astropulse options

[Jaehaerys Targaryen]

Merci d'avance :

AstroPulse OpenCL application currently available in 3 editions: for AMD/ATi, nVidia and Intel GPUs.
It's intended to process SETI@home AstroPulse v6 tasks.

Build revision:1761
Date of revision commit: 2013/01/30 00:39:00

Available command line switches:

-v N :sets level of verbosity of app. N - integer number.

-ffa_block N :sets how many FFA's different period iterations will be processed per kernel call. N should be integer even number less than 32768.

-ffa_block_fetch N: sets how many FFA's different period iterations will be processed per "fetch" kernel call (longest kernel in FFA).
    N should be positive integer number, should be divisor of ffa_block_N.

-unroll N :sets number of data chunks processed per kernel call in main application loop. N should be integer number, minimal possible value is 2.

-skip_ffa_precompute : Results in skipping FFA pre-compute kernel call. Affects performance. Experimentation required if it will increase or decrease performance on particular GPU/CPU combo.

-exit_check :Results in more often check for exit requests from BOINC. If you experience problems with long app suspend/exit use this option.
    Can decrease performance though.

-use_sleep :Results in additional Sleep() calls to yield CPU to other processes. Can affect performance. Experimentation required.

-initial_ffa_sleep N M: In PC-FFA will sleep N ms for short and M ms for large one before looking for results. Can decrease CPU usage.
    Affects performance. Experimentation required for particular CPU/GPU/GPU driver combo. N and M should be integer non-negative numbers.
    Approximation of useful values can be received via running app with -v 2 and -use_sleep switches enabled and analyzing stderr.txt log file.

-sbs N :Sets maximum single buffer size for GPU memory allocations. N should be positive integer and means bigger size in Mbytes.
    For now if other options require bigger buffer than this option allows warning will be issued but memory allocation attempt will be made.

-hp : Results in bigger priority for application process (normal priority class and above normal thread priority).
    Can be used to increase GPU load, experimentation required for particular GPU/CPU/GPU driver combo.

-cpu_lock :Results in CPUs number limitation for particular app instance. Also attempt to bind different instances to different CPU cores will be made.
    Can be used to increase performance under some specific conditions. Can decrease performance in other cases though. Experimentation required.

Here some already obsolete options are listed. They are not tested for proper operation with latest builds and are only listed for completeness:

-gpu_lock :Old way GPU lock enabled. Use -instances_per_device N switch to provide number of instances to run.

-instances_per_device N :Sets allowed number of simultaneously executed GPU app instances (shared with MultiBeam app instances).
    N - integer number of allowed instances.

-disable_slot N: Can be used to exclude N-th GPU (starting from zero) from usage.
    Not tested and obsolete feature, use BOINC abilities to exclude GPUs instead.

These 2 options used together provide BOINC-independent way to limit number of simultaneously
executing GPU apps. Each SETI OpenCL GPU application with these switches enabled will create/check global Mutexes and suspend its process
execution if limit is reached. Awaiting process will consume zero CPU/GPU and rather low amount of memory awaiting when it can continue execution.

These switches can be placed into the file called ap_cmdline.txt also.
Examples of app_info.xml entries (don't forget to replace listed executable name with real executable name and place <app_info> and </app_info>
tags at the beginning and the end of file correspondingly):

for ATi GPU:

    <app>
        <name>astropulse_v6</name>
    </app>
    <file_info>
        <name>AP6_win_x86_SSE2_OpenCL_ATI_r1761.exe</name>
        <executable/>
    </file_info>
   <file_info>
         <name>libfftw3f-3.dll</name>
   </file_info>
   <file_info>
         <name>ap_cmdline_win_x86_SSE2_OpenCL_ATI.txt</name>
   </file_info>
    <app_version>
        <app_name>astropulse_v6</app_name>
        <version_num>604</version_num>
        <platform>windows_intelx86</platform>
        <avg_ncpus>0.04</avg_ncpus>
        <max_ncpus>0.2</max_ncpus>
        <plan_class>ati_opencl_100</plan_class>
        <coproc>
            <type>ATI</type>
            <count>1</count>
        </coproc>
        <file_ref>
            <file_name>AP6_win_x86_SSE2_OpenCL_ATI_r1761.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>libfftw3f-3.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>ap_cmdline_win_x86_SSE2_OpenCL_ATI.txt</file_name>
            <open_name>ap_cmdline.txt</open_name>
        </file_ref>
    </app_version>
    <app_version>
        <app_name>astropulse_v6</app_name>
        <version_num>604</version_num>
        <platform>windows_intelx86</platform>
        <avg_ncpus>0.04</avg_ncpus>
        <max_ncpus>0.2</max_ncpus>
        <plan_class>opencl_ati_100</plan_class>
        <coproc>
            <type>ATI</type>
            <count>1</count>
        </coproc>
        <file_ref>
            <file_name>AP6_win_x86_SSE2_OpenCL_ATI_r1761.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>libfftw3f-3.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>ap_cmdline_win_x86_SSE2_OpenCL_ATI.txt</file_name>
            <open_name>ap_cmdline.txt</open_name>
        </file_ref>
    </app_version>

For NV GPU:

    <app>
        <name>astropulse_v6</name>
    </app>
    <file_info>
        <name>AP6_win_x86_SSE2_OpenCL_NV_r1761.exe</name>
        <executable/>
    </file_info>
   <file_info>
         <name>libfftw3f-3.dll</name>
   </file_info>
   <file_info>
         <name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</name>
   </file_info>
    <app_version>
        <app_name>astropulse_v6</app_name>
        <version_num>604</version_num>
        <platform>windows_intelx86</platform>
        <avg_ncpus>0.04</avg_ncpus>
        <max_ncpus>0.2</max_ncpus>
        <plan_class>cuda_opencl_100</plan_class>
        <coproc>
            <type>CUDA</type>
            <count>1</count>
        </coproc>
        <file_ref>
            <file_name>AP6_win_x86_SSE2_OpenCL_NV_r1761.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>libfftw3f-3.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</file_name>
            <open_name>ap_cmdline.txt</open_name>
        </file_ref>
    </app_version>
    <app_version>
        <app_name>astropulse_v6</app_name>
        <version_num>604</version_num>
        <platform>windows_intelx86</platform>
        <avg_ncpus>0.04</avg_ncpus>
        <max_ncpus>0.2</max_ncpus>
        <plan_class>opencl_nvidia_100</plan_class>
        <coproc>
            <type>CUDA</type>
            <count>1</count>
        </coproc>
        <file_ref>
            <file_name>AP6_win_x86_SSE2_OpenCL_NV_r1761.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>libfftw3f-3.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</file_name>
            <open_name>ap_cmdline.txt</open_name>
        </file_ref>
    </app_version>

For Intel GPU (currently in beta stage):
    <app>
        <name>astropulse_v6</name>
    </app>
    <file_info>
        <name>AP6_win_x86_SSE2_OpenCL_Intel_r1761.exe</name>
        <executable/>
    </file_info>
   <file_info>
         <name>libfftw3f-3.dll</name>
   </file_info>
   <file_info>
         <name>ap_cmdline_win_x86_SSE2_OpenCL_Intel.txt</name>
   </file_info>
    <app_version>
        <app_name>astropulse_v6</app_name>
        <version_num>604</version_num>
        <platform>windows_intelx86</platform>
        <avg_ncpus>0.04</avg_ncpus>
        <max_ncpus>0.2</max_ncpus>
        <plan_class>opencl_intel_100</plan_class>
        <cmdline></cmdline>   
        <coproc>
            <type>intel_gpu</type>
            <count>1</count>
        </coproc>
        <file_ref>
            <file_name>AP6_win_x86_SSE2_OpenCL_Intel_r1761.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>libfftw3f-3.dll</file_name>
        </file_ref>
        <file_ref>
            <file_name>ap_cmdline_win_x86_SSE2_OpenCL_Intel.txt</file_name>
            <open_name>ap_cmdline.txt</open_name>
        </file_ref>
    </app_version>

Known issues:
- With 12.x Catalyst drivers GPU usage can be low if CPU fully used with another loads.
  Same applies to NV drivers past 267.xx and to Intel SDK drivers.
  If you see low GPU usage of zero blanked tasks try to free one or more CPU cores.
- For overflowed tasks found signal sequence not always match CPU version.

