New AstroPulse applications for GPUs.


log in

Advanced search

Message boards : News : New AstroPulse applications for GPUs.

Previous · 1 . . . 6 · 7 · 8 · 9 · 10 · Next
Author Message
ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4067
Credit: 32,889,251
RAC: 7,088
United Kingdom
Message 1332923 - Posted: 30 Jan 2013, 20:14:38 UTC - in response to Message 1332893.
Last modified: 30 Jan 2013, 20:32:50 UTC

Updated AMD driver on one of my newer office computers so I could get it working on WUs the GPU could help with.
Upon reboot of system and BOINC, I got the blue screen of death.
A file by the name of, approximately, atiok3x2.dll was trying to write to rom, and it was stopped.
Before I complain to IT, any thoughts from the BOINC side?
(Yes, the provost knows I run the program. There will be no NEZ type stuff here.)


All AMD Catalyst starting from 12.10 are broken in term of usage with SETI OpenCL apps (at least with OpenCL MB app). And 13.2 beta 3 known to be unable to compile CL files for both apps at all.
So, better to stay with Cat 12.1 or 12.8.


OK, my action was prompted by a notice in BOINC manager, so I went directly to the AMD website and used their auto-detect to see what system I have and let it auto-decide what driver I needed.
So, between classes, or just after the last class of the afternoon, I will have to check what driver it gave me. I think I might find this fairly easily, but, just in case, can you or someone else indicate where to look for what driver it installed?


Fire up Catalyst Control Centre from the Task Bar, then go to Information, then Software, it should then give you a Catalyst version.

If you want to revert to an earlier driver, Uninstall the Catalyst driver, reboot, then Use the AMD Catalyst Uninstall utility (only for Windows 7 at the moment), reboot, then install the older driver,
(The Catalyst driver Installer/uninstaller doesn't remove all the files from System32 and SysWOW64 directories, and earlier driver Installers won't overwrite later APP runtimes)

Claggy

Profile ivan
Volunteer tester
Avatar
Send message
Joined: 5 Mar 01
Posts: 602
Credit: 135,417,094
RAC: 129,963
United Kingdom
Message 1332924 - Posted: 30 Jan 2013, 20:16:52 UTC - in response to Message 1332893.

So, between classes, or just after the last class of the afternoon, I will have to check what driver it gave me. I think I might find this fairly easily, but, just in case, can you or someone else indicate where to look for what driver it installed?

Usually, Control Panel->System->Hardware->Device Manager->Display Adapters->[Your Card (double- or right-click->Properties)]->Driver.
____________

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1332943 - Posted: 30 Jan 2013, 20:50:47 UTC - in response to Message 1332923.

Updated AMD driver on one of my newer office computers so I could get it working on WUs the GPU could help with.
Upon reboot of system and BOINC, I got the blue screen of death.
A file by the name of, approximately, atiok3x2.dll was trying to write to rom, and it was stopped.
Before I complain to IT, any thoughts from the BOINC side?
(Yes, the provost knows I run the program. There will be no NEZ type stuff here.)


All AMD Catalyst starting from 12.10 are broken in term of usage with SETI OpenCL apps (at least with OpenCL MB app). And 13.2 beta 3 known to be unable to compile CL files for both apps at all.
So, better to stay with Cat 12.1 or 12.8.


OK, my action was prompted by a notice in BOINC manager, so I went directly to the AMD website and used their auto-detect to see what system I have and let it auto-decide what driver I needed.
So, between classes, or just after the last class of the afternoon, I will have to check what driver it gave me. I think I might find this fairly easily, but, just in case, can you or someone else indicate where to look for what driver it installed?


Fire up Catalyst Control Centre from the Task Bar, then go to Information, then Software, it should then give you a Catalyst version.

If you want to revert to an earlier driver, Uninstall the Catalyst driver, reboot, then Use the AMD Catalyst Uninstall utility (only for Windows 7 at the moment), reboot, then install the older driver,
(The Catalyst driver Installer/uninstaller doesn't remove all the files from System32 and SysWOW64 directories, and earlier driver Installers won't overwrite later APP runtimes)

Claggy


Looks like they gave Catalyst Version 13.1.
I'll have to worry about this in an hour.
Next question is where to get the old driver (like 12.1, 12.8). Is that easy to locate on the AMD website?

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1332965 - Posted: 30 Jan 2013, 22:07:57 UTC - in response to Message 1332924.

