app_info for AP500, AP503, MB603 and MB608

Message boards : Number crunching : app_info for AP500, AP503, MB603 and MB608
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 4 · 5 · 6 · 7 · 8 · 9 · 10 . . . 12 · Next

AuthorMessage
Profile MadMaC
Volunteer tester
Avatar

Send message
Joined: 4 Apr 01
Posts: 201
Credit: 47,158,217
RAC: 0
United Kingdom
Message 887341 - Posted: 22 Apr 2009, 17:41:12 UTC - in response to Message 887248.  

Ok, I have deleted my cc_config file and restarted boinc and Im crunching 2 AP WU, but no CUDA units, my GPU is idle?? What am I missing, strangley I have noticed that I no longer have an app_info file (yet Im still using R112 for the AP units) Is this a result of the move to 6.6.20?
ID: 887341 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 887591 - Posted: 23 Apr 2009, 14:10:14 UTC - in response to Message 886494.  

It may take awhile before I get to another Astropulse. I got swamped with a bunch of 6.03s while ttc was down. I'd like to clear some of them out before I do another AP. As I said, the Astropulse takes me 60+ hours to complete.


Well, I finished one. The time to completion jumped back up for both the APs and the Cuda work. I decided to make a small change in the FLOPS for my Cuda work. I re-figured it for 6GFlops instead of 5 and the TTC is a lot closer to what it should be.

I also suspended the APs that weren't running so that I am only running one AP and one MB on my CPUs. this may have helped it figure out the right DCF. This may change today though, I managed to download a bunch of shorties for my Cuda this morning. :)



PROUD MEMBER OF Team Starfire World BOINC
ID: 887591 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 887592 - Posted: 23 Apr 2009, 14:13:35 UTC - in response to Message 887341.  

MadMaC,

6.6.20 should not have caused your app_info to disappear. It still needs to be in there somewhere.


PROUD MEMBER OF Team Starfire World BOINC
ID: 887592 · Report as offensive
Terror Australis
Volunteer tester

Send message
Joined: 14 Feb 04
Posts: 1817
Credit: 262,693,308
RAC: 44
Australia
Message 887928 - Posted: 24 Apr 2009, 15:43:21 UTC

If I'm running 2 video cards of different speeds. Which one do I use to calculate the Gflops as per Marks instructions ? The fast one or the slow one ?

Brodo
ID: 887928 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 887932 - Posted: 24 Apr 2009, 15:56:09 UTC - in response to Message 887928.  

If I'm running 2 video cards of different speeds. Which one do I use to calculate the Gflops as per Marks instructions ? The fast one or the slow one ?

Brodo

At message 880869, a long way down (or up - depending on you preferred order) this same thread, I suggested that using the speed of your faster card would give the smoothest DCF / cache fill experience.

But since then, I've realised that this will fill your cache as if you had two fast cards. Be a little cautious with the cache setting if you follow that advice.
ID: 887932 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 887935 - Posted: 24 Apr 2009, 15:59:54 UTC - in response to Message 887932.  

I believe someone also suggested you average the two and figure from that.


PROUD MEMBER OF Team Starfire World BOINC
ID: 887935 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 887938 - Posted: 24 Apr 2009, 16:05:41 UTC - in response to Message 887935.  

I believe someone also suggested you average the two and figure from that.

Did they give any reasoning? My approach, by targeting DCF, is designed never to miss a CPU (AP) deadline in the four-way environment described in this thread title - at the expense, as I've just posted, of possibly missing CUDA deadlines if you over-do the cache setting.
ID: 887938 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 887951 - Posted: 24 Apr 2009, 16:39:58 UTC - in response to Message 887938.  

I can't find the post where I read that now. I think it was just to keep the figure low enough so that BM wouldn't get the idea you had two really fast cards and load you up beyond what you could finish.


PROUD MEMBER OF Team Starfire World BOINC
ID: 887951 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 887953 - Posted: 24 Apr 2009, 16:45:36 UTC - in response to Message 887951.  

I can't find the post where I read that now. I think it was just to keep the figure low enough so that BM wouldn't get the idea you had two really fast cards and load you up beyond what you could finish.


Or having it too high, and WU's taking too long on the slow card and getting maximum time exceeded,
But i've only seen that on Astropulse so far.

Claggy
ID: 887953 · Report as offensive
Profile -=SuperG=-
Avatar

Send message
Joined: 3 Apr 99
Posts: 63
Credit: 89,161,651
RAC: 23
Canada
Message 887954 - Posted: 24 Apr 2009, 16:45:42 UTC - in response to Message 887951.  

