Welcome to MilkyWay@home

Posts by Matt Arsenault

21) Message boards : Number crunching : What does Could not load Ktm32.dll (126) mean? (Message 53975)
Posted 11 Apr 2012 by Matt Arsenault
Post:
Yes you can. It isn't a fatal issue. Your tasks are completing fine.
22) Message boards : Number crunching : What does Could not load Ktm32.dll (126) mean? (Message 53973)
Posted 10 Apr 2012 by Matt Arsenault
Post:
That will happen if you're running on Windows older than Vista. It tries to use transactions to ensure that updating checkpoints won't break if you lose power at the wrong time or something like that, but on Windows you couldn't do that until vista.
23) Message boards : Number crunching : DLL error when attaching to MilkyWay (Message 53919)
Posted 5 Apr 2012 by Matt Arsenault
Post:
The way boinc deals with multifile applications / ones that ship libraries is kind of stupid. It needs to have a name unique to files ever used by the project, but that won't correspond to the actual name of the library. It's supposed to be copying the wrong named dll to the working directory in the slot with the actual name. You seem to be using a BOINC 7.something. Maybe it broke in that release?
24) Message boards : Number crunching : Work done never changes (Message 53819)
Posted 27 Mar 2012 by Matt Arsenault
Post:
You're using an ancient version which doesn't send results back correctly so it can never validate. You need to use a newer version.
25) Message boards : Number crunching : Validate error on all WU of 2 new computers (Message 53804)
Posted 26 Mar 2012 by Matt Arsenault
Post:
You're using the anonymous platform with a really ancient version which won't work anymore. You can either just not use app_info and get the current version, or you have to update what you're using manually.
26) Message boards : Number crunching : gpu app (Message 53790)
Posted 25 Mar 2012 by Matt Arsenault
Post:
3/25/2012 7:03:01 PM | | ATI GPU 0: AMD Radeon HD 6x00 series (Turks) (CAL version 1.4.1664, 1024MB, 864 GFLOPS peak)

This doesn't have doubles so it won't work.
27) Message boards : Number crunching : Done WUs disappearing? (Message 53728)
Posted 21 Mar 2012 by Matt Arsenault
Post:
Results get purged from the database after some period (which I think right now is set to 2 days) so they disappear from the stats pages.
28) Message boards : Number crunching : "WU Freezes BOINC Manager" Redux... (Message 53428)
Posted 25 Feb 2012 by Matt Arsenault
Post:
BOINC freezing isn't really something milkyway can do. Can you try using the sample process thing in Activity Monitor on the boinc processes when it does happen (it's there on Lion)?
29) Message boards : Number crunching : All work Units giving "Computational Error" (Message 53424)
Posted 25 Feb 2012 by Matt Arsenault
Post:
After a 9mth hiatus I thought I'd give MW a shot again but I'm am also getting every single WU erroring out :(.

Had to update drivers so it's now on Cat 12.1, updated BOINC to 6.12.34 (btw I also tried 7.0.18 but it couldn't detect my GPU!, so went back to v6).
Your driver installation is missing the OpenCL dll. The catalyst installer apparently offers an option to not install it (why I have no idea), but reinstall your drivers without unchecking the install APP SDK option in the custom install part.
30) Message boards : Number crunching : getting errors with new v1.02 separation application? (Message 53379)
Posted 22 Feb 2012 by Matt Arsenault
Post:
Error getting number of platform (-1001): CL_PLATFORM_NOT_FOUND_KHR
Something is wrong with your driver installation, but it looks like you've fixed it now since you have tasks succeeding.
31) Message boards : Number crunching : getting errors with new v1.02 separation application? (Message 53352)
Posted 20 Feb 2012 by Matt Arsenault
Post:
actually i tried both <cmdline> --device 0</cmdline> and <cmdline> --device 1</cmdline>. nevertheless, its probably worth a try to remove both <cmdline> --device x</cmdline> from the app_info.xml and <ignore_ati_dev>x</ignore_ati_dev> from the cc_config.xml for now, and add what you suggested to the cc_config.xml file. it appears i have a few more things to experiment with, so i'll try to get on it tonight and report back as soon as i can...

