AMD card doing intel_gpu tasks

Andrew Dicker
Andrew Dicker
Joined: 1 May 13
Posts: 8
Credit: 12,035,093
RAC: 0
Topic 84945

Hi all, i have a machine with multiple GPUs (one from each brand), and the AMD card is doing intel_gpu tasks. Not that i mind (it is... just a little faster), but you might (and it messes with scheduling).

eg: http://albertathome.org/task/1078216
https://albertathome.org/task/1075483
https://albertathome.org/task/1075517

(Sorry for posting this in the drupal forum first, it looked like there had been more activity there recently is all)

Claggy
Claggy
Joined: 29 Dec 06
Posts: 122
Credit: 4,040,969
RAC: 0

AMD card doing intel_gpu tasks

Does your host report an Intel GPU? as the website here does not, has it ever reported an Intel GPU? What do the old logs say?

You might want to try upgrading your hosts to the latest Alpha release of Boinc (7.2.11), the Alpha Boinc 7.2.5 you're running could be the problem, (but unlikely, sounds more like a scheduler problem)
the point of running Alpha releases is you run them, report the Bugs, then move onto the next version, and report the bugs on that version.

Claggy

Alez
Alez
Joined: 8 Apr 13
Posts: 7
Credit: 4,335,055
RAC: 0

ok, I have the same but not

ok, I have the same but not quite the same. The problem is not with the Boinc client I think but the open_cl class part.
I have 3 discrete gpu's and the IGPU enabled. To control then I use a cc_config. If I set exclude gpu flag for NVIDIA it will exclude nVidia GPU 0 as it should. However if I exclude gpu ATI it excludes both the ATI GPU 0 and the intel_gpu.
It would appear that here both open cl class apps for the different gpu's are being treated as the same.

Alez
Alez
Joined: 8 Apr 13
Posts: 7
Credit: 4,335,055
RAC: 0

I just tried the exact same

I just tried the exact same on Einstein and there it works, exclude NVIDIA excludes the nVidia 0 and exclude ATI excludes the ATI 0 but allows the intel_gpu to work.

Claggy
Claggy
Joined: 29 Dec 06
Posts: 122
Credit: 4,040,969
RAC: 0

I've also seen this now, my

I've also seen this now, my i7-2600K/GTX460/HD7770 asked for Nvidia and ATI work and got sent the Intel app:

01/01/2014 12:55:40 | Albert@Home | [sched_op] Starting scheduler request
01/01/2014 12:55:40 | Albert@Home | Sending scheduler request: To fetch work.
01/01/2014 12:55:40 | Albert@Home | Requesting new tasks for NVIDIA and ATI
01/01/2014 12:55:40 | Albert@Home | [sched_op] CPU work request: 0.00 seconds; 0.00 devices
01/01/2014 12:55:40 | Albert@Home | [sched_op] NVIDIA work request: 18821.36 seconds; 0.00 devices
01/01/2014 12:55:40 | Albert@Home | [sched_op] ATI work request: 17770.31 seconds; 0.00 devices
01/01/2014 12:55:46 | Einstein@Home | Finished download of p2030.20130925.G62.79+00.41.S.b2s0g0.00000_1456.bin4
01/01/2014 12:55:46 | Einstein@Home | Started download of p2030.20130925.G62.79+00.41.S.b2s0g0.00000_1462.bin4
01/01/2014 12:55:46 | Albert@Home | Scheduler request completed: got 5 new tasks
01/01/2014 12:55:46 | Albert@Home | [sched_op] Server version 700
01/01/2014 12:55:46 | Albert@Home | Project requested delay of 60 seconds
01/01/2014 12:55:46 | Albert@Home | [sched_op] estimated total CPU task duration: 0 seconds
01/01/2014 12:55:46 | Albert@Home | [sched_op] estimated total NVIDIA task duration: 18283 seconds
01/01/2014 12:55:46 | Albert@Home | [sched_op] estimated total ATI task duration: 27928 seconds
01/01/2014 12:55:46 | Albert@Home | [sched_op] Deferring communication for 00:01:00
01/01/2014 12:55:46 | Albert@Home | [sched_op] Reason: requested by project
01/01/2014 12:55:48 | Albert@Home | Started download of einsteinbinary_BRP5_1.39_windows_x86_64__BRP5-opencl-intel_gpu.exe