I can't find the post where I read that now. I think it was just to keep the figure low enough so that BM wouldn't get the idea you had two really fast cards and load you up beyond what you could finish.


I remember reading this as well.

Before this thread, I had to read lots of threads to figure out how to set up the app_info properly. I do not remember where I read it either... :(

Regards,
G
Boinc Wiki




"Great spirits have always encountered violent opposition from mediocre minds." -Albert Einstein
ID: 887954 · Report as offensive
Terror Australis
Volunteer tester

Send message
Joined: 14 Feb 04
Posts: 1817
Credit: 262,693,308
RAC: 44
Australia
Message 887961 - Posted: 24 Apr 2009, 17:00:22 UTC - in response to Message 887932.  

If I'm running 2 video cards of different speeds. Which one do I use to calculate the Gflops as per Marks instructions ? The fast one or the slow one ?

Brodo

At message 880869, a long way down (or up - depending on you preferred order) this same thread, I suggested that using the speed of your faster card would give the smoothest DCF / cache fill experience.

But since then, I've realised that this will fill your cache as if you had two fast cards. Be a little cautious with the cache setting if you follow that advice.

Thanks Richard - I knew I'd seen the answer somewhere in the dim dark past but couldn't find it again when I needed it.

Am now up and running using the "No AP" version of your app_info file. I only run a 1.5 day cache so time outs are not a problem

Strange, AK_v8 even picked up the 2 units that were half crunched by the stock 6.03 app and carried on with them. I thought it would shred them and start 2 new ones

Brodo
ID: 887961 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 887966 - Posted: 24 Apr 2009, 17:08:20 UTC - in response to Message 887961.  

Strange, AK_v8 even picked up the 2 units that were half crunched by the stock 6.03 app and carried on with them. I thought it would shred them and start 2 new ones

Brodo

No surprise, Lunatics apps are polite and well-mannered like that. :-)

[Well, they are when we get them, and the corresponding app_info, right]
ID: 887966 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65709
Credit: 55,293,173
RAC: 49
United States
Message 888075 - Posted: 24 Apr 2009, 22:48:40 UTC - in response to Message 876502.  
Last modified: 24 Apr 2009, 22:58:07 UTC

I looked in My client_state.xml file and <p_flops> doesn't exist there, I have a GTX295 and I run MB on both the Intel Q9300 Quad Core cpu(3.51GHz) and on the gpus(the GTX295), all this is running under XP x64. Is there anywhere else in Boinc 6.6.20 where It could be located at?

I do have this in client state of course:

</xml_signature>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>1.000000</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>3512976066.004993</flops>
<file_ref>
<file_name>AK_v8_win_x64_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>0.010000</avg_ncpus>
<max_ncpus>0.010000</max_ncpus>
<flops>3512976066.004993</flops>
<plan_class>cuda</plan_class>
<file_ref>
<file_name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</file_name>


Edit: I set My flops level at <flops>12000000000</flops> for the time being as a placeholder, It's better than nothing I suspect.

<app_info>

<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8_win_x64_SSSE3x.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>AK_v8_win_x64_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>

<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</name>
<executable/>
</file_info>
<file_info>
<name>cudart.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft.dll</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
<app_version>

<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<plan_class>cuda</plan_class>
<avg_ncpus>0.010000</avg_ncpus>
<max_ncpus>0.010000</max_ncpus>
<flops>12000000000</flops>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft.dll</file_name>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>

</app_info>
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 888075 · Report as offensive
Fred W
Volunteer tester

Send message
Joined: 13 Jun 99
Posts: 2524
Credit: 11,954,210
RAC: 0
United Kingdom
Message 888077 - Posted: 24 Apr 2009, 22:56:30 UTC - in response to Message 888075.  

I looked in My client_state.xml file and <p_flops> doesn't exist there, I have a GTX295 and I run MB on both the Intel Q9300 Quad Core cpu(3.51GHz) and on the gpus(the GTX295). Is there anywhere else in Boinc 6.6.20 where It could be located at?

I do have this in client state of course:

</xml_signature>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>1.000000</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>3512976066.004993</flops>
<file_ref>
<file_name>AK_v8_win_x64_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>0.010000</avg_ncpus>
<max_ncpus>0.010000</max_ncpus>
<flops>3512976066.004993</flops>
<plan_class>cuda</plan_class>
<file_ref>
<file_name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</file_name>

I'm running my Q9450 (o/c) + GTX295 with the CUDA flops set to 21200000000 and the AK-V8 flops set to 6351283655.

