Message boards :
Number crunching :
I'm getting lots and lots of 'Computation errors'!
Message board moderation
Author | Message |
---|---|
Send message Joined: 28 May 08 Posts: 2 Credit: 18,025,458 RAC: 0 |
My version of BOINC: boinc_6.10.60_windows_intelx86 My ATI graphics driver: 10-2_xp32_dd_ccc_wdm_enu My OS: Windows XP SP3 Application: MilkyWay@Home 0.59 (ati14) I have an ATI 4770 (512 megs) and recently (before that it was working perfectly fine), milkyway@home has been giving me numerous (actually, ALL my WUs were affected) computation errors when my ATI GPU was involved. The task would start but after 2 sec, it keeps giving a computation error message. I've tried resetting and reattaching the project but to no avail. I thought my graphics card was going bad but after testing it on 3DMark06, I found that it was OK. Is anybody else having the same problem as me? |
Send message Joined: 20 Sep 08 Posts: 1391 Credit: 203,563,566 RAC: 0 |
You need to run Catalyst Control Centre 11.3 for the new MW units. Don't drink water, that's the stuff that rusts pipes |
Send message Joined: 4 Oct 08 Posts: 1734 Credit: 64,228,409 RAC: 0 |
I think what Chris is saying is it is best to update and run the driver he mentioned, preferably the APP version. I am running Milkyway error free on an older driver - 10.11 - one ATI card has the APP driver included and the other has yet to have that installed. At least I am getting confirmed results, even if most of them are awaiting at pending ATM. @Chris What version of BOINC Manager are you running? I am still with 6.10.58. Go away, I was asleep |
Send message Joined: 24 Feb 09 Posts: 620 Credit: 100,587,625 RAC: 0 |
You are running 1.4.553 as your driver. Thats way too old, and must be updated. The new application is using the latest Driver elements, and your driver knows nothing of those, and therefore falls over. Go to AMD 4xxx XP 32 Bit APP Driver You need the one marked "Catalyst Software Suite", do not use the one without APP. Regards Zy |
Send message Joined: 20 Sep 08 Posts: 1391 Credit: 203,563,566 RAC: 0 |
@John - 1 x 6.10.60, 5 x 6.10.58 Don't drink water, that's the stuff that rusts pipes |
Send message Joined: 28 May 08 Posts: 2 Credit: 18,025,458 RAC: 0 |
Thanks to everyone for the helpful replies. I've updated the graphics driver to ver. 11.3 and my GPU is processing happily now! :-) Now I just have to wait for the processed WUs to be validated. BTW, is the current default app the same as the optimised app released a few years ago? (mentioned in the sticky thread 'Optimised Apps') |
Send message Joined: 24 Feb 09 Posts: 620 Credit: 100,587,625 RAC: 0 |
Current Stock app is the best available - the Opt App page not yet updated - too much happening at admin end. So looks like you are good to go, happy crunching :) Regards Zy |
Send message Joined: 6 Apr 11 Posts: 7 Credit: 59,288,856 RAC: 0 |
I'm not sure what changed but I cannot upgrade my drivers to accommodate this projects new code. The new 11.3 drivers are absolutely terrible for my 4870x2. I tried them and they work for the project, but if I want to play any games on my machine, it runs like crap, stuff look weird, etc. As a matter of a fact, anything beyond 10.5 hotfix drivers are garbage for a 4870x2. I do hope the developers can accommodate those of us with older hardware and drivers. Before the big server upgrade and database wipe, this project worked excellent on my setup. If anyone has any idea on how to get APP to work with 10.5 hotfix drivers, let me know. Thanks. |
Send message Joined: 17 Aug 10 Posts: 1 Credit: 14,329,307 RAC: 0 |
I will just detach from project. Had enough of all these problems. My idle time will be spent on some other project from this moment on. Too bad, cuz I liked MW and it used to run smoothly without any problems. |
Send message Joined: 7 Jan 08 Posts: 12 Credit: 12,600,231 RAC: 37 |
Is something broken with the stock MAC application? |
Send message Joined: 8 May 10 Posts: 576 Credit: 15,979,383 RAC: 0 |
Is something broken with the stock MAC application?The 0.31 one is getting pretty old (and is supposed there since I've been lazy and not built a 32 bit version or universal binary) and might be what's broken. I suppose I should get around to fixing that. |