So, between classes, or just after the last class of the afternoon, I will have to check what driver it gave me. I think I might find this fairly easily, but, just in case, can you or someone else indicate where to look for what driver it installed?

Usually, Control Panel->System->Hardware->Device Manager->Display Adapters->[Your Card (double- or right-click->Properties)]->Driver.

To follow my last post, and to respond to this, specifically, Driver 9.0.100.2000. It seems the numbers referred to in the post I responded to an hour ago refer to the version of Catalyst/CCC? And I'm assuming that the CCC version and the driver I get might be tied together ... ?

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4067
Credit: 32,889,251
RAC: 7,088
United Kingdom
Message 1332980 - Posted: 30 Jan 2013, 23:29:06 UTC - in response to Message 1332943.
Last modified: 30 Jan 2013, 23:32:10 UTC

Looks like they gave Catalyst Version 13.1.
I'll have to worry about this in an hour.
Next question is where to get the old driver (like 12.1, 12.8). Is that easy to locate on the AMD website?


The old drivers are on a link at the Bottom of the Current driver download page, first choose your OS:

http://sites.amd.com/us/game/downloads/Pages/downloads.aspx

Then go to the bottom of the next page.

Or you can use the links on Hal6000's ATI Driver Version Cheat Sheet

Claggy

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333230 - Posted: 31 Jan 2013, 18:01:30 UTC

Going back to CCC Version 12.1 did NOT change the DRIVER. The driver is still 9.0.100.2000.

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4067
Credit: 32,889,251
RAC: 7,088
United Kingdom
Message 1333267 - Posted: 31 Jan 2013, 19:51:55 UTC - in response to Message 1333230.

Going back to CCC Version 12.1 did NOT change the DRIVER. The driver is still 9.0.100.2000.

Did you use the Catalyst Uninstall Tool? as that deletes the left over files, enabling older drivers to be installed.

Claggy

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333284 - Posted: 31 Jan 2013, 20:46:30 UTC - in response to Message 1333267.

I did a "roll back driver" while in control panel/system/hardware/... .
It is now at 8.821.0.000.
No more blue screens, but BAM does not indicate GPU computing is enabled.
Under Tools/Computing Preferences, use of GPU was already checked (by default?).

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333323 - Posted: 31 Jan 2013, 22:48:51 UTC - in response to Message 1333284.

Searched high and low for an older version of CCC. Found one. Installed. BAM now indicates GPU computing enabled. Now to see if it'll actually download such WUs and process correctly.

Vito
Send message
Joined: 6 Dec 12
Posts: 4
Credit: 136,737
RAC: 174
United States
Message 1333380 - Posted: 1 Feb 2013, 1:34:19 UTC

I joined SETI at the end of the year so I was just getting started when my graphics card began having problems and yesterday I upgraded to an Win 8 I7 with a Nividia GT 640 and late yesterday over 100 items were downloaded to my Computer. I had never seen more than 14 with my old computer. Several of the work units processed much quicker than I could track them. But today there are still about 100 that have been sitting in a Status of Downloading all day and all are SETI@homeEnchanced6.10 (cuda_fermi). I saw at least 2 astropulse process but these 100 are just sitting there. I cannot make the execute.

I have now changed my setting to only download Astropulse 6 I believe.
I hate to just abort them.

Has anyone else seen this.

rob smithProject donor
Volunteer tester
Send message
Joined: 7 Mar 03
Posts: 8302
Credit: 55,197,961
RAC: 75,529
United Kingdom
Message 1333432 - Posted: 1 Feb 2013, 6:21:28 UTC - in response to Message 1333381.

A whole pile of your results have ended in errors.
If you visit the "number crunching" message board you will see a pinned topic which deals with this known problem.
(I suffered the same with my new cruncher, and the solutions offered work, but you may have to try them one at a time to see which works best for you)
____________
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333670 - Posted: 1 Feb 2013, 21:55:16 UTC - in response to Message 1333432.

A whole pile of your results have ended in errors.
If you visit the "number crunching" message board you will see a pinned topic which deals with this known problem.
(I suffered the same with my new cruncher, and the solutions offered work, but you may have to try them one at a time to see which works best for you)


I suspect you were responding to Vito, but ... .

I only had 1 WU for opencl, and it errored out. Another has downloaded. I have not checked its status yet.

Is there of checking the REASON for the error? This WU.

