WARNING: This website is obsolete! Please follow this link to get to the new Albert@Home website!
Running on ATI |
Message boards :
Problems and Bug Reports :
Running on ATI
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
Oliver Behnke Volunteer moderator Project administrator Project developer Send message Joined: 4 Sep 07 Posts: 130 Credit: 8,545,955 RAC: 0 |
Thanks, that's exactly our focus right now. I mean, we not only had to implement the OpenCL version of our algorithm but also update the server backend as well as the client requirements since we are the first project deploying a "real" (fully BOINC compliant) OpenCL app on all three major platforms (Win, Mac, Lin). This is not straight-forward, in particular when AMD's drivers come into play... Again, we first aim at supporting AMD GPUs at all, then we'll start optimizing the app. And hey, our test rigs show a performance drop of 20% only, compared to our CUDA app - not that bad of the first release (but you can't directly compare those results anyway)... Best, Oliver |
Alexone Send message Joined: 21 Nov 11 Posts: 6 Credit: 1,000 RAC: 0 |
Please, look at this WU. http://albert.phys.uwm.edu/workunit.php?wuid=12832 What's wrong? Why Ati open'cl app made invalid result? ATI driver 11.8 |
pragmatic prancing periodic problem child, left Send message Joined: 26 Jan 05 Posts: 1639 Credit: 70,000 RAC: 0 |
I ran that one with Catalysts 11.7; also wonder why we didn't match. Jord. BOINC FAQ Service They say most of your brain shuts down in cryo-sleep. All but the primitive side, the animal side. No wonder I'm still awake. |
Jon Fox Send message Joined: 9 Feb 05 Posts: 7 Credit: 138,252 RAC: 0 |
iMac OS X 10.7.2/7.0.2 BOINC Client all ATI WUs failing with "Error while computing" status. Sample Stderr Output: http://albert.phys.uwm.edu/result.php?resultid=57408 http://albert.phys.uwm.edu/result.php?resultid=51158 -- Jon |
Oliver Behnke Volunteer moderator Project administrator Project developer Send message Joined: 4 Sep 07 Posts: 130 Credit: 8,545,955 RAC: 0 |
As the error messages indicate, your GPU doesn't have enough free memory. Sorry but we can't reliably check this in advance right now... Oliver |
Jon Fox Send message Joined: 9 Feb 05 Posts: 7 Credit: 138,252 RAC: 0 |
That's the way I read the error as well but didn't want to assume -- also wanted to make sure it was reported in case there was an error in the Work Unit assignment algorithms that should have been able to map work to memory configuration of a specific GPU. I'll stop processing work and monitor updates for any enhancements that will provide a better match between work units and GPU configuration. Thanks for the feedback and explanation. -- Jon |
oz Send message Joined: 28 Feb 05 Posts: 10 Credit: 1,285,478 RAC: 0 |
Same Codebase different platform / Compiler .. ? Linux: Intel(R) Xeon(R) CPU X5472 @ 3.00GHz [2] AMD ATI Radeon HD 5700 series (Juniper) (1024MB) driver: 1.4.1646 runtime 45000 - 61000 sec Windows:Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz [2] AMD ATI Radeon HD 5700 series (Juniper) (1024MB) driver: 1.4.1607 runtime 7100 - 8800 sec hmm... . I would expect ~ 8000sec for juniper radeon The successors (cypress, cayman) looking good. |
AMDave Send message Joined: 22 Apr 05 Posts: 3 Credit: 100,000 RAC: 0 |
After several failed attempts to get BOINC 7.0.2 (DEV) working under Ubuntu 11.04 and 11.10, I have returned an Albert OpenCL result under Win7-64. The GPGPU was a HD 5770, latest AMD driver version available 11.12, BOINC client is development version 7.0.2 The standard 500 coblestones was awarded. Run time was 30,744.65 seconds CPU time was 1,304.54 seconds So I make that just over 9 points per 10 minute period. Barely better than a CPU but using bucket-loads more power. The completion time is not quite 1 quarter of one of CPU-core's completion time on the same machine. But the CPU has 6 cores and so can still beat the snot out of the OpenCL performance on the HD 5770. Not really worth it then. Well, not with the current AMD Catalyst driver anyway. Once I get BOINC 7.0.2 working under Linux_AMD64 I should be able to test the different Catalyst drivers. But until then, this is the best I can do. |
