New AstroPulse for GPU ( ATi & NV) released (r1316)

Yeah, at a certain point you get diminishing returns for fine-tuning parameters. I didn't bother pushing unroll too hard, I only used 11 because the HD 6950 has 22 compute units.

For the ffa parameters, 16384/8192 crashes (presumably too much memory required or whatever), and didn't bother trying to push further than 8192/4096.

For the record, the HD 6950 I have only has 1 GiB memory (it cannot use the HD 6970 firmware) and I've set the clocks to 850/1350 (very mild overclock, I know), known to be stable for both MB, AP (no computation errors for, well, as long as I've had the card) and playing games. (Any higher causes freezing/lock-ups in games - I find games break before GPGPU get errors.)
____________
Soli Deo Gloria

I also tested the GTX 260 Core 216 today, just to see if there was any difference from Fermi/Kepler - I had a report typed up but then lost it all, but suffice it to say that the differences between r1316 and r1339 followed the pattern of the rest of my tests. Skipping the FFA pre-compute didn't change CPU time, but overall run-time increased by about 8%.

I used -ffa_block 2048 -ffa_block_fetch 1024. -ffa_block 8192 -ffa_block_fetch 4096 was very similar, but oddly enough, -ffa_block 4096 -ffa_block_fetch 2048 was consistently very slightly slower, but not by significant margin at all.
____________
Soli Deo Gloria

I really wonder if my file is correct app_info :

<app_info>
<file_info>
<name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_intelx86</platform>
<plan_class>cuda_fermi</plan_class>
<cmdline>-instances_per_device 1 -unroll 10 -ffa_block 6144 -ffa_block_fetch 1536 -sbs 256</cmdline>
<coproc>
<type>CUDA</type>
<count>0.51</count>
</coproc>
<file_ref>
<file_name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_x86_64</platform>
<plan_class>cuda_fermi</plan_class>
<cmdline>-instances_per_device 1 -unroll 10 -ffa_block 6144 -ffa_block_fetch 1536 -sbs 256</cmdline>
<coproc>
<type>CUDA</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8b2_win_x64_SSE41.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_intelx86</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse_v6</name>
</app>
<file_info>
<name>ap_6.01r557_SSE2_331_AVX.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<platform>windows_intelx86</platform>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse_v505</name>
</app>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<platform>windows_intelx86</platform>
<cmdline>-v505</cmdline>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<platform>windows_x86_64</platform>
<cmdline>-v505</cmdline>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>Lunatics_x41g_win32_cuda32.exe</name>
<executable/>
</file_info>
<file_info>
<name>cudart32_32_16.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft32_32_16.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda_fermi</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>609</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda23</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
<platform>windows_x86_64</platform>
<plan_class>cuda_fermi</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>609</version_num>
<platform>windows_x86_64</platform>
<plan_class>cuda23</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_x86_64</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
</app_info>

____________


Alliance Francophone

Hi

-instances_per_device 1 has no effect, its default.
Same as -sbs switch.
-sbs switch not needed for now, it only has slightly effect.


<count> should be set to 1 since you only run 1 instance on all GPU apps.
____________

Thanx Fataldog and Wedge009,
Very clear test using different cmd-line parameters.

About your app_info.xml question:

You've got 1: <coproc>
<type>CUDA</type>
<count>0.51</count>
</coproc>

and 1

<coproc>
<type>CUDA</type>
<count>0.5</count>
</coproc>

Why not use the app_info.xml file present in the latest installer
and modify only the AstroPulse part, changing rev.555 to rev.1316.

Which can be found here .
____________


Knight Who Says Ni N!, OUT numbered.................

Hey guys, reading this thread sounds like a foreign language to me, but from what I gather, are we supposed to drop the binary into the setiathome.berkeley.edu directory, and in the app_info change the line:

<name>ap_6.01r557_SSE2_331_AVX.exe</name>

to

<name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</name>

in the 2 places it is referenced?

Hey guys, reading this thread sounds like a foreign language to me, but from what I gather, are we supposed to drop the binary into the setiathome.berkeley.edu directory, and in the app_info change the line:

<name>ap_6.01r557_SSE2_331_AVX.exe</name>

to

<name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</name>

in the 2 places it is referenced?

I am just humbly trying to learn from you guys, but if I am wrong and there is actually a lot more to it than just that, I can wait for the next installer release and just do it the simple way without modifying the app info or anything?
btw, thanks for your hard work on these improvements.

Hello,

So at the beginning of 0.51 else is good?

