If I start BOINC is see:

)	
OpenCL CPU: Intel(R) Xeon(R) CPU E5-2630 v2 @ 2.60GHz (OpenCL driver vendor: Advanced Micro Devices, Inc., driver version 1912.5 (sse2,avx), device version OpenCL 1.2 AMD-APP (1912.5))	

...does this mean that on my system it's possible to use an ATI app which support AVX?
The currently available ATI apps use 'just' SSE2 (on/off the CPU).

No, It means Boinc detects CPU OpenCL support, meaning if anyone ever want to produce a multi-core OpenCL CPU app then they can be sent to supporting hosts.

Claggy

Windows x86 SSE2

Fantastic! Thank you. I'll hook my ThinkPad up to Beta a little later and see how it runs overnight.

If I start BOINC is see:

)	
OpenCL CPU: Intel(R) Xeon(R) CPU E5-2630 v2 @ 2.60GHz (OpenCL driver vendor: Advanced Micro Devices, Inc., driver version 1912.5 (sse2,avx), device version OpenCL 1.2 AMD-APP (1912.5))	

...does this mean that on my system it's possible to use an ATI app which support AVX?
The currently available ATI apps use 'just' SSE2 (on/off the CPU).

No, It means Boinc detects CPU OpenCL support, meaning if anyone ever want to produce a multi-core OpenCL CPU app then they can be sent to supporting hosts.

Claggy

The ATI app would run faster if she would use AVX instead of SSE2 (of the CPU)?

Thanks.

---

Windows x86 SSE2

Fantastic! Thank you. I'll hook my ThinkPad up to Beta a little later and see how it runs overnight.

I would suggest to check SSE vs SSE2 speed before. They both did not tested on target (that is, SSE+SSE2-only) hardware so far. So what is faster - unknown.

---

News about SETI opt app releases: https://twitter.com/Raistmer

The ATI app would run faster if she would use AVX instead of SSE2 (of the CPU)?

Thanks.

No. Only chirping AVX-enabled there and chirping done on GPU. AVX used in FFTW though.

---

News about SETI opt app releases: https://twitter.com/Raistmer

Windows x86 SSE2

Fantastic! Thank you. I'll hook my ThinkPad up to Beta a little later and see how it runs overnight.

I would suggest to check SSE vs SSE2 speed before. They both did not tested on target (that is, SSE+SSE2-only) hardware so far. So what is faster - unknown.

Okay, I've got the SSE running on its first task now. AR is 2.729883, so it should run quicker than a normal AR, but I don't have anything to compare it to with v8. On v7, ARs around 2.7 seemed to run about 5.5 hours on SSE2 vs. about 6.0-6.5 hours on stock.

I'll let SSE run overnight, then switch to SSE2 tomorrow.

EDIT: I see wingman Zalster's TITAN X finished that WU in a little over 4 minutes. My old Pentium M is going to need juuusssst a bit more time than that. ;^)

Good evening all I seem to have a small problem for the last couple of days I have been trying to run v8 beta on my android phone everything seems fine while it is running but if I unplug my phone for any reason all progress is reset to 0 and it starts all over again when I plug it back in(v7 did not do this it resumed where it left off) bit I did find out that if I open boinc and suspend the tasks before I unplug my phone they resume where they left off when I plug it back in it is a little puzzling that it does this. Thanks for any help with this.

---

ME AND MY BOY LOOKING FOR E.T.

OK, I installed the Windows x64 AVX and Windows OpenCL ATi HD5 Signals on GPU build.

Something strange... the CPU time is higher than the Run time with the ATI app: https://setiweb.ssl.berkeley.edu/beta/results.php?hostid=76831&offset=0&show_names=0&state=4&appid=38

The PC have two Xeon CPUs (6 cores/12 threads each), but let it run without HT (for faster GPU app crunching), so real 12 CPU-Cores are available.
It's Win8.1 Pro x64. It looks like the Task-Manager isn't so well there.
100% / 12 = (one core=) 8.33 % of whole CPU(s)

With the SAHv7 ATI r2929 app, in Task-Manager one ATI app used up to 8% CPU(s) (one whole CPU-core) (at the Details tab overview).

Now with SAHv8 ATI r3256 app, one ATI app use up to 11% CPU(s), so ~ 1 core + 1/3 core.

I need to 'worry' because of something?

Thanks.

---

Found the problem with the Android versions. FFTW's attempt to time routines weren't ever finishing, so I'll have to look into FFTW's timing code. It's nice to see it wasn't my fault this time. Till then I can test versions using FFTW_ESTIMATE rather than FFTW_MEASURE for the timing.

I'll definitely get the OpenCL versions out to beta tomorrow.

---

Today 141 tasks completed for my NV 560ti and daily quota is reached.
The card can do some more if needed.


I must have missed the thread about the quota system.
Can anyone please enlighten me.