Best usage tips:

For best performance it is important to free 2 CPU cores running multiple instances.
Freeing at least 1 CPU core is necessity to get enough GPU usage.

command line parameters.
_______________________

High end cards (more than 12 compute units)

-unroll 12 -ffa_block 8192 -ffa_block_fetch 4096 -hp

Mid range cards (less than 12 compute units)

-unroll 10 -ffa_block 6144 -ffa_block_fetch 1536 -hp

entry level GPU (less than 6 compute units)

-unroll 4 -ffa_block 2048 -ffa_block_fetch 1024 -hp

Your mileage might vary.
-----------------------------------------------------

App instances.
______________

On high end cards HD 5850/5870, 6950/6970, 7950/7970 you can run 3 instances.

On mid range cards HD 5770, 6850/6870, 7850/7870 best performance should be running 2 instances.

If you experience screen lags reduce unroll factor and ffa_block_fetch value.

Addendum:
_________

Running multiple cards in a system requires freeing another CPU core.

[modesti]

Je passe sur les lignes de commande et les app_machinchose.xml. Le plus important, AMHA, c'est ça :

Known issues:
- With 12.x Catalyst drivers GPU usage can be low if CPU fully used with another loads.
  Same applies to NV drivers past 267.xx and to Intel SDK drivers.
  If you see low GPU usage of zero blanked tasks try to free one or more CPU cores.
- For overflowed tasks found signal sequence not always match CPU version.

Best usage tips:

For best performance it is important to free 2 CPU cores running multiple instances.
Freeing at least 1 CPU core is necessity to get enough GPU usage.

App instances.
______________

On high end cards HD 5850/5870, 6950/6970, 7950/7970 you can run 3 instances.

On mid range cards HD 5770, 6850/6870, 7850/7870 best performance should be running 2 instances.

If you experience screen lags reduce unroll factor and ffa_block_fetch value.

Addendum:
_________

Running multiple cards in a system requires freeing another CPU core.

Problèmes connus :
- Avec les pilotes Catalyst 12.x, l'utilisation du GPU peut être faible si le CPU est entièrement utilisé par d'autres charges.
  La même chose s'applique aux pilotes NV après 267.xx et aux pilotes Intel SDK.
  Si vous voyez une faible utilisation du GPU pour des tâches « zero blanked » [??], essayez de libérer un ou plusieurs coeurs CPU.
- Pour les tâches débordées, la séquence de signal trouvée ne colle pas toujours avec la version CPU.


Conseil pour une meilleure utilisation :
Pour une meilleure performance, il est important de libérer 2 coeurs CPU qui font lorsque vous faites tourner des instances multiples [plusieurs UT].
Libérer au moins 1 cœur CPU est nécessaire pour obtenir une utilisation GPU suffisante.


Instances d'application.
Sur les cartes haut de gamme HD 5850/5870, 6950/6970, 7950/7970, vous pouvez faire tourner 3 instances [UT].

Sur les cartes moyenne gamme HD 5770, 6850/6870, 7850/7870, la meilleure performance devrait être obtenue en faisant tourner 2 instances [UT].

Si vous avez des lags d'affichage, réduisez le facteur de déroulement et la valeur ffa_block_fetch.


Addendum:
L'utilisation de plusieurs cartes [graphiques] dans un système nécessite la libération d'un cœur CPU supplémentaire.

[Infomat]

Merci pour l'info    et la trad  .

[Jaehaerys Targaryen]

Merci pour l'info    et la trad  .

Merci Modesti

[modesti]

En me relisant, j'ai remarqué une erreur que j'ai corrigée (texte barré et souligné)

[Infomat]

Merci pour le SAV