Salut Yannock73,
Essaie en français si tu peux.
Je pense que ce sera plus facile et plus clair et je pourrais peut-être t'aider.
Et puis, tu as des cracks dans ton équipe aussi. Je pense notamment à [AF>Amis des Lapins] Phil1966

---

[CSF] Thomas H.V. Dupont
Founder of the team CRUNCHERS SANS FRONTIERES 2.0
www.crunchersansfrontieres

yannock73, ce que tu décris est le fonctionnement normal de l'application sur un GPU ATI (utilisation d'un seul GPU et de tous les coeurs CPU).
N'oublie pas que l'application est en phase de développement.

Oh des frenchies ;)
Bon si ça parait déjà "normal" ça me va.
Mais c'est quand même handicapant pour les autres taches si GPUGrip monopolise toute l'attention de BOINC pour 1 tache, et qui plus est, utilise toute les ressources et tout les CPUs et ainsi empêche les autres d'avancer. Mais bon c'est une beta donc à suivre.
J’étais sur SETI il y a .... pfiuuu trop longtemps, puis j'ai arrêter.
Je reviens suite a un reportage et je redécouvre le calcul partagé.
BOINC, c'est franchement extra et la plupart des projet sont vraiment intéressants. J’espère que ma modeste contribution fera vraiment avancer la science :)

I would like to read the AMD Answeres too so can we plz go back to English? ^^

---

DSKAG Austria Research Team: http://www.research.dskag.at

I would like to read the AMD Answeres too so can we plz go back to English? ^^

I used https://translate.google.com/ copy and paste text. Auto detects language.

---

Oh des frenchies ;)
Bon si ça parait déjà "normal" ça me va.
Mais c'est quand même handicapant pour les autres taches si GPUGrip monopolise toute l'attention de BOINC pour 1 tache, et qui plus est, utilise toute les ressources et tout les CPUs et ainsi empêche les autres d'avancer. Mais bon c'est une beta donc à suivre.
J’étais sur SETI il y a .... pfiuuu trop longtemps, puis j'ai arrêter.
Je reviens suite a un reportage et je redécouvre le calcul partagé.
BOINC, c'est franchement extra et la plupart des projet sont vraiment intéressants. J’espère que ma modeste contribution fera vraiment avancer la science :)

Welcome back alors Yannock73 !
Cruncher un jour, cruncher toujours :)

---

[CSF] Thomas H.V. Dupont
Founder of the team CRUNCHERS SANS FRONTIERES 2.0
www.crunchersansfrontieres

Is there a possibility of an AMD Beta App specific badge, perhaps with multiple levels based on contribution, similar to the scientific publication badges? (Eg. Bronze, Silver, Gold, etc)

hello ! :)
why using radeons only by now ? wy nvidia privigege all this time
nvifdia dont have double check, then dont apear and fix the errors...
was not better to use double check processing ?
at my olld time of boinc the only one board that worked was the hds radeons... bcs of its non error result, it apear the error before, and dont send an fake result.. as in nividia ones at the time...
i dont find the otion for amd gpu.. only amd cpu gpu there...
and im only geting this cpus 8 core things.. that are not soo good in here, its using 1/3 1/2 of the cpus power... maybe it was reduced bcs of users overheating ?
all projects i got 28-35mips, with this 8core task im geitng 15mips, and the processor are cold as an ice..
also i want to use the gpu... where is the option ?

hello ! :)
why using radeons only by now ? wy nvidia privigege all this time
nvifdia dont have double check, then dont apear and fix the errors...
was not better to use double check processing ?
at my olld time of boinc the only one board that worked was the hds radeons... bcs of its non error result, it apear the error before, and dont send an fake result.. as in nividia ones at the time...
i dont find the otion for amd gpu.. only amd cpu gpu there...
and im only geting this cpus 8 core things.. that are not soo good in here, its using 1/3 1/2 of the cpus power... maybe it was reduced bcs of users overheating ?
all projects i got 28-35mips, with this 8core task im geitng 15mips, and the processor are cold as an ice..
also i want to use the gpu... where is the option ?

I will start the answers...first MOST Nvidia gpu's have dual precision built in, unlike AMD cards that only certain models have it.

Second the project WANTS to use the cpu AND the gpu, not JUST the gpu, so using all or most of your cpu cores is a project choice to do it.

As for not using your Nvidia gpu for this set of units, they are designed for AMD gpu's, NOT Nvidia gpu's. You can still use your Nvidia cards on the regular gpu units.

