The O2-All Sky Gravitational Wave Search on GPUs - discussion thread.

Jim1348
Jim1348
Joined: 19 Jan 06
Posts: 463
Credit: 257957147
RAC: 0

archae86 wrote:But the reason

archae86 wrote:
But the reason your RX 570 Windows system has failed 100% of 1.07 tasks is not the brand of AMD on the card, nor the model.  Based on elapsed times, I don't think you are running 3X or 4X, and don't have a guess as to what might be wrong.

It is not overclocked, running cool (yes, at 1X), and has the support of two i7-4771 cores.  So it seems to be hit or miss with these cards.  Maybe it is the drivers?  At any rate, something needs to be done to make it reliable, or not many people will be able to make a go of it.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4312
Credit: 250433667
RAC: 35039

First of all, thank you all

First of all, thank you all for participating in this Beta test! It is pretty important for us to verify that this GPU application version produces scientifically valid results. In that sense every bit of information helps, particularly the 'invalid' results. The invalid results tell us that there is work left to do, and (hopefully) give an indication where. The 'valid' results are way less interesting, although they provide a reference for the rate of the 'invalid's.

BM

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4312
Credit: 250433667
RAC: 35039

Jim1348 wrote: How about a

Jim1348 wrote:
How about a CUDA version?  I don't know if that would simplify development

Not at all. Actually it would double the amount of code and thus the effort required to maintain it. We need the OpenCL version anyway for AMD and (possibly later) Intel GPU, as well as for internal (non-E@H) use.

BM

MarkJ
MarkJ
Joined: 28 Feb 08
Posts: 437
Credit: 139002861
RAC: 0

cecht wrote:My one host had

cecht wrote:
My one host had regularly been getting usually two new O2AS20-500 v1.07 tasks about every 1 hr 15 min, but stopped getting new work at 8:19am CDT ("Scheduler request completed: got 0 new tasks"). My other host continues to get O2AS20-500 v1.07 tasks. (I just now reset Project preferences for the "lazy host" and it can get binary pulsar #1 GPU work, which it is running now.) The lack of v1.07 work appears to be a host-specific issue, but I can't figure it out. help?

I don't think its host specific as I have been getting 0 new tasks on most work requests and I am running the Linux CPU (1.01) app only. I'm seeing this on 4 hosts. They occasionally get a single task.

Betreger
Betreger
Joined: 25 Feb 05
Posts: 992
Credit: 1588459068
RAC: 762074

Bernd, thanx for the

Bernd, thanx for the feedback. I shall keep on crunching on my Nvidea card knowing that I'm helping to develop the app. 

Matt White
Matt White
Joined: 9 Jul 19
Posts: 120
Credit: 280798376
RAC: 0

cecht wrote:My one host had

cecht wrote:
My one host had regularly been getting usually two new O2AS20-500 v1.07 tasks about every 1 hr 15 min, but stopped getting new work at 8:19am CDT ("Scheduler request completed: got 0 new tasks"). My other host continues to get O2AS20-500 v1.07 tasks. (I just now reset Project preferences for the "lazy host" and it can get binary pulsar #1 GPU work, which it is running now.) The lack of v1.07 work appears to be a host-specific issue, but I can't figure it out. help?

Craig, is the host that is giving us issues allowed to run beta test? Did one of the settings get inadvertently changed? If you are still using an app config, did you try to re read it? I assume you gave the box a kick to the head (reboot). :)

Other than that, I'm stumped. My server downloaded a few binary pulsar #1 GPU tasks the other day, but it still has v1.07 GW tasks in the work queue. The AMD/LINUX box is crunching along on GW work. both CPU and GPU types.

Clear skies,
Matt
Matt White
Matt White
Joined: 9 Jul 19
Posts: 120
Credit: 280798376
RAC: 0

Bernd Machenschalk

Bernd Machenschalk wrote:
First of all, thank you all for participating in this Beta test! It is pretty important for us to verify that this GPU application version produces scientifically valid results. In that sense every bit of information helps, particularly the 'invalid' results. The invalid results tell us that there is work left to do, and (hopefully) give an indication where. The 'valid' results are way less interesting, although they provide a reference for the rate of the 'invalid's.

Thanks for the update, Bernd. It is refreshing to know that even invalid results are helpful, and are contributing to the effort.

Clear skies,
Matt
cecht
cecht
Joined: 7 Mar 18
Posts: 1533
Credit: 2900492224
RAC: 2186390

Matt White wrote:Craig, is

Matt White wrote:
Craig, is the host that is giving us issues allowed to run beta test? Did one of the settings get inadvertently changed? If you are still using an app config, did you try to re read it? I assume you gave the box a kick to the head (reboot). :)

Yes; no; yes; yes. Thanks for the reminders though. I've reset preferences to again try feeding GW GPU tasks to that host; am waiting to see what happens when its FGRBPG1 task queue runs out.

UPDATE: I'm getting GW 1.07 tasks again (with a few FGRBPG1 tasks left in the queue). Problem solved!, whatever the problem was. Next time, I'll let a "problem" simmer for a few days before getting worked up about it being a problem.

Ideas are not fixed, nor should they be; we live in model-dependent reality.

Betreger
Betreger
Joined: 25 Feb 05
Posts: 992
Credit: 1588459068
RAC: 762074

I have increased my

I have increased my production of GW CPU work and the accompanying high rate of invalid results. I hope this is making Bernd happy. 

n12365
n12365
Joined: 4 Mar 16
Posts: 26
Credit: 6491436572
RAC: 0

Bernd Machenschalk

Bernd Machenschalk wrote:
First of all, thank you all for participating in this Beta test! It is pretty important for us to verify that this GPU application version produces scientifically valid results. In that sense every bit of information helps, particularly the 'invalid' results. The invalid results tell us that there is work left to do, and (hopefully) give an indication where. The 'valid' results are way less interesting, although they provide a reference for the rate of the 'invalid's.

Is running the GPU application at 1x helpful?  I can almost double my output by processing two tasks at time, but I assume it is easier to debug if only one task at a time is being processed.

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.