2014-01-01 12:55:38.5362 [PID=19651] Request: [USER#xxxxx] [HOST#8143] [IP xxx.xxx.xxx.164] client 7.2.34
2014-01-01 12:55:38.5582 [PID=19651] [send] Not using matchmaker scheduling; Not using EDF sim
2014-01-01 12:55:38.5582 [PID=19651] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
2014-01-01 12:55:38.5582 [PID=19651] [send] CUDA: req 18821.36 sec, 0.00 instances; est delay 0.00
2014-01-01 12:55:38.5582 [PID=19651] [send] ATI: req 17770.31 sec, 0.00 instances; est delay 0.00
2014-01-01 12:55:38.5582 [PID=19651] [send] work_req_seconds: 0.00 secs
2014-01-01 12:55:38.5582 [PID=19651] [send] available disk 97.73 GB, work_buf_min 172800
2014-01-01 12:55:38.5582 [PID=19651] [send] active_frac 0.984597 on_frac 0.900038
2014-01-01 12:55:38.5582 [PID=19651] [send] p_vm_extensions_disabled: no
2014-01-01 12:55:38.5583 [PID=19651] [send] CPU 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
2014-01-01 12:55:38.5697 [PID=19651] [send] [AV#649] not reliable; cons valid 1 < 10
2014-01-01 12:55:38.5697 [PID=19651] [send] set_trust: cons valid 1 < 10, don't use single replication
2014-01-01 12:55:38.5697 [PID=19651] [send] [AV#710] not reliable; cons valid 1 < 10
2014-01-01 12:55:38.5697 [PID=19651] [send] set_trust: cons valid 1 < 10, don't use single replication
2014-01-01 12:55:38.5697 [PID=19651] [send] [AV#711] not reliable; cons valid 1 < 10
2014-01-01 12:55:38.5697 [PID=19651] [send] set_trust: cons valid 1 < 10, don't use single replication
2014-01-01 12:55:38.5697 [PID=19651] [send] [AV#712] not reliable; cons valid 4 < 10
2014-01-01 12:55:38.5697 [PID=19651] [send] set_trust: cons valid 4 < 10, don't use single replication
2014-01-01 12:55:38.5697 [PID=19651] [send] [AV#713] not reliable; cons valid 2 < 10
2014-01-01 12:55:38.5697 [PID=19651] [send] set_trust: cons valid 2 < 10, don't use single replication
2014-01-01 12:55:38.5697 [PID=19651] [send] [AV#716] not reliable; cons valid 0 < 10
2014-01-01 12:55:38.5698 [PID=19651] [send] set_trust: cons valid 0 < 10, don't use single replication
2014-01-01 12:55:38.5698 [PID=19651] [send] [AV#721] not reliable; cons valid 0 < 10
2014-01-01 12:55:38.5698 [PID=19651] [send] set_trust: cons valid 0 < 10, don't use single replication
2014-01-01 12:55:38.5698 [PID=19651] [send] [AV#728] not reliable; cons valid 0 < 10
2014-01-01 12:55:38.5698 [PID=19651] [send] set_trust: cons valid 0 < 10, don't use single replication
2014-01-01 12:55:38.5698 [PID=19651] [send] [AV#737] not reliable; cons valid 0 < 10
2014-01-01 12:55:38.5698 [PID=19651] [send] set_trust: cons valid 0 < 10, don't use single replication
2014-01-01 12:55:38.5698 [PID=19651] [send] [AV#766] not reliable; cons valid 1 < 10
2014-01-01 12:55:38.5698 [PID=19651] [send] set_trust: cons valid 1 < 10, don't use single replication
2014-01-01 12:55:38.5698 [PID=19651] [send] [AV#768] not reliable; cons valid 2 < 10
2014-01-01 12:55:38.5698 [PID=19651] [send] set_trust: cons valid 2 < 10, don't use single replication
2014-01-01 12:55:38.5698 [PID=19651] [mixed] sending non-locality work first
2014-01-01 12:55:38.5749 [PID=19651] [version] get_app_version(): getting app version for WU#520566 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_2011) appid:21
2014-01-01 12:55:38.5749 [PID=19651] [version] looking for version of einsteinbinary_BRP4
2014-01-01 12:55:38.5750 [PID=19651] [version] Checking plan class 'BRP4X64'
2014-01-01 12:55:38.5759 [PID=19651] [version] reading plan classes from file '/BOINC/projects/AlbertAtHome/plan_class_spec.xml'
2014-01-01 12:55:38.5759 [PID=19651] [version] parsed project prefs setting 'also_run_cpu' : true : 1.000000
2014-01-01 12:55:38.5759 [PID=19651] [version] project prefs setting 'also_run_cpu' (1.000000) prevents using plan class.
2014-01-01 12:55:38.5759 [PID=19651] [version] [AV#588] app_plan() returned false
2014-01-01 12:55:38.5760 [PID=19651] [version] Checking plan class 'BRP4SSE'
