| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: BOINC 7.0.64/7.0.65 released for the public |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 50
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You'd really need to ask that question in the GPU support forum. Don't know... new drivers installing requires most often the uninstalling of the old. As for I5 internal GPU HD3000... not even sure it was ever supported. Only recently was latest client able to see the HD 4000... of course, OpenCL needs installing for Intel and AMD/ATI.
|
||
|
|
LuNaTiC-GR
Cruncher Greece Joined: Sep 21, 2008 Post Count: 9 Status: Offline Project Badges: 
|
Ok ill make the question again @ gpu forum.
----------------------------------------I see posters before mine have the symbol B for beta testing... i wanna ask i never get something for beta test even i enable that option at homepage of WCG ... is there needed to do something else or just happend? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
These forum are very lenient. Practically no moderation [a few fulminate at any moderation, where 4 letter words definitely get you a mod.edit, no matter how important anyone thinks themselves]. Only rare attempts are made to enforce 'on topic'. Most other places, moved or archived or closed, if a repeat. This OT really belongs in the Beta forum ;>), but here is a nutshell answer...
----------------------------------------Beta tests arrive by themselves [when there is one] if enrolled and client asking for [any] work at the right time. Given the limited number in a test [from 1000-5000 tasks typically] it usually passes by most without ever knowing. Oh and it helps to be in the USA or stay up very late [you being in Greece and me in Italy leads mostly to being too late to force the hand]. Those tests most often start like the 11th hour of the afternoon, and on Friday would you believe. Chances may have increased for a wider audience since the enforced backoff for those hammering the Update button, having set the profile to practically only get BETA stream, to now 121 seconds. Was 11 seconds and for a while 31 seconds.[/OT] [Edit 1 times, last edit by Former Member at May 12, 2013 3:57:33 PM] |
||
|
|
LuNaTiC-GR
Cruncher Greece Joined: Sep 21, 2008 Post Count: 9 Status: Offline Project Badges:
|
Thank you
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
A number of users have [for Linux], reported a bug at the devs mail list... the GPU not commencing computing when set to start after being idle for X minutes. When the bug was introduced has not been determined yet, but previous general port [7.0.27] did not have the issue, repeat, on Linux [USB keyboard/mouse detection problem suspected]. Workaround: Downgrade or set Run GPU Always when you know you're going to be away.
As for the "Doing the dirty" series, chapter 3, continue running GFAM/DSFL 50/50 with help of app_config.xml and one device is perilously close to getting an idle core. Oldest for GFAM is now due in 4 days, oldest for DSFL is due in 8.25 day i.e. then number for GFAM arriving is continuing to be higher. Why intervention has as yet not been needed is that the DSFL tasks have been inserting a number of long once. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Chapter 4 of, well, it's really dirty dancing with the feeders and the joystick of the techs [how they set the balance], sets the ratios.
----------------------------------------Linux host: 2 GFAM running, 57 buffered for 10.4 days 2 DSFL running, 1 buffered for 3:45 hours W7-32 Duo: 1GFAM running, 8 buffered for 2.75 days 1 DSFL running, 1 buffered for 10.8 hours W7-64 Heptagon [1 core disabled] 3 GFAM running, 53 buffered for 13.75 days 3 DSFL running, 7 buffered for 2.5 days 1 CEP2 running, 2 buffered for... 1.16 days Started the app_config.xml enforced experiment of 50/50 for the 2 sciences on May 8 and now May 17 hosts 1 and 2 are getting perilously close to an idle threads situation, particular the Linux box. Time to switch on the external 'back-up' projects, those set to zero share but having work fetches soon as cores go idle... so is the theory with 7.0.64/65... something else to proof if working or broken... and fetch is partially broken we know now of 7.0.xx in a mixed CPU/GPU environ, supposedly fixed now in the first out 7.1.1 but that is really really alpha... I'll be testing soon but it is *not* supported in production, at all!. Revert to devs at Berkeley for comments/discussion. (Honestly if it is fixed in 7.1.1 it's a strong canditate for backporting. We'll have to live for longer with 7.0.xx before a next 'recommended' will be considered.) edit: closing ) [Edit 1 times, last edit by Former Member at May 17, 2013 11:17:17 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Chapter 5, a short one, and it's happened... no more DSFL on the Linux box, and slots now filled with SIMAP+Malariacontrol as zero share backup projects. An overabundance of GFAM waiting to make use of the 2 core slots they're allowed to use. Will this continue to work? We'll see.