By cons whenever I boinc spears, the application:

AP6_win_x86_SSE2_OpenCL_NV_r1316.exe is deleted?! why

I really wonder if my file is correct app_info :

<app_info>
<file_info>
<name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_intelx86</platform>
<plan_class>cuda_fermi</plan_class>
<cmdline>-instances_per_device 1 -unroll 10 -ffa_block 6144 -ffa_block_fetch 1536 -sbs 256</cmdline>
<coproc>
<type>CUDA</type>
<count>0.51</count>
</coproc>
<file_ref>
<file_name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_x86_64</platform>
<plan_class>cuda_fermi</plan_class>
<cmdline>-instances_per_device 1 -unroll 10 -ffa_block 6144 -ffa_block_fetch 1536 -sbs 256</cmdline>
<coproc>
<type>CUDA</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>AP6_win_x86_SSE2_OpenCL_NV_r1316.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8b2_win_x64_SSE41.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_intelx86</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse_v6</name>
</app>
<file_info>
<name>ap_6.01r557_SSE2_331_AVX.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<platform>windows_intelx86</platform>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>601</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse_v505</name>
</app>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<platform>windows_intelx86</platform>
<cmdline>-v505</cmdline>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<platform>windows_x86_64</platform>
<cmdline>-v505</cmdline>
<file_ref>
<file_name>ap_6.01r557_SSE2_331_AVX.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>Lunatics_x41g_win32_cuda32.exe</name>
<executable/>
</file_info>
<file_info>
<name>cudart32_32_16.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft32_32_16.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda_fermi</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>609</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda23</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
<platform>windows_x86_64</platform>
<plan_class>cuda_fermi</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>609</version_num>
<platform>windows_x86_64</platform>
<plan_class>cuda23</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_x86_64</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
</app_info>

`You still have 2 AstroPulse entries, 1 for CPU and
1 for GPU!!!

Can't have both, so you'll have to remove the CPU entry from
your app_info.xml file!

That's probably why the NV_r1316 executable is removed by BOINC.
IIRC, there's a minimum version of BOINC (7.0.xx?)?

Also running 1 instance_per_device, no 0.5 or 0.51 CUDA device.
Just 1, which is default, so nothing has to be set.
____________


Knight Who Says Ni N!, OUT numbered.................

I really wonder if my file is correct app_info :

`You still have 2 AstroPulse entries, 1 for CPU and
1 for GPU!!!

Can't have both, so you'll have to remove the CPU entry from
your app_info.xml file!

That's probably why the NV_r1316 executable is removed by BOINC.
IIRC, there's a minimum version of BOINC (7.0.xx?)?

Also running 1 instance_per_device, no 0.5 or 0.51 CUDA device.
Just 1, which is default, so nothing has to be set.

`You still have 2 AstroPulse entries, 1 for CPU and
1 for GPU!!!

Can't have both, so you'll have to remove the CPU entry from
your app_info.xml file!

That's probably why the NV_r1316 executable is removed by BOINC.
IIRC, there's a minimum version of BOINC (7.0.xx?)?

Also running 1 instance_per_device, no 0.5 or 0.51 CUDA device.
Just 1, which is default, so nothing has to be set.

I do not receive despites Astropulse unit below? units but enhanced 6.10


____________


Alliance Francophone

knowing that I have a GTX 470 ZOTAC AMP edition, I see what parameters change in app_info.

I use boinc 7.0.28
____________


Alliance Francophone

Fred, there's no reason why you can't run CPU Astropulse and GPU Astropulse at the same time, Hell, i've run CPU Astropulse, ATI Astropulse and Nvidia Astropulse all at the same time.

There's no minimum Boinc version for running NV_r1316 eithier. (Subject to Boinc being able to detect CAL or Cuda devices, so that probably means Boinc 6.4.x should work)

Claggy

I noticed today 42 Exit status 197 (0xc5) EXIT_TIME_LIMIT_EXCEEDED!?

Will AV scan my computer tomorrow, SmartDoctor had turned the GPUs from
850MHz to 500MHz.? And wasn't fully started?
And a cmd screen was open and couldn't been closed.
This all 'happened' out of the blue?!
Had to change start-menu to take care about that.
____________


Knight Who Says Ni N!, OUT numbered.................

Will this be added to the lunatics app soon?

You receive units in astro nvidia GPU? For me despites fair to have that checked GPU and astro v6 in my preferences ... still nothing

Edit: made ​​with google translation (French-English)
____________


Alliance Francophone