Manual setting of GPU flags?

First, love the program. It has gotten me folding again. Now, my issue...My current setup runs a XFX Radeon HD 5870 for graphics rendering and an EVGA GTX 260 core 216 SC for PhysX (I am running the drivers that Nvidia forgot to put in the ATI block). My issue is that for some reason any GPU client I make is being pushed to the GTX 260. To see if it was my system, I downloaded the normal GPU3 F@H client, set the flags (including "-gpu 0"), and my 5870 was folding perfectly. Is there anyway to manually change the flag on one of the GPU clients to "-gpu 0" so that I can fold with both my GPUs and CPU within this program, or do I need to run this prog for CPU and the 260 (-gpu 1) while using the normal client for the 5870? The program runs the client perfectly for the 260 and the SMP client (Core i7 920), just not the 5870. Thanks.

Sherman J. Buster

have you tried running the -gpu 1 flag?

How can you set the flag when using the FAH GPU Tracker V2 on an individual client? I cannot find where the flags are located at.

sigvoror wrote:How can you set the flag when using the FAH GPU Tracker V2 on an individual client? I cannot find where the flags are located at.

To change the -gpu x flag you will need to edit client.cfg in one of the client folders.

Ok, client.cfg:

[settings]
username=Anonymous
team=0
passkey=
asknet=no
machineid=2

[http]
active=no
host=localhost
port=8080
usereg=no

Would crash program if I tried editing the file...so I copied and overwrote the client.cfg file from the F@H client I downloaded myself (the one that runs on the 5870). New client.cfg looks like this:

[settings]
username=sigvoror
team=33
passkey=
asknet=no
machineid=2
bigpackets=normal
extra_parms=-gpu 0

[http]
active=no
host=localhost
port=8080
usereg=no
proxy_name=
proxy_passwd=

[core]
priority=0
cpuusage=100
disableassembly=no
nocpulock=0
checkpoint=15

[power]
battery=no



This is log file from work folder (FAH GPU Tracker V2\GPU0\work:


*------------------------------*
Folding@Home GPU Core - Beta
Version 1.24 (Mon Feb 9 11:00:12 PST 2009)

Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
Build host: amoeba
Board Type: AMD
Core :
Preparing to commence simulation
- Assembly optimizations manually forced on.
- Not checking prior termination.
- Expanded 70377 -> 360060 (decompressed 511.6 percent)
Called DecompressByteArray: compressed_data_size=70377 data_size=360060, decompressed_data_size=360060 diff=0
- Digital signature verified

Project: 5742 (Run 4, Clone 358, Gen 241)

Assembly optimizations on if available.
Entering M.D.
Got kill signal -- issuing INTERRUPTED core shutdown

Folding@home Core Shutdown: INTERRUPTED

The Tracker LOG file (FAH GPU Tracker V2\) looks like this:


[00:36:07] FAH GPU Tracker v3.21 Startup
[00:36:07] --------------------------------------------------
[00:36:07] Detected OS: Windows 7
[00:36:08] Detected CPU: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz
[00:36:08] 4 Cores, 8 Threads, 2,668 MHz
[00:36:11] Applied new config settings
[00:36:17] Starting GPU 0
[00:36:17] Started new WU on GPU0
[00:36:35] Stopped GPU 0
[00:39:01] Starting GPU 0
[00:39:01] Started new WU on GPU0
[00:39:36] GPU 0 closed unexpectedly, restarting
[00:39:36] Starting GPU 0
[00:39:36] Resumed existing WU on GPU0
[00:39:56] GPU 0 closed unexpectedly, restarting
[00:39:56] Starting GPU 0
[00:39:56] Resumed existing WU on GPU0
[00:40:17] GPU 0 closed unexpectedly, restarting
[00:40:17] Starting GPU 0
[00:40:17] Resumed existing WU on GPU0
[00:40:38] GPU 0 closed unexpectedly, restarting
[00:40:38] Starting GPU 0
[00:40:38] Resumed existing WU on GPU0
[00:40:58] GPU 0 closed unexpectedly, restarting
[00:40:58] Starting GPU 0
[00:40:58] Resumed existing WU on GPU0
[00:41:04] Stopped GPU 0
[00:42:36] Applied new config settings
[00:45:34] Starting GPU 0
[00:45:34] Resumed existing WU on GPU0
[00:45:38] Stopped GPU 0
[00:46:19] Starting GPU 0
[00:46:19] Resumed existing WU on GPU0
[00:46:38] GPU 0 closed unexpectedly, restarting
[00:46:39] Starting GPU 0
[00:46:39] Resumed existing WU on GPU0
[00:46:59] GPU 0 closed unexpectedly, restarting
[00:46:59] Starting GPU 0
[00:46:59] Resumed existing WU on GPU0
[00:47:06] Stopped GPU 0
[00:50:53] Starting GPU 0
[00:50:53] Resumed existing WU on GPU0
[00:51:12] GPU 0 closed unexpectedly, restarting
[00:51:12] Starting GPU 0
[00:51:12] Resumed existing WU on GPU0
[00:51:32] GPU 0 closed unexpectedly, restarting
[00:51:32] Starting GPU 0
[00:51:32] Resumed existing WU on GPU0
[00:51:45] Stopped GPU 0

Hopefully this will help you out and you can see what I am doing wrong. I will stop posting on the [H] website forum about this, and only through this post. Oh yeah, this is a shot of my Tracker settings:

If I remember correctly the R700 was for the 4XX0 and the R800 which is the 5000 series isnt even folding right now. GPU3 is only running on the nvidia 400 series. That is why you are getting the failed WU on GPU0. The R700 cannot run the GPU3 client.

CadBane wrote:If I remember correctly the R700 was for the 4XX0 and the R800 which is the 5000 series isnt even folding right now. GPU3 is only running on the nvidia 400 series. That is why you are getting the failed WU on GPU0. The R700 cannot run the GPU3 client.

This isn't correct. There is no -forcegpu ati_r800 flag. In order to fold on ATI 5xxx you need to use forcegpu ati_r700. It is possible to fold on 5xxx cards right now using this method.

jedi95 is, of course, correct. Like I said, I can get the normal client running perfectly for my 5870. Still haven't been able to get it running through the Tracker, though. Ah well, I will keep messing around with it as time allows. If I get it running right, I will let you know what I did to do so.

sigvoror wrote:jedi95 is, of course, correct. Like I said, I can get the normal client running perfectly for my 5870. Still haven't been able to get it running through the Tracker, though. Ah well, I will keep messing around with it as time allows. If I get it running right, I will let you know what I did to do so.

In the near future I will be able to properly test this type of configuration, which should enable me to fix whatever is causing this problem. I should have what I need by the end of the week.