S@H version 7.02 released for ATI/AMD GPUs under windows.

Message boards : News : S@H version 7.02 released for ATI/AMD GPUs under windows.
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · Next

AuthorMessage
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1667
Credit: 13,187,693
RAC: 12,570
Sweden
Message 45540 - Posted: 11 Apr 2013, 10:59:04 UTC - in response to Message 45538.  

Soon we'll have apps that compile O.K on Cats 13.2 & Cat 13.3 (Both on APP runtime 1124.2) once Alpha testing is complete.

Claggy


But will it also at the same time compile OK on earlier Cats, or will we be having the reversed problem then? They compile fine on newer Cats, but not OK on earlier Cats?
ID: 45540 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 45541 - Posted: 11 Apr 2013, 11:39:52 UTC - in response to Message 45540.  

Soon we'll have apps that compile O.K on Cats 13.2 & Cat 13.3 (Both on APP runtime 1124.2) once Alpha testing is complete.

Claggy


But will it also at the same time compile OK on earlier Cats,

Yes, it should do.

Claggy
ID: 45541 · Report as offensive
TRuEQ & TuVaLu
Volunteer tester
Avatar

Send message
Joined: 28 Jan 11
Posts: 619
Credit: 2,580,051
RAC: 0
Sweden
Message 45543 - Posted: 11 Apr 2013, 14:47:07 UTC - in response to Message 45530.  
Last modified: 11 Apr 2013, 14:47:48 UTC

Now I am getting grey hairs.
When I returned to my computor, BM had exited.

When starting it again 2 Seti Beta tasks had error.

- exit code 194 (0xc2) both of them.

link to the 2 tasks.

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=13391253

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=13391259




So, BOINC just killed fully completed task. And looks like it did that few times in a row.



Happend only once.

And now I'm not getting any tasks. Driver 12.1 might be the cause of that.
I'll upgrade to 12.8 soon.

But most tasks of the non-supposed-to-send-to-opencl-1.1-drivers(7.02) worked.
ID: 45543 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,129,477
RAC: 6,673
United States
Message 45545 - Posted: 11 Apr 2013, 15:30:34 UTC - in response to Message 45531.  
Last modified: 11 Apr 2013, 15:30:52 UTC

In ideal world - yes, they should not. But in reality big companies say something like: "Either eat that shit we throw to you and sit silently or PAY, PAY PAY".


In the end it will be AMD that pays for not caring about their users. Releasing driver sets that cause bluescreens, unexplained reboots or just fail to work is not a good advertisement for their graphics cards. After what they put me through with Angela's computer, I would not buy one, nor would I recommend that anyone buy one.

Since some of our users will place the blame for such problems on SETI@home rather than on AMD, we are facing some risks with AMD apps that we don't typically face with NVIDIA apps.
ID: 45545 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 45547 - Posted: 11 Apr 2013, 17:01:22 UTC - in response to Message 45545.  

we are facing some risks with AMD apps that we don't typically face with NVIDIA apps.

Not quite true.
Do you know what problems current CUDA stock app has with latest NV GPUs?
Here on beta opt app used that cured from those already, but not on main.
Do you know that some driver version range of NV drivers unsuitable for SETI too?
Look for compatibility tricks users have to do... When I said "big company" it touches NV too. Maybe not in such big degree as AMD but...

ID: 45547 · Report as offensive
William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 606
Credit: 588,843
RAC: 0
Message 45548 - Posted: 11 Apr 2013, 17:14:06 UTC

That may well be, but it took NV exactly one bugrep to put in the compatibility value fix.
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 45548 · Report as offensive
TRuEQ & TuVaLu
Volunteer tester
Avatar

Send message
Joined: 28 Jan 11
Posts: 619
Credit: 2,580,051
RAC: 0
Sweden
Message 45550 - Posted: 11 Apr 2013, 18:02:15 UTC

I can't help it.
I like ATI.
ID: 45550 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 45577 - Posted: 14 Apr 2013, 14:46:35 UTC - in response to Message 45524.  
Last modified: 14 Apr 2013, 14:59:09 UTC

Right now we are limited as the ATI developers recommend, OpenCL runtime 938.2 and below which means Catalyst 12.8 and below.

