline during last week while I was away so I didn't help much. I only have one folding right now and hopefully the other one will be back online this evening.

 

[ddrueding]

Work 1 with Beta units = 303k PPD
Work 2 without Beta units = 105k PPD
Work 2 with Beta units = 56k PPD

Help?

 

[P5-133XL]

Need more details (like logs).

My immediate reaction is OC'ing. Check clock rates of your cards. Core_17 is very sensitive to OC'ing and will cause clock rates to spontaneously drop and stay that way till rebooted. I really maintain systray monitoring of clock rates, %GPU usage, and temp for all my video cards.

 

[ddrueding]

I'll grab some logs tomorrow. No manual OC'ing, but I believe the cards carry a higher clock from the mfgr. I have the most insane fan configuration on them at the moment with additional fans pushing air into the coolers, pulling it out, and moving tons of air around the boards/sinks. Unless the stock cooling system is completely inadequate, there is no way the cards are overheating.

 

[ddrueding]

Found the logs. No errors. But I also downloaded the EVGA Precision software so I could monitor the GPUs. Turns out they were thermally throttling while the onboard fan was running @ 35% (!?). Got that turned up a bit using the fan curve programming in the software and all is well. While I was here I couldn't help but start fooling around, and am now running the cards 25%+ overclocked (1306MHz). I'm assuming any issues will appear in the FAHControl logs file even with the "Warnings & Errors" box ticked?

 

[P5-133XL]

No, not necessarily. OC'ing can cause VRM's causing the clockrate to drop like a rock which won't show up in the logs. There is also a failure mode that causes the logs to just stop recording though the client thinks the cores are progressing but in reality they are not. Both of these seem to be related to OC'ing.

 

[ddrueding]

I'm watching the clockrate via GPU-Z and Precision, so they are fine. Ended up getting an error when the WU was submitted at 1306MHz, but they are going through fine @ 1201MHz. That machine (work2) is currently doing 385k PPD. Work1 is managing 301k PPD.

 

[CougTek]

My Radeon HD 7970 does 115K ppd on a Project 7810 unit. Even if I'm below average for that project given the card I have, I won't complain. 115K ppd is 115K ppd.

 

[CougTek]

...and now my Radeon HD 7950 (another system) is doing over 200K ppd on a project 8902 unit. This is good.