Que of 50 tasks is ok with me.

As Zalster said.
NV app .3314 is hogging a full cpu core...but only on Beta.
I for some stupid reason ran a few tasks om main with same app and there I had no hogging a full cpu core problem. App used cpu core as it supposed to.

http://setiathome.berkeley.edu/forum_thread.php?id=78702&postid=1753925#1753925

1 error to report. Ran on main with .3299 ATI HD V8
It ran on the main card that is connected to the screen.
computor has 1.75GB of RAM available for windows. And uses often 3.5GB of swapping page file. so comp is often heavy loaded. At the moment the error occured i was participating in an online auction which might have an activeX or Java component that was in use at the time. Browser was Firefox. Comp is running ATI 5970(2Gpu's) + 1 ATI 5850 and the 5850 was running the task that errored.

Why I find this intresting is that normally I get a driver restart or bluescreen when this happend before. This time I only got a "computation error". :)

It might be wu that is fault here.
Anyone intrested in rerunning this wu to see if wu is ok?

link to error task:
http://setiathome.berkeley.edu/results.php?hostid=7358242&offset=0&show_names=0&state=6&appid=

0012f074 00406637 00000001 00000010 00000000 00000000 MB8_win_x86_SSE2_OpenCL_ATi_HD5!checkpoint+0x75 (d:\r\sah_v7_opt\akv8\client\seti.cpp:711)
0012fcd8 0041c32b 005083c0 00000000 0046cdd0 00000010 MB8_win_x86_SSE2_OpenCL_ATi_HD5!seti_analyze+0x0 (d:\r\sah_v7_opt\akv8\client\analyzefuncs.cpp:5858)

similar crash was for NV build in offline Claggy's test

---

News about SETI opt app releases: https://twitter.com/Raistmer

So the wu is broken?

0012f074 00406637 00000001 00000010 00000000 00000000 MB8_win_x86_SSE2_OpenCL_ATi_HD5!checkpoint+0x75 (d:\r\sah_v7_opt\akv8\client\seti.cpp:711)
0012fcd8 0041c32b 005083c0 00000000 0046cdd0 00000010 MB8_win_x86_SSE2_OpenCL_ATi_HD5!seti_analyze+0x0 (d:\r\sah_v7_opt\akv8\client\analyzefuncs.cpp:5858)

similar crash was for NV build in offline Claggy's test

My first v8-task on the Android tab seems to slow down almost to a complete stop on its last five percent.

The task was downloaded last Thursday morning. On Saturday it was already at 94.2 percent. Since then it seems to be going slower and slower, and has spent the whole day today (Tuesday) going from 99.6 to 99.7.

If we're lucky it will complete during the next couple of days.

This behaviour may of course be "by design", but if it isn't, then there may be something fishy in the processing of the last five percents.

My first v8-task on the Android tab seems to slow down almost to a complete stop on its last five percent.

The task was downloaded last Thursday morning. On Saturday it was already at 94.2 percent. Since then it seems to be going slower and slower, and has spent the whole day today (Tuesday) going from 99.6 to 99.7.

If we're lucky it will complete during the next couple of days.

This behaviour may of course be "by design", but if it isn't, then there may be something fishy in the processing of the last five percents.

Have you not been reading this thread, All the Android for Arm apps have been deprecated because they didn't work,
they didn't get past the wisdom generation, what you are seeing Boinc's estimation of progress:

https://github.com/BOINC/boinc/commit/9136a369d4e15cc727c06b55b50c833e184bf9fc

client: if app doesn't report fraction done, estimate it

https://github.com/BOINC/boinc/commit/34f252870310b18c7cbe3e71573daff6b01e768c

client: if app doesn't report fraction done, estimate fraction done in a way that converges to but never reaches 100%.

You might as well abort them, they aren't making any progress, If you restart the host, you'll find the app hasn't even made it to the first checkpoint.

Claggy

Have you not been reading this thread, All the Android for Arm apps have been deprecated because they didn't work,
they didn't get past the wisdom generation, what you are seeing Boinc's estimation of progress:

Reading, yes. Understanding... eh.. no :-)

I restarted the host, verified that they were at 0.0% and then aborted them.

Thanks for clarifying.

I just released fixed versions about a minute ago and will send a message to Matteo and Steven (the creators of FFTW) about the problem. The fix is pretty easy, but could cause problems on SGI machines (if they still exist).

---

I just released fixed versions about a minute ago and will send a message to Matteo and Steven (the creators of FFTW) about the problem. The fix is pretty easy, but could cause problems on SGI machines (if they still exist).

I just tapped "update" and got two fresh tasks. Let's hope these work better then.

I now use v8.000 on main. V8.02 here on beta.
And v8.05 is now as "stock".

Pease tell us/me what version we/I should run.
And I will switch to it as soon as i can.