Send message Joined: 12 Aug 09 Posts: 262 Credit: 92,631,041 RAC: 0 |
I will just detach from project. Had enough of all these problems. My idle time will be spent on some other project from this moment on. Too bad, cuz I liked MW and it used to run smoothly without any problems. And it will do again. Have a little patience as we all had and give some input for the developers to adjust the software. All projects have there issues from time to time. Greetings from, TJ |
Send message Joined: 29 Jun 08 Posts: 10 Credit: 50,922,967 RAC: 0 |
OS: Windows XP 32Bit SP3 CPU: Athlon XP2400 BOINC: 6.10.60 Application: MilkyWay@Home 0.62 (ati14) ATI graphics driver: 11.3 ATI HD3850 (AGP) Link to host: Link to Host Card worked flawlessly before. Now only errors. I had the driver reinstalled during the validation failure dilemma. From the error messages i assume DLL problem. Anyone an idea what be wrong? |
Send message Joined: 8 May 10 Posts: 576 Credit: 15,979,383 RAC: 0 |
OS: Windows XP 32Bit SP3Since the database clears things out way too quickly, I don't see any problems with the current version there. |
Send message Joined: 15 Jul 08 Posts: 383 Credit: 729,293,740 RAC: 0 |
Since the database clears things out way too quickly, I don't see any problems with the current version there. This is a problem. Is there a way you can have invalid WUs stay in the database longer so we can at least see if any of our machines are having problems? As it is now a box can be throwing errors with no way for us to know it. Edit: Also have caught a few invalids that say there were too many invalid results and the WU might be bad. This level was set too low IMO and with the number of machines erroring out WUs it looks like some good WUs are getting this message because by chance they're sent out to 3 machines that aren't returning good results and 1 that is. |
Send message Joined: 29 Jun 08 Posts: 10 Credit: 50,922,967 RAC: 0 |
Below are the errors i get. Any one with an idea is more than welcome. What i already did till now: - Reinstalled Catalyst 11.3 - Reinstalled Stream SDK 2.3 - Windowsupdate (which updated .NET and VC) - reseted project several times The job runs till the end and then errors out without an error message in BOINC. So it doesn't stop immediately at the beginning. Errormessages from WU details: <core_client_version>6.10.60</core_client_version> <![CDATA[ <message> - exit code -1073741795 (0xc000001d) </message> <stderr_txt> <search_application> milkywayathome_client separation 0.62 Windows x86 double CAL++ </search_application> Found 1 CAL devices Chose device 0 Device target: CAL_TARGET_670 Revision: 41 CAL Version: 1.4.1332 Engine clock: 669 Mhz Memory clock: 829 Mhz GPU RAM: 512 Wavefront size: 64 Double precision: CAL_TRUE Compute shader: CAL_FALSE Number SIMD: 4 Number shader engines: 1 Pitch alignment: 256 Surface alignment: 256 Max size 2D: { 8192, 8192 } Estimated iteration time 573.033445 ms Target frequency 30.000000 Hz, polling mode 1, using responsiveness factor of 1.000000 Dividing into 20 chunks Integration range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 } Using { 1, 20 } chunk(s) of size { 1400, 80 } Integration time = 545.853849 s, average per iteration = 852.896639 ms Unhandled Exception Detected... - Unhandled Exception Record - Reason: Illegal Instruction (0xc000001d) at address 0x004051F9 Engaging BOINC Windows Runtime Debugger... ******************** BOINC Windows Runtime Debugger Version 6.10.58 Dump Timestamp : 04/15/11 11:13:20 Install Directory : F:\BOINC\Program\ Data Directory : F:\BOINC\Data Project Symstore : Loaded Library : F:\BOINC\Program\\dbghelp.dll Loaded Library : F:\BOINC\Program\\symsrv.dll Loaded Library : F:\BOINC\Program\\srcsrv.dll LoadLibraryA( F:\BOINC\Program\\version.dll ): GetLastError = 126 Loaded Library : version.dll Debugger Engine : 4.0.5.0 Symbol Search Path: F:\BOINC\Data\slots\1;F:\BOINC\Data\projects\milkyway.cs.rpi.edu_milkyway ModLoad: 00400000 0008f000 