Message boards :
News :
New Separation Runs
Message board moderation
Author | Message |
---|---|
Send message Joined: 10 Apr 19 Posts: 408 Credit: 120,203,200 RAC: 0 |
Hey Everyone, I'm taking down the following runs: de_modfit_70_bundle5_3s_south_pt2_2 de_modfit_71_bundle5_3s_south_pt2_2 de_modfit_72_bundle5_3s_south_pt2_2 de_modfit_73_bundle5_3s_south_pt2_2 de_modfit_74_bundle5_3s_south_pt2_2 de_modfit_75_bundle5_3s_south_pt2_2 de_modfit_76_bundle5_3s_south_pt2_2 de_modfit_77_bundle5_3s_south_pt2_2 de_modfit_78_bundle5_3s_south_pt2_2 de_modfit_79_bundle5_3s_south_pt2_2 de_modfit_80_bundle5_3s_south_pt2_2 de_modfit_81_bundle5_3s_south_pt2_2 de_modfit_82_bundle5_3s_south_pt2_2 de_modfit_83_bundle5_3s_south_pt2_2 de_modfit_84_bundle5_3s_south_pt2_2 de_modfit_85_bundle5_3s_south_pt2_2 de_modfit_86_bundle5_3s_south_pt2_2 de_modfit_70_bundle5_3s_south_pt2_3and putting up the following runs: de_modfit_70_bundle5_3s_south_pt2_4 de_modfit_71_bundle5_3s_south_pt2_4 de_modfit_72_bundle5_3s_south_pt2_4 de_modfit_73_bundle5_3s_south_pt2_4 de_modfit_74_bundle5_3s_south_pt2_4 de_modfit_75_bundle5_3s_south_pt2_4 de_modfit_76_bundle5_3s_south_pt2_4 de_modfit_77_bundle5_3s_south_pt2_4 de_modfit_78_bundle5_3s_south_pt2_4 de_modfit_79_bundle5_3s_south_pt2_4 de_modfit_80_bundle5_3s_south_pt2_4 de_modfit_81_bundle5_3s_south_pt2_4 de_modfit_82_bundle5_3s_south_pt2_4 de_modfit_83_bundle5_3s_south_pt2_4 de_modfit_84_bundle5_3s_south_pt2_4 de_modfit_85_bundle5_3s_south_pt2_4 de_modfit_86_bundle5_3s_south_pt2_4This new batch can be identified with the "_4" at the end rather than "_2" or "_3". As always, you'll see WUs from the old runs for the next week or so as those runs terminate. Please let me know if you experience any issues with the new runs. Best, Tom |
Send message Joined: 12 Jun 10 Posts: 57 Credit: 6,174,408 RAC: 41 |
Thanks for the update Tom. Is there a way to monitor progress of the project? |
Send message Joined: 11 Jan 19 Posts: 1 Credit: 237,183 RAC: 0 |
Hello , There's an error in the task code exec , i have been forced to cancel all my tasks and won't accept new tasks until you can ensure it's fixed. Greetings |
Send message Joined: 24 Jan 11 Posts: 715 Credit: 555,873,019 RAC: 42,416 |
You are mistaken. The current tasks are fine. Nobody but you has any issues. |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
You are mistaken. The current tasks are fine. Nobody but you has any issues. ... excellent answer - short + clear ... |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
Hello , ... nice posting ! what kind of error ? what forces you to cancel them ? my GPU separation tasks are doing fine - except for the ones running at the moment on my old gtx750Ti ! for some reason they are getting "computation error" cheers |
Send message Joined: 18 Nov 22 Posts: 84 Credit: 640,530,847 RAC: 0 |
it pays to look at the specific error to understand whats happening or might be happening. your errors are "Using device 4 on platform 0 Found 4 CL devices Requested device is out of range of number found devices Failed to select a device (1): MW_CL_ERROR" this host also reports five (5) GPUs. does it actually have 5? or is a GPU spoofing technique in play to get more tasks? if spoofing, then this issue could be that you are not using a max_concurrent statement to limit running tasks. if you don't do this, then BOINC will think there are devices available to run tasks that dont truly exist and you can get error like this. if you're not spoofing, then it's likely a problem with the drivers. a quick check could be to simply remove and reinstall the drivers. |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
Thanks for your help . Appreciate your time to look into the error reasons. NO spoofing - never ever ... The old lady has real 5 GPUs. Option max_concurrent is nicely set to 5. They have been runnung for weeks ok. The driver I'm using is the newest/last one that supports XP. Everything seems fine. No overheating. Memory is ok for both CPU + GPU. CPUs are running way less than 100%. Just the GPUs are at full 100% - which never had been a problem on all my rigs. Have installed the latest XP supported driver several times - of course with "clean" install. Probably (?) one of the 750Ti dying or losing interest to crunch? After restarting everything is OK. I noticed that the Internet connection had gone way down south ... I'll keep an eye on it. I'm completely at loss here - "nice" new feeling for me. cheers S-F-V |
