MultiGPU folding problem

Denithor

Diamond Member
Apr 11, 2004
6,298
23
81
Have two GPU clients running on the folding rig in my sig.

The first client works perfectly, turns in WU, downloads another, keeps on cranking.

The second, however, finishes a WU, turns it in, then just seems to start folding the same WU over again (it shows it working, progress being made, but no points are accumulated and Fahmon shows zero ppd after the initial WU is turned in. If I manually stop & restart the client it will download a new WU & crunch through it, but then the cycle starts again.

Any suggestions?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,426
16,291
136
Sounds like they are using the same client or macine id. Stop the second. Delete client.cfg, queue.dat, work folder, unitinfo.*. Tehn run again and config, and make sure that you specify "advanced options='y', and make sure you select a machineid that number one gpu is not using (it in the top of the log, right after you start it, somewhere on fahlog.txt) Default is 2, so it number one is machineid=2, then select 3 for the second folder gpu. Then restart.

I have 7 boxes configured this way, and it seems to work for me.
 

Denithor

Diamond Member
Apr 11, 2004
6,298
23
81
Thanks, I'll give that a try this afternoon when I get home. You can have the same machine ID as long as they are on different machines, correct? IE - machineid=2 & machineid=3 on first rig and machineid=2 on second rig (look at my sig for why).
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,426
16,291
136
right, you can have up to 16 machine id's on the same physical server.
 

GLeeM

Elite Member
Apr 2, 2004
7,199
128
106
Originally posted by: Markfw900
right, you can have up to 16 machine id's on the same physical server.

Yup, each "client" has a unique ID. It is made up of "MachineID" plus "UserID". The UserID is downloaded from Stanford and stored in the Registry, it is different for each computer.
 

Denithor

Diamond Member
Apr 11, 2004
6,298
23
81
Originally posted by: Markfw900
Sounds like they are using the same client or macine id. Stop the second. Delete client.cfg, queue.dat, work folder, unitinfo.*. Tehn run again and config, and make sure that you specify "advanced options='y', and make sure you select a machineid that number one gpu is not using (it in the top of the log, right after you start it, somewhere on fahlog.txt) Default is 2, so it number one is machineid=2, then select 3 for the second folder gpu. Then restart.

I have 7 boxes configured this way, and it seems to work for me.

Everything else done & it's working again, I will keep an eye on it to see if it restarts correctly. However, where do I set the advanced options=y ? I don't see that anywhere.
 

Insidious

Diamond Member
Oct 25, 2001
7,649
0
0
I think they mean as you are answering the configuration questions, one of them will ask you if you want to modify the advanced options. When you answer Y, it will ask some more questions.

-Sid
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,426
16,291
136
Here is my last config section text:
-------------------------------------------------------------------------
Launch directory: C:\folding_gpu2
Executable: C:\folding_gpu2\Folding@home-Win32-GPU.exe
Arguments: -gpu 1 -forcegpu nvidia_g80

[02:20:13] Configuring Folding@Home...

User name [Anonymous]? Mark_F_Williams
Team Number [0]? 198
Passkey []?
Ask before fetching/sending work (no/yes) [no]?
Use proxy (yes/no) [no]?
Acceptable size of work assignment and work result packets (bigger units
may have large memory demands) -- 'small' is <5MB, 'normal' is <10MB, and
'big' is >10MB (small/normal/big) [normal]? big
Change advanced options (yes/no) [no]? yes
Core Priority (idle/low) [idle]?
CPU usage requested (5-100) [100]?
Disable highly optimized assembly code (no/yes) [no]?
Pause if battery power is being used (useful for laptops) (no/yes) [no]?
Interval, in minutes, between checkpoints (3-30) [15]?
Memory, in MB, to indicate (3071 available) [3071]?
Set -advmethods flag always, requesting new advanced
scientific cores and/or work units if available (no/yes) [no]?
Ignore any deadline information (mainly useful if
system clock frequently has errors) (no/yes) [no]?
Machine ID (1-16) [2]? 3
Launch automatically, install as a service in this directory (yes/no) [no]?
The following options require you to restart the client before they take effect
Disable CPU affinity lock (no/yes) [no]?
Additional client parameters []?
IP address to bind core to (for viewer) []?

[02:21:03] - Ask before connecting: No
[02:21:03] - User name: Mark_F_Williams (Team 198)
[02:21:03] - User ID: C8415D27794DB31
[02:21:03] - Machine ID: 3
[02:21:03]

---------------------------------------------------------------------------
Note where I answered yes to change advanced options, then I selected 3
when asked for the machineid, since this is for the second card in the box.

Aside from username , you should answer all the questions the same as I did for your second client, and make sure you use " -gpu 1 -forcgpu nvidia_g80" at the end of your shortcut. Mine looks like the below:

C:\folding_gpu2\Folding@home-Win32-GPU.exe -gpu 1 -forcegpu nvidia_g80
 

Denithor

Diamond Member
Apr 11, 2004
6,298
23
81
Bringing this back up because it started happening again a week ago.

Since then, I've gone back through deleting the WU & stuff again. It works once and then stops again.

I've also completely uninstalled the GPU clients and reinstalled from scratch the latest copy from the download site. Again, this works for a couple of cycles but then GPU2 just stops with a completed WU and sits there (doesn't turn in, doesn't download another and obviously doesn't start crunching again).

Note I'm using the graphical interface version, not the console client, does that make any difference?

I'm not using the -forcegpu line on mine, and it should be G92 for my 8800GS/9600GSO/8800GTS right? Do you think that could make a difference?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,426
16,291
136
First, I recommend the console version. Second, I can;t even get my second client to even start without the -forcegpu nvidia_g80, use it for sure as well as the -gpu 1 flag.
 

Denithor

Diamond Member
Apr 11, 2004
6,298
23
81
Should the line be -forcegpu nvidia_g80 or -forcegpu nvidia_g92 for use with G92-based cards?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,426
16,291
136
there is no 92 hint. The g80 works fine on g92 cards
 

Denithor

Diamond Member
Apr 11, 2004
6,298
23
81
Well, I think I've gotten it working now.

Installed the console version, set it up exactly as you specified and both are running.

One question - are the two DOS/command prompt boxes going to sit on my tray whenever the clients are running?