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333765 - Posted: 2 Feb 2013, 2:23:49 UTC - in response to Message 1333670.
Last modified: 2 Feb 2013, 2:24:20 UTC

Clarification, for two words I left out:

Is there (a way) of checking the REASON for the error? This WU.

Profile Zapped SparkyProject donor
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 30 Aug 08
Posts: 7330
Credit: 1,237,633
RAC: 1,298
United Kingdom
Message 1333777 - Posted: 2 Feb 2013, 3:00:00 UTC

If you have a problem it would be much better to try to solve it over in the number crunching forum.
____________
In an alternate universe, it was a ZX81 that asked for clothes, boots and motorcycle.

Client error 418: I'm a teapot

Tropical Goldfish Fish 15: Squeaky bras 'R us

Illusions of normality sufferer

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333798 - Posted: 2 Feb 2013, 4:13:22 UTC - in response to Message 1333777.

If you have a problem it would be much better to try to solve it over in the number crunching forum.


Suggestions (link?) to a current thread dealing with such issues? Would rather put it in the "right place" than start a new thread.

Profile Zapped SparkyProject donor
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 30 Aug 08
Posts: 7330
Credit: 1,237,633
RAC: 1,298
United Kingdom
Message 1333810 - Posted: 2 Feb 2013, 4:32:23 UTC - in response to Message 1333798.

If you have a problem it would be much better to try to solve it over in the number crunching forum.


Suggestions (link?) to a current thread dealing with such issues? Would rather put it in the "right place" than start a new thread.

The right place is number crunching, just start a new thread.
____________
In an alternate universe, it was a ZX81 that asked for clothes, boots and motorcycle.

Client error 418: I'm a teapot

Tropical Goldfish Fish 15: Squeaky bras 'R us

Illusions of normality sufferer

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2642
Credit: 6,000,259
RAC: 4,280
Bulgaria
Message 1333820 - Posted: 2 Feb 2013, 4:54:20 UTC - in response to Message 1333798.

If you have a problem it would be much better to try to solve it over in the number crunching forum.


Suggestions (link?) to a current thread dealing with such issues? Would rather put it in the "right place" than start a new thread.

The problem is you don't follow the links.

How else you can explain your "Searched high and low for an older version of CCC. Found one."
after the 2 posts by Claggy with 3 proper links and explanations what to do?


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333890 - Posted: 2 Feb 2013, 9:39:43 UTC - in response to Message 1333820.

If you have a problem it would be much better to try to solve it over in the number crunching forum.


Suggestions (link?) to a current thread dealing with such issues? Would rather put it in the "right place" than start a new thread.

The problem is you don't follow the links.

How else you can explain your "Searched high and low for an older version of CCC. Found one."
after the 2 posts by Claggy with 3 proper links and explanations what to do?


Your response is NOT helpful.

I made it quite clear I was at work. That means this had low priority. Preparing for teaching a class or grading for it obviously MUST carry higher priority.

I followed TWO links provided by Claggy.
One was to AMDs spot for download of DRIVERS.
Hence, the blue screen of death I mentioned.
I disabled the GPU so that would stop and so other WUs could be crunched in the meantime, and until I could get another chance to try to sort things out.
Rather than using a Fix/Uninstall feature Claggy mentioned, I used "Roll back driver" when I followed CONTROL/SYSTEM/HARDWARE (as mentioned by another poster, and as I stated before).
The second link I followed was to the "HAL6000" cheat sheet. From that, I figured that I needed a CCC version that matched the driver version. I uninstalled the CCC I had, and downloaded and installed the earlier version that matched the driver.
Upon doing that, and re-enabling the GPU, then re-booting BAM, BAM indicated GPU computing was enabled and downloaded 1 WU for it to work on, among other non-GPU associated WUs.

I've been with S@H since '99, and BOINCing since '05. I've followed along with a number of tweaks, and usually get it right on the first attempt. If not, the second or third. Sometimes, the issue isn't even anything any BOINCer has ever encountered, yet I get it straightened out. But, I ask.

So, thank you so much for your cordial and informative (and WRONG) post (I followed TWO ... count 'em, TWO) of the links.
Pardon me if I am not going to stay around at the office for a few more hours to get it straightened out on day one, and pardon me for having a job that takes priority over tweaking BOINC to do more.

Profile Sarge
Volunteer tester
Avatar
Send message
Joined: 25 Aug 99
Posts: 9219
Credit: 1,410,520
RAC: 1,629
United States
Message 1333891 - Posted: 2 Feb 2013, 9:41:04 UTC - in response to Message 1333810.

