Our team got passed by HPCS

Uber_Tiny

Uber_Tiny

New member
I don't understand this though.

According to the information, HPCS is getting a PPD of 37,336,429. Rage is only getting 1,205,887 PPD. The confusing part is how HPCS is only producing about half the units that Rage is. :confused:

How are they coming up with those points?
 
the vast majority of the units our team turns in are EUEs, incomplete work units.

we actually turn in around 100 complete work units per update.
 
Ryoko said:
the vast majority of the units our team turns in are EUEs, incomplete work units.

we actually turn in around 100 complete work units per update.
Click to expand...

How can we tell if the WU's returned are EUE's? And what does EUE stand for? :-)
 
Ryoko said:
the vast majority of the units our team turns in are EUEs, incomplete work units.

we actually turn in around 100 complete work units per update.
Click to expand...

I fold big stuff on my GPUs. As far as I am am able to tell. Everything gets completed.
 
I was more thinking along the lines of trying to increase my points. I am not convinced since I reinstalled a couple of weeks ago that my GPUs are being used to their best ability or rather getting small jobs. Hence I was quizzing how you determine whether you are getting big or not.

Or maybe a better question would be... How do you force the GPU to get high point workloads ?
V7.1.43 is the version I downloaded to use.
 
RuinatioN said:
I was more thinking along the lines of trying to increase my points. I am not convinced since I reinstalled a couple of weeks ago that my GPUs are being used to their best ability or rather getting small jobs. Hence I was quizzing how you determine whether you are getting big or not.

Or maybe a better question would be... How do you force the GPU to get high point workloads ?
V7.1.43 is the version I downloaded to use.
Click to expand...

ah, now that i can't answer... what kind of gpus are you using? perhaps someone who knows more about gpu folding can help...
 
AMD 5870's (Cypress chips).

I just get a strange feeling that before the reinstall and client update they was getting more points.

You could give some tips or stuff to check to make sure I've set the CPU correctly. Not overclocked it..... yet.
 
RuinatioN. What graphics card do you have? If you have a graphics card AMD HD 5xxx or 6xxx series:

Change your V7 configuration in Fahcontrol.exe as follows: Under Configure + Slots + (choose GPU) + Edit
Scroll down to Extra slot options (expert only) and click Add
Under Name: enter client-type and in Value:, enter advanced
Click OK twice and Save

Can you please post the "System" section from your log so that we can see the system configuration?
Example:

*********************** Log Started 2012-01-16T08:11:18 ************************
08:11:18:************************* Folding@home Client *************************
08:11:18: Website: http://folding.stanford.edu/
08:11:18: Copyright: (c) 2009-2012 Stanford University
08:11:18: Author: Joseph Coffland <[email protected]>
08:11:18: Args: --lifeline 4880 --command-port=36330
08:11:18: Config: C:/Users/Student/AppData/Roaming/FAHClient/config.xml
08:11:18:******************************** Build ********************************
08:11:18: Version: 7.1.43
08:11:18: Date: Jan 2 2012
08:11:18: Time: 12:33:05
08:11:18: SVN Rev: 3223
08:11:18: Branch: fah/trunk/client
08:11:18: Compiler: Intel(R) C++ MSVC 1500 mode 1200
08:11:18: Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
08:11:18: /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT
08:11:18: Platform: win32 XP
08:11:18: Bits: 32
08:11:18: Mode: Release
08:11:18:******************************* System ********************************
08:11:18: CPU: Intel(R) Core(TM) i7-2630QM CPU @ 2.00GHz
08:11:18: CPU ID: GenuineIntel Family 6 Model 42 Stepping 7
08:11:18: CPUs: 8
08:11:18: Memory: 7.95GiB
08:11:18: Free Memory: 5.97GiB
08:11:18: Threads: WINDOWS_THREADS
08:11:18: On Battery: false
08:11:18: UTC offset: -8
08:11:18: PID: 5728
08:11:18: CWD: C:/Users/Student/AppData/Roaming/FAHClient
08:11:18: OS: Windows 7 Home Premium
08:11:18: OS Arch: AMD64
08:11:18: GPUs: 1
08:11:18: GPU 0: ATI:4 Whistler XT [AMD Radeon HD 6770M Series]
08:11:18: CUDA: Not detected
08:11:18:Win32 Service: false

<config>
<!-- FahCore Control -->
<checkpoint v='10'/>

<!-- Folding Slot Configuration -->
<gpu v='true'/>

<!-- Logging -->
<verbosity v='5'/>

<!-- Network -->
<proxy v=':8080'/>

<!-- User Information -->
<passkey v='********************************'/>
<team v='64'/>
<user v='xxxxx'/>

<!-- Folding Slots -->
<slot id='0' type='SMP'>
<client-type v='bigadv'/>
<cpus v='-1'/>
<max-packet-size v='big'/>
</slot>
<slot id='1' type='GPU'>
<client-type v='advanced'/>
<gpu-index v='0'/>
</slot>
</config>


Un saludo.