Tex1954 Send message Joined: 1 Dec 11 Posts: 10 Credit: 25,500 RAC: 0 |
NEW 12.1 drivers fixed the HD6990 "high CPU usage" problem... work great! http://support.amd.com/us/kbarticles/Pages/Catalyst121Previewdriver.aspx 9-) |
AMDave Send message Joined: 22 Apr 05 Posts: 3 Credit: 100,000 RAC: 0 |
I saw an intriguing trend in the ATI HD 5770 WUs. the first finished at 33k+ seconds the second finished at 28K+seconds the third finished at 11k+ seconds I'm waiting on the fourth to see where this is going. |
Oliver Behnke Volunteer moderator Project administrator Project developer Send message Joined: 4 Sep 07 Posts: 130 Credit: 8,545,955 RAC: 0 |
NEW 12.1 drivers fixed the HD6990 "high CPU usage" problem... work great! This is great news, finally! Can others confirm that, too? I'll update our test rigs to check it myself... Oliver |
pragmatic prancing periodic problem child, left Send message Joined: 26 Jan 05 Posts: 1639 Credit: 70,000 RAC: 0 |
I haven't had the high CPU usage problem since the 11.11 drivers. I even put in a whole new videocard yesterday, so the drivers for it are installed clean. Using the 11.12 Catalysts here. Your app is using 1% CPU cycles every so many seconds. Jord. BOINC FAQ Service They say most of your brain shuts down in cryo-sleep. All but the primitive side, the animal side. No wonder I'm still awake. |
zombie67 [MM] Send message Joined: 10 Oct 06 Posts: 130 Credit: 30,924,459 RAC: 0 |
I am running 7.0.7, 12.1, XP64, C2Q, 5870, and the task is still consuming a about 40% of a CPU core. Edit: Some more info. When I reserve a core, GPU load is ~60%. When I do not reserve a core, GPU load drops to ~0%. Edit 2: Same behavior with 7.0.7, 12.1, Win7 64, i7 920, dual 5870s. Dublin, California Team: SETI.USA |
pragmatic prancing periodic problem child, left Send message Joined: 26 Jan 05 Posts: 1639 Credit: 70,000 RAC: 0 |
Ok, there's a very weird thing going on here. When I run work on my HD6850, and do not touch BOINC, don't suspend it or worse, exit it and later restart it, then most all tasks will run and finish in between 5,000 and 6,400 seconds. Do I suspend BOINC and resume it, or exit it and restart it, then work will become very slow and sluggish, running for multiple hours, up to 20K seconds. Before the GPU load will be sustained at around 70%, all the way through. After the GPU load is sporadic, coming and going but mostly zero. Like this: Date , GPU Core Clock [MHz] , GPU Memory Clock [MHz] , GPU Temperature [°C] , Fan Speed (%) [%] , Fan Speed (RPM) [RPM] , GPU Load [%] , GPU Temp. #1 [°C] , GPU Temp. #2 [°C] , GPU Temp. #3 [°C] , Memory Usage (Dedicated) [MB] , Memory Usage (Dynamic) [MB] , VDDC [V] 2012-01-08 02:17:58 , 625.0 , 1000.0 , 51.0 , 33 , 1131 , 0 , 52.0 , 53.0 , 53.5 , 495 , 71 , 0.000 , 2012-01-08 02:17:59 , 500.0 , 700.0 , 51.0 , 33 , 1133 , 23 , 51.5 , 52.0 , 52.5 , 495 , 71 , 0.000 , 2012-01-08 02:18:00 , 625.0 , 1000.0 , 51.0 , 33 , 1133 , 0 , 51.5 , 52.0 , 52.5 , 495 , 71 , 0.000 , 2012-01-08 02:18:01 , 625.0 , 1000.0 , 51.0 , 33 , 1130 , 0 , 51.5 , 52.5 , 53.0 , 495 , 71 , 0.000 , 2012-01-08 02:18:11 , 500.0 , 700.0 , 51.0 , 33 , 1139 , 10 , 51.0 , 51.0 , 52.0 , 495 , 71 , 0.000 , 2012-01-08 02:18:12 , 625.0 , 1000.0 , 51.0 , 33 , 1133 , 0 , 51.5 , 52.0 , 53.0 , 495 , 71 , 0.000 , .... all zero GPU load here ... 