2014-01-01 12:55:38.5760 [PID=19651] [version] parsed project prefs setting 'also_run_cpu' : true : 1.000000
2014-01-01 12:55:38.5761 [PID=19651] [version] project prefs setting 'also_run_cpu' (1.000000) prevents using plan class.
2014-01-01 12:55:38.5761 [PID=19651] [version] [AV#598] app_plan() returned false
2014-01-01 12:55:38.5761 [PID=19651] [version] returning NULL; platforms:
2014-01-01 12:55:38.5761 [PID=19651] [version] windows_x86_64
2014-01-01 12:55:38.5761 [PID=19651] [version] windows_intelx86
2014-01-01 12:55:38.5761 [PID=19651] [version] get_app_version(): getting app version for WU#520566 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_2011) appid:21
2014-01-01 12:55:38.5761 [PID=19651] [version] get_app_version(): getting app version for WU#529183 (PA0043_015B1_316) appid:27
2014-01-01 12:55:38.5761 [PID=19651] [version] looking for version of einsteinbinary_BRP5
2014-01-01 12:55:38.5762 [PID=19651] [version] Checking plan class 'BRP5-opencl-ati'
2014-01-01 12:55:38.5762 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.5762 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.5762 [PID=19651] [version] Checking plan class 'BRP5-opencl-intel_gpu'
2014-01-01 12:55:38.5762 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.5762 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.5763 [PID=19651] [version] [AV#729] (BRP5-opencl-intel_gpu) adjusting projected flops based on PFC avg: 38.50G
2014-01-01 12:55:38.5763 [PID=19651] [version] Best version of app einsteinbinary_BRP5 is [AV#729] (38.50 GFLOPS)
2014-01-01 12:55:38.5763 [PID=19651] [send] est delay 0, skipping deadline check
2014-01-01 12:55:38.6481 [PID=19651] [send] Sending app_version einsteinbinary_BRP5 7 139 BRP5-opencl-intel_gpu; projected 38.50 GFLOPS
2014-01-01 12:55:38.6483 [PID=19651] [send] est. duration for WU 529183: unscaled 11687.60 scaled 13188.82
2014-01-01 12:55:38.6483 [PID=19651] [HOST#8143] Sending [RESULT#1278610 PA0043_015B1_316_0] (est. dur. 13188.82s (3h39m48s81)) (max time 233751.91s (64h55m51s91))
2014-01-01 12:55:38.6490 [PID=19651] [version] get_app_version(): getting app version for WU#528708 (p2030.20130202.G202.59-01.95.N.b5s0g0.00000_2752) appid:29
2014-01-01 12:55:38.6490 [PID=19651] [version] looking for version of einsteinbinary_BRP4G
2014-01-01 12:55:38.6491 [PID=19651] [version] Checking plan class 'BRP4G-opencl-ati'
2014-01-01 12:55:38.6491 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6491 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6491 [PID=19651] [version] [AV#721] (BRP4G-opencl-ati) adjusting projected flops based on PFC avg: 69.49G
2014-01-01 12:55:38.6491 [PID=19651] [version] Best version of app einsteinbinary_BRP4G is [AV#721] (69.49 GFLOPS)
2014-01-01 12:55:38.6491 [PID=19651] [send] est. duration for WU 528708: unscaled 4029.32 scaled 4546.87
2014-01-01 12:55:38.6491 [PID=19651] [send] [WU#528708] meets deadline: 13188.82 + 4546.87 < 1209600
2014-01-01 12:55:38.6553 [PID=19651] [send] Sending app_version einsteinbinary_BRP4G 7 134 BRP4G-opencl-ati; projected 69.49 GFLOPS
2014-01-01 12:55:38.6556 [PID=19651] [send] est. duration for WU 528708: unscaled 4029.32 scaled 4546.87
2014-01-01 12:55:38.6556 [PID=19651] [HOST#8143] Sending [RESULT#1276650 p2030.20130202.G202.59-01.95.N.b5s0g0.00000_2752_0] (est. dur. 4546.87s (1h15m46s87)) (max time 80586.49s (22h23m06s48))
2014-01-01 12:55:38.6560 [PID=19651] [version] get_app_version(): getting app version for WU#520567 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_2012) appid:21
2014-01-01 12:55:38.6561 [PID=19651] [version] get_app_version(): getting app version for WU#520553 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_1998) appid:21
2014-01-01 12:55:38.6561 [PID=19651] [version] get_app_version(): getting app version for WU#529178 (PA0043_015B1_306) appid:27
2014-01-01 12:55:38.6561 [PID=19651] [version] returning cached version: [AV#729]