F:\BOINC\Data\projects\milkyway.cs.rpi.edu_milkyway\milkyway_0.62_windows_intelx86__ati14.exe (-nosymbols- Symbols Loaded) Linked PDB Filename : ... bunch of DLL references ... *** Dump of the Process Statistics: *** - I/O Operations Counters - Read: 9, Write: 0, Other 1075 - I/O Transfers Counters - Read: 0, Write: 465, Other 0 - Paged Pool Usage - QuotaPagedPoolUsage: 67352, QuotaPeakPagedPoolUsage: 72936 QuotaNonPagedPoolUsage: 24000, QuotaPeakNonPagedPoolUsage: 24320 - Virtual Memory Usage - VirtualSize: 113598464, PeakVirtualSize: 170954752 - Pagefile Usage - PagefileUsage: 16211968, PeakPagefileUsage: 73465856 - Working Set Size - WorkingSetSize: 2232320, PeakWorkingSetSize: 74489856, PageFaultCount: 20270 *** Dump of thread ID 2732 (state: Ready): *** - Information - Status: Base Priority: Unknown, Priority: Unknown, , Kernel Time: 36562500.000000, User Time: 32343750.000000, Wait Time: 69675.000000 - Unhandled Exception Record - Reason: Illegal Instruction (0xc000001d) at address 0x004051F9 - Registers - eax=003f6c20 ebx=00e1f874 ecx=003f6f10 edx=00000002 esi=003f6c20 edi=00000000 eip=004051f9 esp=00e1f838 ebp=00e1f860 cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000 efl=00010212 - Callstack - ChildEBP RetAddr Args to Child 00e1f860 004053b7 00000001 025619e1 00000010 00e1f8d8 milkyway_0.62_windows_intelx86_!+0x0 00e1f8d0 00405ccd 00000000 00405ccd 00e1f998 003f6c20 milkyway_0.62_windows_intelx86_!+0x0 00e1f96c 00405d99 00e1fc00 003f6c20 003f6e00 00e1fb98 milkyway_0.62_windows_intelx86_!+0x0 00e1fa00 00403ea0 00e1fc00 003f6e00 00e1fb98 00e1fa60 milkyway_0.62_windows_intelx86_!+0x0 00e1fb14 0040419b 00e1fc00 003f6c20 003f6e00 003f5580 milkyway_0.62_windows_intelx86_!+0x0 00e1fb54 00401661 003f3050 00e1fc00 003f6c20 00e1fbb8 milkyway_0.62_windows_intelx86_!+0x0 00e1fd80 0040185f 0047c200 0043dcbe 60c721ec 003f2eb0 milkyway_0.62_windows_intelx86_!+0x0 00e1ff78 0043e203 00000014 003f32c8 003f3520 60c72390 milkyway_0.62_windows_intelx86_!+0x0 00e1ffc0 7c817077 00000074 00000000 7ffdd000 c000001d milkyway_0.62_windows_intelx86_!+0x0 00e1fff0 00000000 0043e259 00000000 00000000 00000000 kernel32!RegisterWaitForInputIdle+0x0 *** Dump of thread ID 2984 (state: Waiting): *** - Information - Status: Wait Reason: ExecutionDelay, , Kernel Time: 0.000000, User Time: 0.000000, Wait Time: 69673.000000 - Registers - eax=0430e234 ebx=00000000 ecx=00000000 edx=0255ffac esi=00000000 edi=0255ff60 eip=7c90e514 esp=0255ff30 ebp=0255ff88 cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000 efl=00000206 - Callstack - ChildEBP RetAddr Args to Child 0255ff88 7c802455 00000064 00000000 0255ffb4 00431e04 ntdll!KiFastSystemCallRet+0x0 0255ff98 00431e04 00000064 004342f0 003f2e28 00000000 kernel32!Sleep+0x0 0255ffb4 7c80b729 00000000 004342f0 003f2e28 00000000 milkyway_0.62_windows_intelx86_!+0x0 0255ffec 00000000 00431df0 00000000 00000000 00905a4d kernel32!GetModuleFileNameA+0x0 *** Debug Message Dump **** *** Foreground Window Data *** Window Name : Window Class : Window Process ID: 0 Window Thread ID : 0 Exiting... </stderr_txt> ]]> |
Send message Joined: 29 Jun 08 Posts: 10 Credit: 50,922,967 RAC: 0 |
Now did a rollback of BOINC to 6.10.58 because of this LoadLibraryA( F:\BOINC\Program\\version.dll ): GetLastError = 126 also didn't help. Btw Collatz currently runs without problems on this host. Is it possible that the binary is complied with SSE2 enabled? The AthlonXP doesn't have this but then i would have expected the WU to fail right away and not run till the end and then error out. |
Send message Joined: 20 Sep 08 Posts: 1391 Credit: 203,563,566 RAC: 0 |
Double precision: CAL_TRUE I think the message above is a clue ? If this is an 3850 AGP card you may need to run the CCC Hotfix version. |
Send message Joined: 29 Jun 08 Posts: 10 Credit: 50,922,967 RAC: 0 |
Double precision: CAL_TRUE Did this. The Hofix version is installed. Strange though why the program would run for 10 minutes with progress bar changing if he doesn't find a compute shader? |
Send message Joined: 20 Sep 08 Posts: 1391 Credit: 203,563,566 RAC: 0 |
Indeed, I'm sure our resident Expert Matt will have more of an an idea. |
©2024 Astroinformatics Group