If you have a problem it would be much better to try to solve it over in the number crunching forum.


Suggestions (link?) to a current thread dealing with such issues? Would rather put it in the "right place" than start a new thread.

The right place is number crunching, just start a new thread.


I get the first.
I don't like starting new threads if one already exists that might already be dealing with the same/similar issue(s).

Profile Raistmer
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 16 Jun 01
Posts: 3396
Credit: 46,355,973
RAC: 9,968
Russia
Message 1333899 - Posted: 2 Feb 2013, 10:08:15 UTC - in response to Message 1333891.
Last modified: 2 Feb 2013, 10:23:12 UTC

Proper thread to discuss issues with new app is release thread of that app.
Here is release thread for stock app.
Release thread for r1761 is here: http://setiathome.berkeley.edu/forum_thread.php?id=70723

EDIT: regarding error you got - no idea. Need debugger for that on your host perhaps. Also, try to run app offline (i.e., via bench for example) will it crash ?
If yes, then we can try with latest build (or very verbose build) and pdb file for it. Just report in rev1761 thread.

EDIT2: cause you got that error with stock app better to answer right here:

single pulses: 0
repetitive pulses: 30
percent blanked: 9.53
class T_remove_radar: total=3.91e+009, N=1, <>=3.91e+009, min=3.91e+009, max=3.91e+009
class T_main_loop_L1: total=5.04e+013, N=111, <>=4.54e+011, min=4.51e+011, max=4.58e+011
class T_FFT_forward: total=2.64e+010, N=909312, <>=2.91e+004, min=7.58e+003, max=3.25e+007
class T_remove_radar_randomize: total=1.94e+012, N=1817736, <>=1.07e+006, min=2.88e+002, max=2.75e+007
class T_build_chirp_table: total=0.00e+000, N=0, <>=0.00e+000, min=1.84e+019, max=0.00e+000
class T_DataWrite: total=1.82e+009, N=88800, <>=2.05e+004, min=3.27e+003, max=4.65e+005
class T_DataWrite_ns: total=0, N=0, <>=0, min=0 max=0
class T_oclReadBuf: total=3.14e+007, N=909312, <>=3.40e+001, min=2.40e+001, max=3.38e+005
class T_ChirpWrite: total=0.00e+000, N=0, <>=0.00e+000, min=1.84e+019, max=0.00e+000
class T_ChirpWrite_ns: total=0, N=0, <>=0, min=0 max=0
class T_dechirp: total=2.45e+010, N=909312, <>=2.70e+004, min=1.01e+004, max=6.90e+006
class Dechirp_ns: total=0, N=0, <>=0, min=0 max=0
class Half_ns: total=0, N=0, <>=0, min=0 max=0
class T_PC_single_pulse_kernel_FFA_update: total=4.83e+013, N=909312, <>=5.31e+007, min=5.16e+007, max=1.28e+008
class PC_ns: total=0, N=0, <>=0, min=0 max=0
class T_oclReadBuf: total=3.14e+007, N=909312, <>=3.40e+001, min=2.40e+001, max=3.38e+005
class T_oclWriteBuf: total=1.86e+009, N=88800, <>=2.09e+004, min=3.38e+003, max=4.66e+005
class T_FFT_inverse: total=1.12e+010, N=909312, <>=1.23e+004, min=5.61e+003, max=6.87e+006
class T_ffa: total=2.08e+009, N=1, <>=2.08e+009, min=2.08e+009, max=2.08e+009
class T_GPU_buffer_read_backs: total=0, N=0, <>=0, min=0 max=0
OCL_ZERO_COPY USE_OPENCL OPENCL_WRITE USE_INCREASED_PRECISION SMALL_CHIRP_TABLE COMBINED_DECHIRP_KERNEL
rev 1316


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x0040A1FA read attempt to address 0x00399E0C

Engaging BOINC Windows Runtime Debugger...


Scientific processing was finished completely. Error occured at some cleanup stage.

EDIT3: Cause I see time to time reports that AP crashed after processing finish I created special thread for such reports: http://setiathome.berkeley.edu/forum_thread.php?id=70742 you can subscribe on it to get updates on this issue.
____________

Previous · 1 . . . 6 · 7 · 8 · 9 · 10 · Next

Message boards : News : New AstroPulse applications for GPUs.

Copyright © 2014 University of California