2012-01-08 02:19:04 , 625.0 , 1000.0 , 52.0 , 33 , 1130 , 26 , 52.0 , 53.0 , 53.5 , 495 , 70 , 0.000 , 2012-01-08 02:19:05 , 625.0 , 1000.0 , 52.0 , 33 , 1131 , 27 , 52.0 , 52.0 , 53.0 , 495 , 71 , 0.000 , 2012-01-08 02:19:06 , 625.0 , 1000.0 , 52.0 , 33 , 1132 , 25 , 52.0 , 52.5 , 53.0 , 495 , 70 , 0.000 , 2012-01-08 02:19:07 , 625.0 , 1000.0 , 51.0 , 33 , 1129 , 26 , 51.5 , 52.0 , 52.5 , 495 , 71 , 0.000 , 2012-01-08 02:19:08 , 625.0 , 1000.0 , 51.0 , 33 , 1129 , 25 , 51.5 , 51.5 , 52.5 , 495 , 71 , 0.000 , 2012-01-08 02:19:09 , 625.0 , 1000.0 , 52.0 , 33 , 1131 , 23 , 51.5 , 52.5 , 53.0 , 495 , 71 , 0.000 , 2012-01-08 02:19:10 , 500.0 , 700.0 , 51.0 , 33 , 1135 , 0 , 51.0 , 51.0 , 51.5 , 495 , 71 , 0.000 , 2012-01-08 02:19:11 , 625.0 , 1000.0 , 51.0 , 33 , 1130 , 0 , 51.5 , 52.0 , 52.5 , 495 , 71 , 0.000 , 2012-01-08 02:19:12 , 625.0 , 1000.0 , 52.0 , 33 , 1128 , 0 , 52.0 , 52.5 , 53.0 , 495 , 71 , 0.000 , Then when I do a full exit & restart of BOINC, it'll go into full overdrive again to catch up. Back to the sustained ~70%. Date , GPU Core Clock [MHz] , GPU Memory Clock [MHz] , GPU Temperature [°C] , Fan Speed (%) [%] , Fan Speed (RPM) [RPM] , GPU Load [%] , GPU Temp. #1 [°C] , GPU Temp. #2 [°C] , GPU Temp. #3 [°C] , Memory Usage (Dedicated) [MB] , Memory Usage (Dynamic) [MB] , VDDC [V] 2012-01-08 02:25:42 , 625.0 , 1000.0 , 51.0 , 33 , 1127 , 68 , 51.5 , 52.5 , 53.0 , 495 , 70 , 0.000 , 2012-01-08 02:25:42 , 625.0 , 1000.0 , 51.0 , 33 , 1126 , 69 , 51.5 , 53.0 , 53.5 , 495 , 70 , 0.000 , 2012-01-08 02:25:43 , 625.0 , 1000.0 , 52.0 , 33 , 1125 , 70 , 51.5 , 53.0 , 53.5 , 495 , 70 , 0.000 , 2012-01-08 02:25:44 , 625.0 , 1000.0 , 51.0 , 33 , 1125 , 72 , 52.0 , 52.5 , 53.0 , 495 , 71 , 0.000 , 2012-01-08 02:25:46 , 625.0 , 1000.0 , 52.0 , 33 , 1126 , 73 , 52.0 , 53.0 , 54.0 , 495 , 103 , 0.000 , 2012-01-08 02:25:46 , 625.0 , 1000.0 , 52.0 , 33 , 1127 , 74 , 52.0 , 53.0 , 53.5 , 495 , 71 , 0.000 , 2012-01-08 02:25:47 , 625.0 , 1000.0 , 51.0 , 33 , 1125 , 76 , 52.0 , 53.0 , 53.5 , 495 , 71 , 0.000 , 2012-01-08 02:25:49 , 625.0 , 1000.0 , 52.0 , 33 , 1123 , 76 , 52.0 , 53.5 , 54.0 , 495 , 71 , 0.000 , Jord. BOINC FAQ Service They say most of your brain shuts down in cryo-sleep. All but the primitive side, the animal side. No wonder I'm still awake. |
TRuEQ & TuVaLu Send message Joined: 11 Sep 06 Posts: 75 Credit: 615,315 RAC: 0 |
Ok, there's a very weird thing going on here. I tried suspending and then resumed.... I don't get the slow and sluggish thing... I run boinc 7.0.3 I also have marked "Leave task in memory while suspended". So tasks can resume without being read to memory when resumed. |
KAMasud Send message Joined: 6 Oct 06 Posts: 12 Credit: 1,500 RAC: 0 |
Ok! Got my first valid WU on a ATI GPU. Yaaaaa! Soon there might be a second. Ageless. Plz keep your beak shut, dont want any comment from you. |
KAMasud Send message Joined: 6 Oct 06 Posts: 12 Credit: 1,500 RAC: 0 |
Ok! Second one validated both against NVIDA. Here is what i have done. Went to AMD site and got their under development OpenCL and APP SDK installed those two and now i dont know where my Catlyst Control Center has gone to? Any way now getting valid units at Albert and Poem. |
TRuEQ & TuVaLu Send message Joined: 11 Sep 06 Posts: 75 Credit: 615,315 RAC: 0 |
I use 11.12 Catalyst suite and I have a working Control Center and the APP for OpenCl that is included with the driver. I don't need to install the separate SDK to run POEM and Albert or SETI OpenCl. |
Oliver Behnke Volunteer moderator Project administrator Project developer Send message Joined: 4 Sep 07 Posts: 130 Credit: 8,545,955 RAC: 0 |
Nice, the friends at SETI also noticed corrupt results seemingly caused by certain driver/SDK combinations. Maybe AMD finally fixed/improved their software. I'll try to build a new set of apps using the latest SDK for the latest drivers ASAP. Best, Oliver |
TRuEQ & TuVaLu Send message Joined: 11 Sep 06 Posts: 75 Credit: 615,315 RAC: 0 |
Any chance that the new apps will work on 4xxx ATI cards?? |