Hmm. My ATI OSX/ML machine can't get work. 7.0.31.
____________

Got some. Thanks!
____________

I did 1 task.

http://albert.phys.uwm.edu/results.php?hostid=3755

Was faster then the older one.

Good work.

http://albert.phys.uwm.edu/workunit.php?wuid=98764
Looking good so far. Great work.
Stephan


____________

Crunched roughly 50 so far with no problems; using an HD 6850 with Boinc v.7.28 and OC'd i5 2500K (4.5GHz).

Time reduced a lot from average 3,800secs to 2,100secs doing 1xWU at a time...now if you could reduce the CPU Test WUs from 16,000+ secs I wouldn't mind. :)

Congratulations!

My pc did a BRP 1.28 in ~1980 sec. HD6900 Series GPU, i7 with free cpu, 7.0.31, CCC 12.6
On HD5850 a wu takes ~3450sec.
AMD-apps are grown up now.
____________

Phantastic!



The amd is a HD6950 running a single wu @ 82% gpu load, the nVidia is a GTX680, unknown settings.

edit: uups, picture is missing...


wu 997334
task 279100 runtime 1,981.83 Binary Radio Pulsar Search v1.28 (opencl-ati)
task 279099 runtime 11,162.70 Binary Radio Pulsar Search v1.25 (BRP3cuda32)
____________

Hi,

actually the app should do even faster, did you free a CPU core on the machine with the HD 5850 as well? I have a HD 5850 that is doing a single unit in ca 1800 sec (!!) (although that is on a comparatively fast dual core 3 GHZ Core2 (no hyperthreading!) with only one other task running on the CPU.

Cheers
HB

Seeing as there are few projects that will crunch WUs on my MacBook's GPU, I'm always interested to see how it pans on a new release.

1. the time estimates are way off. initial estimate was >50 hours. task crunched in about 7 hours. IIRC this is down from 12 hours.
2. the CPU utilization was 0.564 of a core. I think this is high (Collatz grabs just 0.02 of a core). Even so I think this is a reduction
3. i still only got 500 credits. for 7 hours crunching Collatz gives me ~6,500

I'm not whining about the credits - I read, understood and accepted the notice. But this is a feedback thread after all.

I enjoy the credit game as much as the next cruncher but I don't always pick the highest paying project. But i'd like a decent whack even so. What i'd like to know is if and when these apps move to production, what will be the going rate? I have a very slight preference for astronomy projects over mathematical projects but 500 credits won't get my GPU. I know, I'm just one and there's plenty more dedicated fish in the sea. nonetheless, this is my feedback.

I'll keep coming back and crunching a WU each time there's a new app. After all, you went to the trouble of making an OSX app, the least I can do is test it.

Thanks!

I could not resist ..

I 'borrowed' my old testsystem and tried some wu's.
AMD A8 3870 APU: wu 100169 runtime 6,489.60
HD5830 runtime ~ 2,915.96
____________

Are the CUDA version released?
I am seeing the BRP4 v1.28 version is running on my BOINC menager

I am looking forward to testing out BRP4 v1.28 for Linux. Thanks for the updates!
____________

The speed is amazing.Going from 3800s to 2200s for single task,and 3500s for 2 task.That is on PCIe x16 v1.1 not on v3.0

Great work

Yesterday I upgraded my GPU to a factory over clocked GTX 660Ti.

I've run about 15 units, two at a time, and observed a speedup of about 700s compared to the BRP4 v1.25 on Einstein. At the same time the CPU-time per result has decreased by about 400s, thats more than half on my system.

v1.28 on Albert, x2, run time ~2170s and CPU time ~360s.
v1.25 on Einstein, x2, run time ~2900 and CPU time ~770s.

GPU-load also increased from ~80% to 95+%.

Great work!
____________

How bout a single work unit?I want to see if a 660Ti is faster than a 580gtx/

How bout a single work unit?I want to see if a 660Ti is faster than a 580gtx/

BRP3Cuda32
GTX550Ti 1793 / 303 single wu
i3 win7/64 7.0.31 x16 slot
____________

I ran the new CUDA 1.28 app via one of my Windows systems today. I have not been able to get much work today but the two tasks that ran via my GTX 580, completed at 834 seconds each. This is with one task running at a time. GPU load was at approximately 90-91% while running one task.

If memory serves me right, the previous application ran at around 1360 seconds per task with the 1.25 app via this system. This is a very decent improvement in performance. Thanks for the work put into optimizing the BRP4 applications.

How bout a single work unit?I want to see if a 660Ti is faster than a 580gtx/

Here are some preliminary numbers for the GTX 680.

One task per GPU

System #1 - Single GPU

x16 3.0 - 721 seconds

System #2 - Multi GPU

x16 3.0 - 785 seconds
x8 3.0 - 901 seconds

Overall, the performance looks great so far. I want to do some more testing with multiple tasks running at once, different PCI-E configurations, and with the CPU dedicated for BRP4 GPU only. The above tests were done with ~50% CPU load from running other CPU tasks at the same time.

I have a possible finding--not even a little bit sure--but am posting in case others might spot such a thing.

I've got two different hosts with the same GPU, a GTX 460. Neither has had downclocking problems for some weeks, but I found both downclocked severely today, with the problem persisting through system reboot.

It might just barely be possible that running the current Albert BRP1.28 CUDA ap, or on an even less likely note the Albert 0.29 Gamma Ray Pulsar application--or switching back and forth from those to the current Einstein applications was involved.

More likely something else in my system's history was the problem, but I thought I'd post the suspicion in case someone else sees something.

I'm not even sure what the true downclock frequency was, as different sources reported different numbers, but it was either 405 MHz or less--the reduction in power consumption and GPU temperature, while reporting exceedingly high GPU utilization, but making very slow progress on the WU was persuasive that downclocking was at hand.
____________

I have sometimes the same issue.It only happens when i overclock the gpu +130Mhz from baseline.After several hours the screen goes blank and with gpu-z it shows core speed of 400Mhz then i proceed with reboot and everything goes to normal.
I think it some fail safe method that Nvidia uses.Also I noticed when boinc is not running the core speed drops to 50mhz and goes to 800 instantaneous if a demanding gpu work is needed

Another idea: What I'm always a bit suspicious about (BOINC-performance-wise) is OSX's power saving feature to automatically switch between a CPU build-in GPU (e.g. the Intel stuff in Sandy/Ivy Bridge CPUs) and the dedicated graphics card. Could this be the case that this feature is set differently on those two hosts perhaps?

After some time I decided to look for Albert again, this time with my nvidia. The reason was:
30-Aug-2012 08:53:39 [Albert@Home] Requesting new tasks for ATI
30-Aug-2012 08:53:44 [Albert@Home] Scheduler request completed: got 0 new tasks

To my surprise there was work:

30-Aug-2012 10:41:08 [Albert@Home] Started download of einsteinbinary_BRP4_1.28_i686-pc-linux-gnu__BRP3cuda32nv270

Hmmm. It looks like there is not only a 1.28 OpenCL binary and it looks like we can expect a newer CUDA binany. At the moment we have 1.24 in Einstein.

Bernd or Bikeman, can you please give a bit more information? :)
Thanks,
Stephan
PS: Ah ... I see that there was an bit of information. But mostly for Windows. Now let's see what my Linux box can do.
____________

The older cards are also running well with the new version.

8800GT G92 512 MB - x16 slot @ 5.0 GT/s

1.28: 2940 seconds
1.24: ~3600 seconds