Works for me.

F.
ID: 888077 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 888078 - Posted: 24 Apr 2009, 22:57:46 UTC - in response to Message 888075.  

Look for P_FPOPS


PROUD MEMBER OF Team Starfire World BOINC
ID: 888078 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65709
Credit: 55,293,173
RAC: 49
United States
Message 888080 - Posted: 24 Apr 2009, 23:02:26 UTC - in response to Message 888078.  
Last modified: 24 Apr 2009, 23:08:37 UTC

Look for P_FPOPS

Ok I found It and It's 3512976066.004993 so now what? multiply It by 0.2?

Edit: Ok I'm trying this out: <flops>00702595213</flops>
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 888080 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 888081 - Posted: 24 Apr 2009, 23:09:56 UTC - in response to Message 888080.  

Look for P_FPOPS

Ok I found It and It's 3512976066.004993 so now what? multiply It by 0.2?


NO, that's CPU flops, get GPU flops from your Boinc log on start-up,
It'll be something like 87 (Est.Gflops) times it by 1000000000 then by 0.2

Claggy
ID: 888081 · Report as offensive
Fred W
Volunteer tester

Send message
Joined: 13 Jun 99
Posts: 2524
Credit: 11,954,210
RAC: 0
United Kingdom
Message 888083 - Posted: 24 Apr 2009, 23:13:27 UTC - in response to Message 888080.  

Look for P_FPOPS

Ok I found It and It's 3512976066.004993 so now what? multiply It by 0.2?

Edit: Ok I'm trying this out: <flops>00702595213</flops>

You need the P_FPops for CPU: That should be multiplied by 1.75.
Your GPU will be the same as mine - Est 106 GFlops - so you should use 21200000000 for that.

F.
ID: 888083 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65709
Credit: 55,293,173
RAC: 49
United States
Message 888085 - Posted: 24 Apr 2009, 23:22:09 UTC - in response to Message 888083.  

Look for P_FPOPS

Ok I found It and It's 3512976066.004993 so now what? multiply It by 0.2?

Edit: Ok I'm trying this out: <flops>00702595213</flops>

You need the P_FPops for CPU: That should be multiplied by 1.75.
Your GPU will be the same as mine - Est 106 GFlops - so you should use 21200000000 for that.

F.

Ok now where do I put the flops for the cpu in the app info file?

How's this?

And the flops is the fpops number after being multiplied by 1.75.

<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8_win_x64_SSSE3x.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<flops>6147708115</flops>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>AK_v8_win_x64_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 888085 · Report as offensive
Fred W
Volunteer tester

Send message
Joined: 13 Jun 99
Posts: 2524
Credit: 11,954,210
RAC: 0
United Kingdom
Message 888088 - Posted: 24 Apr 2009, 23:30:48 UTC - in response to Message 888085.  
Last modified: 24 Apr 2009, 23:31:01 UTC

Here's a snip of mine:

     <app_version>
        <app_name>setiathome_enhanced</app_name>
        <version_num>603</version_num>
        <platform>windows_x86_64</platform>
        <avg_ncpus>1.000000</avg_ncpus>
        <max_ncpus>1.000000</max_ncpus>
        <flops>6351283655</flops>
        <file_ref>
            <file_name>AK_v8_win_x64_SSSE3x.exe</file_name>
            <main_program/>
        </file_ref>
    </app_version>  
    <app_version>
        <app_name>setiathome_enhanced</app_name>
        <version_num>608</version_num>
        <platform>windows_x86_64</platform>
        <avg_ncpus>0.150000</avg_ncpus>
        <max_ncpus>0.150000</max_ncpus>
        <flops>21200000000</flops>
        <plan_class>cuda</plan_class>
        <file_ref>
            <file_name>setiathome_6.08_windows_intelx86__cuda.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>cudart.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>cufft.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>libfftw3f-3-1-1a_upx.dll</file_name>
        </file_ref>
        <coproc>
            <type>CUDA</type>
            <count>1</count>
        </coproc>
    </app_version>


F.
ID: 888088 · Report as offensive
Previous · 1 . . . 4 · 5 · 6 · 7 · 8 · 9 · 10 . . . 12 · Next

Message boards : Number crunching : app_info for AP500, AP503, MB603 and MB608


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.