2014-01-01 12:55:38.6561 [PID=19651] [send] est. duration for WU 529178: unscaled 11687.60 scaled 13188.82
2014-01-01 12:55:38.6562 [PID=19651] [send] [WU#529178] meets deadline: 17735.69 + 13188.82 < 1209600
2014-01-01 12:55:38.6604 [PID=19651] [send] Sending app_version einsteinbinary_BRP5 7 139 BRP5-opencl-intel_gpu; projected 38.50 GFLOPS
2014-01-01 12:55:38.6607 [PID=19651] [send] est. duration for WU 529178: unscaled 11687.60 scaled 13188.82
2014-01-01 12:55:38.6607 [PID=19651] [HOST#8143] Sending [RESULT#1278601 PA0043_015B1_306_1] (est. dur. 13188.82s (3h39m48s81)) (max time 233751.91s (64h55m51s91))
2014-01-01 12:55:38.6614 [PID=19651] [version] get_app_version(): getting app version for WU#528547 (p2030.20130202.G202.59-01.95.N.b5s0g0.00000_176) appid:29
2014-01-01 12:55:38.6614 [PID=19651] [version] have ATI version but no more ATI work needed
2014-01-01 12:55:38.6614 [PID=19651] [version] looking for version of einsteinbinary_BRP4G
2014-01-01 12:55:38.6614 [PID=19651] [version] Checking plan class 'BRP4G-opencl-ati'
2014-01-01 12:55:38.6614 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6614 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6614 [PID=19651] [version] [AV#721] Don't need ATI jobs, skipping
2014-01-01 12:55:38.6614 [PID=19651] [version] Checking plan class 'BRP4G-cuda32'
2014-01-01 12:55:38.6614 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6615 [PID=19651] [version] driver version required max: 29053, supplied: 33193
2014-01-01 12:55:38.6615 [PID=19651] [version] [AV#723] app_plan() returned false
2014-01-01 12:55:38.6615 [PID=19651] [version] Checking plan class 'BRP4G-cuda32-nv301'
2014-01-01 12:55:38.6615 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6615 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 10.00, projected_flops: 3.460719e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6615 [PID=19651] [version] Checking plan class 'BRP4G-opencl-ati'
2014-01-01 12:55:38.6615 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6615 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6615 [PID=19651] [version] [AV#720] Don't need ATI jobs, skipping
2014-01-01 12:55:38.6615 [PID=19651] [version] [AV#716] (BRP4G-cuda32-nv301) adjusting projected flops based on PFC avg: 62.01G
2014-01-01 12:55:38.6615 [PID=19651] [version] Best version of app einsteinbinary_BRP4G is [AV#716] (62.01 GFLOPS)
2014-01-01 12:55:38.6615 [PID=19651] [send] est delay 0, skipping deadline check
2014-01-01 12:55:38.6688 [PID=19651] [send] Sending app_version einsteinbinary_BRP4G 2 133 BRP4G-cuda32-nv301; projected 62.01 GFLOPS
2014-01-01 12:55:38.6690 [PID=19651] [send] est. duration for WU 528547: unscaled 4515.40 scaled 5095.38
2014-01-01 12:55:38.6691 [PID=19651] [HOST#8143] Sending [RESULT#1280622 p2030.20130202.G202.59-01.95.N.b5s0g0.00000_176_2] (est. dur. 5095.38s (1h24m55s38)) (max time 90307.95s (25h05m07s94))
2014-01-01 12:55:38.6693 [PID=19651] [version] get_app_version(): getting app version for WU#520563 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_2008) appid:21
2014-01-01 12:55:38.6694 [PID=19651] [version] get_app_version(): getting app version for WU#520554 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_1999) appid:21
2014-01-01 12:55:38.6694 [PID=19651] [version] get_app_version(): getting app version for WU#529178 (PA0043_015B1_306) appid:27
2014-01-01 12:55:38.6694 [PID=19651] [version] have ATI version but no more ATI work needed
2014-01-01 12:55:38.6695 [PID=19651] [version] looking for version of einsteinbinary_BRP5
2014-01-01 12:55:38.6695 [PID=19651] [version] Checking plan class 'BRP5-opencl-ati'
2014-01-01 12:55:38.6695 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6695 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6695 [PID=19651] [version] [AV#728] Don't need ATI jobs, skipping
2014-01-01 12:55:38.6695 [PID=19651] [version] Checking plan class 'BRP5-opencl-intel_gpu'