I can confirm that that is working, I'm on OpenCL runtime 1016.4 at the moment and the Server isn't sending me work, while it is sending me Cuda42 work for my Nvidia GPU.

Edit: Off topic, but related, I did send you an email saying minimum driver version for Nvidia OpenCL Astropulse should be 263.00, as the Nvidia Astropulse app never worked (during testing) on 260.99 drivers,
But there is still at least one host getting OpenCL Astropulse work at Seti Main with 260.99 drivers, and all he is doing is producing inconclusive/invalid results:

Computer 4819000

Claggy
ID: 45577 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 45580 - Posted: 14 Apr 2013, 22:39:45 UTC

Maybe we have workaround for Cat 12.10 issues now. If alpha testing will be successful then range of supported Cat drivers will be increased.
ID: 45580 · Report as offensive
Profile cAnDYmanS@H-Beta
Volunteer tester
Avatar

Send message
Joined: 24 May 12
Posts: 38
Credit: 436,379
RAC: 0
Romania
Message 45590 - Posted: 15 Apr 2013, 15:18:30 UTC

Only getting AP 6.05s and they seem to be broken... Are there any S@H's 7.02 in the pipe?
Per aspera, ad astra!

ID: 45590 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 45591 - Posted: 15 Apr 2013, 18:22:29 UTC - in response to Message 45590.  

Are there any S@H's 7.02 in the pipe?

Yes, but to receive them for the Current app you're got to have APP runtime 938.2 or earlier, with later runtimes the Current app eithier produces inconclusive work, or the app causes driver restarts.

Or wait for a week or so, for a fixed app to be Alpha tested, then released.

Claggy
ID: 45591 · Report as offensive
Profile Silver Streak
Volunteer tester

Send message
Joined: 11 Jun 06
Posts: 6
Credit: 711,507
RAC: 0
United States
Message 45592 - Posted: 16 Apr 2013, 5:16:38 UTC - in response to Message 45591.  

I finally got all the right drivers and runtime apps and successfully crunched some S@H 7.02's

The Problem: The scheduler sends both ( AP 6.04 ati opencl 100) which runs normal and validates. And ( AP 6.05 cal ati) which runs only on the cpu.

4/14/2013 12:10:25 AM | | OpenCL: AMD/ATI GPU 0: AMD Radeon HD 7700 series (Cape Verde) (driver version CAL 1.4.1720 (VM), device version OpenCL 1.2 AMD-APP (938.2), 1024MB, 984MB available, 3504 GFLOPS peak)
4/14/2013 12:10:25 AM | | OpenCL: AMD/ATI GPU 1: AMD Radeon HD 7700 series (Cape Verde) (driver version CAL 1.4.1720 (VM), device version OpenCL 1.2 AMD-APP (938.2), 1024MB, 984MB available, 3504 GFLOPS peak)

Should I reset?
ID: 45592 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 45593 - Posted: 16 Apr 2013, 9:06:16 UTC - in response to Message 45592.  

I finally got all the right drivers and runtime apps and successfully crunched some S@H 7.02's

The Problem: The scheduler sends both ( AP 6.04 ati opencl 100) which runs normal and validates. And ( AP 6.05 cal ati) which runs only on the cpu.

4/14/2013 12:10:25 AM | | OpenCL: AMD/ATI GPU 0: AMD Radeon HD 7700 series (Cape Verde) (driver version CAL 1.4.1720 (VM), device version OpenCL 1.2 AMD-APP (938.2), 1024MB, 984MB available, 3504 GFLOPS peak)
4/14/2013 12:10:25 AM | | OpenCL: AMD/ATI GPU 1: AMD Radeon HD 7700 series (Cape Verde) (driver version CAL 1.4.1720 (VM), device version OpenCL 1.2 AMD-APP (938.2), 1024MB, 984MB available, 3504 GFLOPS peak)

Should I reset?