thanks,
Eric
I should have said --device 0 before. Removing that and using BOINC 7 should work
32) Message boards : News : Separation updated to 1.00 (Message 53349)
Posted 20 Feb 2012 by Matt Arsenault
Post:
I haven't tried 290 anything yet
33) Message boards : News : Separation updated to 1.00 (Message 53346)
Posted 20 Feb 2012 by Matt Arsenault
Post:
[quoteWell Ok I see someone else has gone all Beta or maybe that should be Alpha, I'm getting 97-98% cpu use(295) and about 78%(285) while doing a gpu app.[/quote]It's yet another Nvidia driver problem introduced in 270.something
34) Message boards : Number crunching : Inactive Processes in Task Manager after Closing BOINC (tasks cannot be killed) (Message 53315)
Posted 19 Feb 2012 by Matt Arsenault
Post:
How often does this happen? When did it start happening?

Once in a while I see them get stuck in the drivers but it's pretty rare (except on the 7970 where DWM crashes every few hours and tasks get stuck)
35) Message boards : Number crunching : HIGH CPU USAGE with new 1.02 OpenCL tasks (Message 53312)
Posted 19 Feb 2012 by Matt Arsenault
Post:
The workaround is enabled if you're using Nvidia drivers newer than 270.xx or AMD drivers that are 11.7 or 11.8. It seems to have come back again for AMD.

The problem is they are busy waiting when you use clFinish or clWaitForEvents to wait for the GPU to complete running things.
36) Message boards : Number crunching : tasks being sent to wrong gpu card (Message 53311)
Posted 19 Feb 2012 by Matt Arsenault
Post:
The GTX460 definitely does. The fraction just means that is how fast it is compared to single.

The multi GPU support in BOINC can only be described as completely broken, which is why I think you need to create the cc_config.xml with the <use_all_gpus> or whatever it is option to even do so. Your GPUs get folded into whatever BOINC decides is the "most capable" GPU which is completely wrong.

You can either use app_info with the --device argument (looks like it should be <cmdline>--device 1</cmdline> in your case if you are talking about this system http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=285405) or use the project device exclusion in cc_config.xml.
37) Message boards : Number crunching : All work Units giving "Computational Error" (Message 53309)
Posted 19 Feb 2012 by Matt Arsenault
Post:
The checkpoint file update is done using transactions if it's available on the system so that you don't lose a checkpoint in the event of a power failure at the wrong moment or anything like taht.

I remember checking and adding a fallback if the FS didn't support transactions but it looks like it's not there now. I'll readd that I guess but moving it to an NTFS partition would avoid the problem.
38) Message boards : Number crunching : getting errors with new v1.02 separation application? (Message 53307)
Posted 19 Feb 2012 by Matt Arsenault
Post:
The GPU exclusion should also work with 6.12
39) Message boards : Number crunching : getting errors with new v1.02 separation application? (Message 53306)
Posted 19 Feb 2012 by Matt Arsenault
Post:
The reason is how BOINC handles device indexing. If you look the first one is using BOINC 7 and the second one with the error is using 6.12.34. Reupgprade to a BOINC 7 (I think 7.0.15 is the newest), or since you are using app_info already you could add <cmdline> --device 0</cmdline>to force it to use that GPU.

You have a kind of weird case where you have 2 GPUs that support both CAL but only 1 supports OpenCL.

The 4290 is based on an R600 core and doesn't support OpenCL. The older version of BOINC will give a device index of 1 to use the other GPU based on the CAL detection which would include that. The OpenCL device BOINC provides in 7 is correct and uses the OpenCL capable GPU.
40) Message boards : Number crunching : All work Units giving "Computational Error" (Message 53222)
Posted 16 Feb 2012 by Matt Arsenault
Post:
Ever since 1.00 update been getting comp errors.

AMD Phenom II x4 965
ATI 4890 x 2 crossfire - CCC driver 11.9
Win 7 64-bit
BOINC 6.12.34 9x64)
Was chugging along fine before then.
All 1.02 openCl_amd_ati are failing at the 1:07 ; 1:08 mark.

Tried that Arkyan file - it broke Milkyway.

2/16/2012 7:29:24 AM | Milkyway@Home | Message from server: Your app_info.xml file doesn't have a usable version of MilkyWay@Home N-Body Simulation.

Suspending project until fix to prevent masses of comp errors.
Is your BOINC data directory on some weird filesystem or a network share or something?


Previous 20 · Next 20

©2021 Astroinformatics Group