2014-01-01 12:55:38.6695 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6695 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6695 [PID=19651] [version] [AV#729] Don't need ATI jobs, skipping
2014-01-01 12:55:38.6696 [PID=19651] [version] Checking plan class 'BRP5-cuda32'
2014-01-01 12:55:38.6696 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6696 [PID=19651] [version] driver version required max: 29053, supplied: 33193
2014-01-01 12:55:38.6696 [PID=19651] [version] [AV#725] app_plan() returned false
2014-01-01 12:55:38.6696 [PID=19651] [version] Checking plan class 'BRP5-cuda32-nv301'
2014-01-01 12:55:38.6696 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6696 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 10.00, projected_flops: 3.460719e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6696 [PID=19651] [version] Checking plan class 'BRP5-opencl-ati'
2014-01-01 12:55:38.6696 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.6696 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.6696 [PID=19651] [version] [AV#738] Don't need ATI jobs, skipping
2014-01-01 12:55:38.6696 [PID=19651] [version] [AV#737] (BRP5-cuda32-nv301) adjusting projected flops based on PFC avg: 33.52G
2014-01-01 12:55:38.6696 [PID=19651] [version] Best version of app einsteinbinary_BRP5 is [AV#737] (33.52 GFLOPS)
2014-01-01 12:55:38.6697 [PID=19651] [send] [HOST#8143] [WU#529178 PA0043_015B1_306] WU is infeasible: Already in reply
2014-01-01 12:55:38.6697 [PID=19651] [version] get_app_version(): getting app version for WU#528587 (p2030.20130202.G202.59-01.95.N.b5s0g0.00000_816) appid:29
2014-01-01 12:55:38.6697 [PID=19651] [version] returning cached version: [AV#716]
2014-01-01 12:55:38.6697 [PID=19651] [send] est. duration for WU 528587: unscaled 4515.40 scaled 5095.38
2014-01-01 12:55:38.6697 [PID=19651] [send] [WU#528587] meets deadline: 5095.38 + 5095.38 < 1209600
2014-01-01 12:55:38.6710 [PID=19651] [RESULT#1280337] expected to be unsent; instead, state is 4
2014-01-01 12:55:38.6711 [PID=19651] [version] get_app_version(): getting app version for WU#520573 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_2018) appid:21
2014-01-01 12:55:38.6711 [PID=19651] [version] get_app_version(): getting app version for WU#520568 (p2030.20130202.G202.81-01.04.N.b3s0g0.00000_2013) appid:21
2014-01-01 12:55:38.6712 [PID=19651] [version] get_app_version(): getting app version for WU#529179 (PA0043_015B1_308) appid:27
2014-01-01 12:55:38.6712 [PID=19651] [version] returning cached version: [AV#737]
2014-01-01 12:55:38.6712 [PID=19651] [send] est. duration for WU 529179: unscaled 13425.01 scaled 15149.40
2014-01-01 12:55:38.6712 [PID=19651] [send] [WU#529179] meets deadline: 5095.38 + 15149.40 < 1209600
2014-01-01 12:55:38.7413 [PID=19651] [send] Sending app_version einsteinbinary_BRP5 2 139 BRP5-cuda32-nv301; projected 33.52 GFLOPS
2014-01-01 12:55:38.7418 [PID=19651] [send] est. duration for WU 529179: unscaled 13425.01 scaled 15149.40
2014-01-01 12:55:38.7419 [PID=19651] [HOST#8143] Sending [RESULT#1278603 PA0043_015B1_308_1] (est. dur. 15149.40s (4h12m29s39)) (max time 268500.27s (74h35m00s26))
2014-01-01 12:55:38.7431 [PID=19651] [send] don't need more work
2014-01-01 12:55:38.7432 [PID=19651] [mixed] sending locality work second
2014-01-01 12:55:38.7433 [PID=19651] [locality] [HOST#8143] removing file rand_PAS.bank.v3 from file_infos list
2014-01-01 12:55:38.7433 [PID=19651] [locality] [HOST#8143] removing file JPLEPH.405 from file_infos list
2014-01-01 12:55:38.7434 [PID=19651] [locality] [HOST#8143] removing file stochastic_full.bank from file_infos list
2014-01-01 12:55:38.7434 [PID=19651] [send] don't need more work
2014-01-01 12:55:38.7456 [PID=19651] [credit] created host_app_version record (8143, 729)
2014-01-01 12:55:38.7509 [PID=19651] Sending reply to [HOST#8143]: 5 results, delay req 60.00
2014-01-01 12:55:38.7514 [PID=19651] Scheduler ran 0.222 seconds