----------------------------------------edit: P.S. That's at the point where 9 days 22 hours worth of GFAM remained, 55 tasks. [Edit 1 times, last edit by Former Member at May 17, 2013 2:51:48 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Chapter 6 is sad, because waking up to idling cores is not good. -No- DSFL on the Quad + Heptagon and somehow the backup projects had not been called on either [a 7.0.64/65 scheduler bug?]. No, actually there was so much GFAM that the buffers were simply full to the setting [2.5 days] and then backup projects are not being asked either. Not Spock logical, to not do so with cores idling. Anyway, intervened, hit My Grid > My Project and will let the DSFL feeder have exclusivity for a day or 2 until GFAM overstock is depleted.
BTW, the backups have in past week accumulated well over 4.5 days of computing time due DSFL undersupply [in the chosen mix]. Think this could be called WCG's loss... SIMAP's and Malariacontrol's gain. |
||
|
|
branjo
Master Cruncher Slovakia Joined: Jun 29, 2012 Post Count: 1892 Status: Offline Project Badges: 
|
I now have HD 4000 recognized: 05-May-2013 21:08:48 [---] Running under account Jóhannes 05-May-2013 21:08:48 [---] Processor: 8 GenuineIntel Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz [Family 6 Model 58 Stepping 9] 05-May-2013 21:08:48 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 popcnt aes syscall nx lm vmx tm2 pbe 05-May-2013 21:08:48 [---] OS: Microsoft Windows 8: Professional with Media Center x64 Edition, (06.02.9200.00) 05-May-2013 21:08:48 [---] Memory: 14.96 GB physical, 29.96 GB virtual 05-May-2013 21:08:48 [---] Disk: 465.42 GB total, 217.88 GB free 05-May-2013 21:08:48 [---] Local time is UTC +0 hours 05-May-2013 21:08:48 [---] CUDA: NVIDIA GPU 0: GeForce GTX 650 Ti (driver version 314.22, CUDA version 5.0, compute capability 3.0, 2048MB, 1973MB available, 1586 GFLOPS peak) 05-May-2013 21:08:48 [---] OpenCL: NVIDIA GPU 0: GeForce GTX 650 Ti (driver version 314.22, device version OpenCL 1.1 CUDA, 2048MB, 1973MB available, 1586 GFLOPS peak) 05-May-2013 21:08:48 [---] OpenCL: Intel GPU 0: Intel(R) HD Graphics 4000 (driver version 9.17.10.2932, device version OpenCL 1.1, 728MB, 728MB available, 45 GFLOPS peak) 05-May-2013 21:08:48 [---] Config: use all coprocessors presumably the intel GPU was in a sleep mode or something all I had to was to connect the monitor to the DIVI input for the hd4000 card at the back of my computer ! thanks Rob for your help and all your valuable input. Johannes Thanks jgis  [ot]btw: have you already received (and crunched) any intel_gpu WU?[/ot] [ot]Collatz has OpenCL intel_gpu WU's (solo_collatz), running them w/o problems concurrently with OpenCL ati WU's[/ot] Cheers   Crunching@Home since January 13 2000. Shrubbing@Home since January 5 2006  |
||
|
|
Jord
Advanced Cruncher Joined: Dec 30, 2005 Post Count: 148 Status: Offline Project Badges: 
|
Just download the new version and it seems i cant get a gpu wu to analyze . Why is that happening and how i can enable my i5 internal gpu to work together with my discrete one ? The Intel HD 3000 is a graphics processing unit, but not a general purpose graphics processing unit. GPU but not GPGPU. Meaning that it can show graphics, but cannot be used for OpenCL calculations. Thanx in advance *i dled new nvidia + hd 3000 drivers before 10mins... still no luck on working :( Thus far only the Intel HD2500 and HD4000 graphics chips have OpenCL support. These can be found on Ivy Bridge i5 and i7 CPUs. Intel HD 2000 and 3000 on Sandy Bridge i5 and i7 CPUs is not supported and will not be supported either. (Sekerob) I noted it's only the beginning and per their official Wiki updated through April 3, 2013 it says... Thanks, I knew there were places that still had to be adjusted. Just did that.
Tears in my eyes
How they fall like rain to the floor |
||
|
|
|