1) It doesn't run only on CPU. BUT GPU usage is low.
2) As i said before in another thread there is no mechanism currently to separate Brook+ tasks from OpenCL tasks for same GPU. You can just deselect AP task in web settings and recive only OpenCL work for OpenCL MB7 application.
ID: 45593 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,129,477
RAC: 6,673
United States
Message 45600 - Posted: 17 Apr 2013, 17:13:35 UTC - in response to Message 45593.  

I hope to get the mechanism we need implemented in the next week.
ID: 45600 · Report as offensive
TRuEQ & TuVaLu
Volunteer tester
Avatar

Send message
Joined: 28 Jan 11
Posts: 619
Credit: 2,580,051
RAC: 0
Sweden
Message 45601 - Posted: 17 Apr 2013, 19:13:55 UTC

I wrote this in another thread:

http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=1987&postid=45594#45594

I don't know if it is of any helpt though....maybe.
ID: 45601 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1667
Credit: 13,187,693
RAC: 12,570
Sweden
Message 45613 - Posted: 20 Apr 2013, 20:29:54 UTC
Last modified: 20 Apr 2013, 20:39:21 UTC

Tasks in cache:

SETI@home Beta Test 7.02 SETI@home v7 (ati_opencl_sah) 931 [Tasks]
SETI@home Beta Test 7.02 SETI@home v7 (ati_opencl_pre129) 483 [Tasks]
SETI@home Beta Test 7.00 SETI@home v7 96 [Tasks]


I'm setting the resource share to "0" for Beta on http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=57178. It will of course finish all those 7.02 WUs, and the 7.00 CPU WU's.

If there's a new app coming, I will change the resource share, so it will get the new app to test.
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 45613 · Report as offensive
Profile cAnDYmanS@H-Beta
Volunteer tester
Avatar

Send message
Joined: 24 May 12
Posts: 38
Credit: 436,379
RAC: 0
Romania
Message 45622 - Posted: 22 Apr 2013, 18:48:06 UTC - in response to Message 45613.  

Meanwhile in Russia...Just kidding. I only have a couple AP 6.04s, which is strange to me because AP 6.04 is the live app on S@H. Well, I will crunch them nonetheless.

Cheers!
Per aspera, ad astra!

ID: 45622 · Report as offensive
Profile cAnDYmanS@H-Beta
Volunteer tester
Avatar

Send message
Joined: 24 May 12
Posts: 38
Credit: 436,379
RAC: 0
Romania
Message 45662 - Posted: 28 Apr 2013, 5:17:44 UTC

Eric, can you please look into this: http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2028?

I know it's AP, but seems to be getting quite serious.

Thanks!
Per aspera, ad astra!

ID: 45662 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1667
Credit: 13,187,693
RAC: 12,570
Sweden
Message 45673 - Posted: 30 Apr 2013, 19:17:59 UTC - in response to Message 45613.  

Tasks in cache:

SETI@home Beta Test 7.02 SETI@home v7 (ati_opencl_sah) 931 [Tasks]
SETI@home Beta Test 7.02 SETI@home v7 (ati_opencl_pre129) 483 [Tasks]
SETI@home Beta Test 7.00 SETI@home v7 96 [Tasks]


I'm setting the resource share to "0" for Beta on http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=57178. It will of course finish all those 7.02 WUs, and the 7.00 CPU WU's.

If there's a new app coming, I will change the resource share, so it will get the new app to test.


OK, I'm done with all 7.02s, all 1414 of them. Only 3 errors, and those were my own fault, for playing around too much with micro managing :-)

Only 27, 7.00 CPU task left now.

No problems whatsoever with CPU 7.00 or ATI 7.02 on my Q8200/ATI HD7870.

Now I will let the computer run on main for a while at 100%.
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 45673 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 45721 - Posted: 6 May 2013, 21:25:54 UTC

I see the r1831 HD5 and non-HD5 apps have made it onto Beta as version 7.03 in pre and post Cat 13.1 app versions. :-)

Claggy
ID: 45721 · Report as offensive
Previous · 1 · 2 · 3 · 4 · Next

Message boards : News : S@H version 7.02 released for ATI/AMD GPUs under windows.


 
©2020 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.