OK :...
Only now with trottle i see that it is using the gpu about 60%...
and is using "8 cpus" at almost like 10%... thats why the cpu are so cold..
i i was wanting to get amd gpu task.. bu i already having it... then ?
bcs i was having this msg: "project has no apps for amd/ati gpus"
but in fact is that the using (8 cpu) in fact uses the gpu as well... right ?
or my gpu crashed looping in something before ? lollolo

OK... i see what is happening now:
See: this tasks arent using 8cpus, at max it use half of total...
then.... another task start to cover the iddle and crush all...
for example now: it was runing the "8 cpu" task ias i told before... and boinc started another task:/ POem@ (0,8cpu + 1gpu) .. and now i thing it is using the gpu and leaving again the Gpu gripd with processors.. OR>>> the problem was me that not see this poem task before, and the gpu activity was from it ?
so.. the cpu and gpu adm tasks are working on ati ?
am i taking use of my gpu on this project ?
Bcs, if it is only about cpus, i cold use, but IF the task uses it, and not 30%... please help to understand better or to fix it...

OK :...
Only now with trottle i see that it is using the gpu about 60%...
and is using "8 cpus" at almost like 10%... thats why the cpu are so cold..
i i was wanting to get amd gpu task.. bu i already having it... then ?
bcs i was having this msg: "project has no apps for amd/ati gpus"
but in fact is that the using (8 cpu) in fact uses the gpu as well... right ?
or my gpu crashed looping in something before ? lollolo

Yes you ARE doing the AMD gpu app, it is NOT like at other projects where most of it runs on ONLY the gpu, here without the cpu cores it essentially won't run at all. This is by design, iow the Project wants it that way.

try can app_config

for i5

<app_config>
    <app>
    <name>amdmd</name>
    <max_concurrent>2</max_concurrent>
    <gpu_versions>
    <gpu_usage>1</gpu_usage>
    <cpu_usage>1</cpu_usage>
    </gpu_versions>
    </app>
    <app_version>
    <app_name>amdmd</app_name>
    <plan_class>mt</plan_class>
    <avg_ncpus>1</avg_ncpus>
    <ngpus>1</ngpus>
    <cmdline>--nthreads 2</cmdline>
    </app_version>
    </app_config>

for i7

 <app_config>
    <app>
    <name>amdmd</name>
    <max_concurrent>3</max_concurrent>
    <gpu_versions>
    <gpu_usage>1</gpu_usage>
    <cpu_usage>1</cpu_usage>
    </gpu_versions>
    </app>
    <app_version>
    <app_name>amdmd</app_name>
    <plan_class>mt</plan_class>
    <avg_ncpus>1</avg_ncpus>
    <ngpus>1</ngpus>
    <cmdline>--nthreads 2</cmdline>
    </app_version>
    </app_config>

for i7 advanced -> WU release every 3 min (WU sometimes bad start)

 <app_config>
    <app>
    <name>amdmd</name>
    <max_concurrent>4</max_concurrent>
    <gpu_versions>
    <gpu_usage>1</gpu_usage>
    <cpu_usage>1</cpu_usage>
    </gpu_versions>
    </app>
    <app_version>
    <app_name>amdmd</app_name>
    <plan_class>mt</plan_class>
    <avg_ncpus>1</avg_ncpus>
    <ngpus>1</ngpus>
    <cmdline>--nthreads 1</cmdline>
    </app_version>
    </app_config>

works only for newly counting WU

I don't think I'd use --nthreads 2 with avg_ncpus 1 (first two examples). BOINC will seriously over-commit the CPU if it has other CPU tasks available to run - CPU apps will be stalled by contention and task-switching.

You're right
all boinc only for GRID GPU

What Richard said is correct, because a user would not want to overcommit their system.

I am attached to 30 projects, and know quite a bit about it :)

No more WU for AMD ? My i7 with a Radeon is not receiving anything (since, friday i guess, didn't check when it has stopped)

No more WU for AMD ? My i7 with a Radeon is not receiving anything (since, friday i guess, didn't check when it has stopped)

+1. Same here.
New work units stopped during Saturday for me I think and nothing since.
:-(

New work units stopped during Saturday for me I think and nothing since.
:-(

Is the AMD beta test finished?
Are you planning a stable release of this client (and a stable queue of wus)?

Any ETA for this application to be available for systems with multiple gpu's?

My secondary system running an i7 920 & 7950 is chugging along just fine.