 
Thanks for the tip. I have added the setting... Will see what happens from here.
Below log is after the addition of "advanced" to the GPU.

*********************** Log Started 2012-01-24T02:00:20 ************************
02:00:20:************************* Folding@home Client *************************
02:00:20: Website: http://folding.stanford.edu/
02:00:20: Copyright: (c) 2009-2012 Stanford University
02:00:20: Author: Joseph Coffland <[email protected]>
02:00:20: Args: --lifeline 11276 --command-port=36330
02:00:20: Config: C:/Users/MyUser/AppData/Roaming/FAHClient/config.xml
02:00:20:******************************** Build ********************************
02:00:20: Version: 7.1.43
02:00:20: Date: Jan 2 2012
02:00:20: Time: 12:33:05
02:00:20: SVN Rev: 3223
02:00:20: Branch: fah/trunk/client
02:00:20: Compiler: Intel(R) C++ MSVC 1500 mode 1200
02:00:20: Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
02:00:20: /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT
02:00:20: Platform: win32 XP
02:00:20: Bits: 32
02:00:20: Mode: Release
02:00:20:******************************* System ********************************
02:00:20: CPU: Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz
02:00:20: CPU ID: GenuineIntel Family 6 Model 42 Stepping 7
02:00:20: CPUs: 8
02:00:20: Memory: 15.98GiB
02:00:20: Free Memory: 12.97GiB
02:00:20: Threads: WINDOWS_THREADS
02:00:20: On Battery: false
02:00:20: UTC offset: 0
02:00:20: PID: 18292
02:00:20: CWD: C:/Users/MyUser/AppData/Roaming/FAHClient
02:00:20: OS: Windows 7 Professional
02:00:20: OS Arch: AMD64
02:00:20: GPUs: 2
02:00:20: GPU 0: ATI:4 Radeon HD 5870 (Cypress)
02:00:20: GPU 1: ATI:4 Radeon HD 5870 (Cypress)
02:00:20: CUDA: Not detected
02:00:20:Win32 Service: false
02:00:20:***********************************************************************
02:00:20:<config>
02:00:20: <!-- Folding Slot Configuration -->
02:00:20: <gpu v='true'/>
02:00:20:
02:00:20: <!-- Network -->
02:00:20: <proxy v=':8080'/>
02:00:20:
02:00:20: <!-- User Information -->
02:00:20: <passkey v='********************************'/>
02:00:20: <team v='64'/>
02:00:20: <user v='RuinatioN'/>
02:00:20:
02:00:20: <!-- Folding Slots -->
02:00:20: <slot id='0' type='GPU'>
02:00:20: <client-type v='advanced'/>
02:00:20: </slot>
02:00:20: <slot id='1' type='GPU'>
02:00:20: <client-type v='advanced'/>
02:00:20: <cuda-index v='0'/>
02:00:20: </slot>
02:00:20: <slot id='2' type='SMP'>
02:00:20: <cpus v='-1'/>
02:00:20: </slot>
02:00:20:</config>
 
Last edited:
One thing I have read and there doesnt seem to be any conclusive or maybe consistent views is that the CPU should not be maxed out to allow CPU time to put data to the GPUs. That said I have left mine at -1 which defaults to using 100% (4c/8t) of the i7-2600k. The suggestion would mean dropping that to 3c/6t. though.
 
RuinatioN said:
One thing I have read and there doesnt seem to be any conclusive or maybe consistent views is that the CPU should not be maxed out to allow CPU time to put data to the GPUs. That said I have left mine at -1 which defaults to using 100% (4c/8t) of the i7-2600k. The suggestion would mean dropping that to 3c/6t. though.
Click to expand...

gpu folding is less efficient on amd video cards. each gpu client uses just about all the cpu power it can get (one core per client) so you would probably be better off setting your smp client to -smp6. you may be able to get ~10k ppd per card, and ~10k ppd for the smp client (if your 2600k is stock).
 
Was expecting closer to the 30k mark than the 20k I seem to be getting. I will play about with configuration and see if I can find a good balance of what will work :)
 
Yes I have. It has improved effeciency slightly but not drastically. Feels like the WU are small and not much in terms of points/credit. That could just be expecting more.
 
I applied the tweak above to my 6870. I'll post back if I notice any changes. Any other tweaks I can try? I'm using the 7.1.43 client as well. System specs are up to date at left.
 
You must log in or register to reply here.
Back
Top