Send message Joined: 28 May 17 Posts: 76 Credit: 4,398,910,125 RAC: 156 |
You really should consider upgrading to Window 10 at the least or a modern distro of Linux. Windows XP is extremely old. Not saying it is the root cause of the errors, but it definitely could be. |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
You really should consider upgrading to Window 10 at the least or a modern distro of Linux. In addition several Projects have stopped making apps that work with Windows XP and earlier meaning his choice of Project to use the gpu to crunch for. |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
You really should consider upgrading to Window 10 at the least or a modern distro of Linux. I don't know if you noticed that I have eight (8) tiny rigs running on Win10? This "test" XP is my little hobby - it works fine - since years. Fast and nice. I probably messed something up while trying out a few changes. I reversed some of those "tuning" bits ... So now the little old and shabby XP with the great and fast GTX750Tis (real 5 of them) is again churning away happily. So, anyway, many thanks for the tips. cheers S-F-V |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
You really should consider upgrading to Window 10 at the least or a modern distro of Linux. +1 Yep - I choose my projects for my test rig using my own self-determined decision/will. So it's basically somewhat fun for me to find out what "still" works or not! |
Send message Joined: 19 Jul 10 Posts: 627 Credit: 19,303,095 RAC: 1,183 |
Probably (?) one of the 750Ti dying or losing interest to crunch?My GTX 275 needs a reboot every 3-4 weeks, like yesterday for example, otherwise it starts generating lots of errors, mostly (-6) CL_OUT_OF_HOST_MEMORY or sometimes something with the number (-30). It would still crunch Moo! without any issues in that condition, but it doesn't like those bundles here, so it messes up the loading of last 2-3 WUs out of the 5 we get here as a bundle. |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
That's interesting as I also have 5+ 750Ti gpu's but mine are sitting on a shelf with no pc to put them in, I got them cheap, all but 1 are 1gb cards while the 1 has 2gb of onboard memory. I bought mine to put into a mining rig but the rig was too big for where I wanted to put it so I never used it. |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
Well, I have about 10 750s + 8 960s. I'm using the ASUS Nvidia Strix 750TIs (all 4GB) on an ASROCK Z68 Extreme4 Gen3 board (of which I have 4 spares). This board I like, because it has a "real" hardware FLOPPY connection. Which I use to play around in VBox insatlling all the old OS2s and win3.1 and the whole lot of whatever. It's fun for me!! Separation task takes about 11min on the 750s - so with 5 I get a nice throughput ... -- so who needs those expensive + big GPUs? ;-) ... The are nice and quiet and don't get hot (never over 60C). The ASUS GTX960s (4GB) are more or less all over the other rigs. cheers S-F-V |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
Probably (?) one of the 750Ti dying or losing interest to crunch?My GTX 275 needs a reboot every 3-4 weeks, like yesterday for example, otherwise it starts generating lots of errors, mostly (-6) CL_OUT_OF_HOST_MEMORY or sometimes something with the number (-30). It would still crunch Moo! without any issues in that condition, but it doesn't like those bundles here, so it messes up the loading of last 2-3 WUs out of the 5 we get here as a bundle. Well, the 750 GPUs seem to be OK. Like you say a restart did it. I will watch the time span that you mentioned. Strange things happen in outer space. cheers S-F-V |
©2024 Astroinformatics Group