Claggy

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 143
Credit: 5,409,572
RAC: 0

Task 1278610 / WU 529183 was

Message 79828 in response to message 79827

Task 1278610 / WU 529183 was sent as an intel_gpu app:

Quote:
2014-01-01 12:55:38.5761 [PID=19651] [version] get_app_version(): getting app version for WU#529183 (PA0043_015B1_316) appid:27
2014-01-01 12:55:38.5761 [PID=19651] [version] looking for version of einsteinbinary_BRP5
2014-01-01 12:55:38.5762 [PID=19651] [version] Checking plan class 'BRP5-opencl-ati'
2014-01-01 12:55:38.5762 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.5762 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.5762 [PID=19651] [version] Checking plan class 'BRP5-opencl-intel_gpu'
2014-01-01 12:55:38.5762 [PID=19651] [version] parsed project prefs setting 'gpu_util_brp' : true : 1.000000
2014-01-01 12:55:38.5762 [PID=19651] [version] host_flops: 4.432157e+09, speedup: 15.00, projected_flops: 5.603312e+10, peak_flops: 4.432157e+09, peak_flops_factor: 0.21
2014-01-01 12:55:38.5763 [PID=19651] [version] [AV#729] (BRP5-opencl-intel_gpu) adjusting projected flops based on PFC avg: 38.50G
2014-01-01 12:55:38.5763 [PID=19651] [version] Best version of app einsteinbinary_BRP5 is [AV#729] (38.50 GFLOPS)
2014-01-01 12:55:38.5763 [PID=19651] [send] est delay 0, skipping deadline check
2014-01-01 12:55:38.6481 [PID=19651] [send] Sending app_version einsteinbinary_BRP5 7 139 BRP5-opencl-intel_gpu; projected 38.50 GFLOPS
2014-01-01 12:55:38.6483 [PID=19651] [send] est. duration for WU 529183: unscaled 11687.60 scaled 13188.82
2014-01-01 12:55:38.6483 [PID=19651] [HOST#8143] Sending [RESULT#1278610 PA0043_015B1_316_0] (est. dur. 13188.82s (3h39m48s81)) (max time 233751.91s (64h55m51s91))


That may be the fastest app in the database, but it can't be the best app if the hardware doesn't support it!

I didn't want to spam the boards with my stats - just milestone theads - but apparently signatures are no longer optional. Follow the link if you're interested.

http://www.boincsynergy.com/images/stats/comb-3475.jpg

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 143
Credit: 5,409,572
RAC: 0

Might like to refer to

Message 79829 in response to message 79828

Might like to refer to message 128489 by Heinz-Bernd at Einstein:

Quote:
Currently, the Intel OpenCL app is just a enamed [renamed ?] AMD/ATI OpenCL app. That is the beauty of OpenCL (compile once, run anywhere, so to speak..), it's a bit sad that this doesn't work with the newest Intel drivers. We'll check if we can find a workaround or fix this somehow.


Conversely, the Intel app should work just fine on AMD/ATI.....

I didn't want to spam the boards with my stats - just milestone theads - but apparently signatures are no longer optional. Follow the link if you're interested.

http://www.boincsynergy.com/images/stats/comb-3475.jpg

Claggy
Claggy
Joined: 29 Dec 06
Posts: 122
Credit: 4,040,969
RAC: 0

RE: Might like to refer to

Message 79830 in response to message 79829

Quote:

Might like to refer to message 128489 by Heinz-Bernd at Einstein:

Quote:
Currently, the Intel OpenCL app is just a enamed [renamed ?] AMD/ATI OpenCL app. That is the beauty of OpenCL (compile once, run anywhere, so to speak..), it's a bit sad that this doesn't work with the newest Intel drivers. We'll check if we can find a workaround or fix this somehow.

Conversely, the Intel app should work just fine on AMD/ATI.....


I thought it was the same, I was just confirming other reports.

Claggy

Andrew Dicker
Andrew Dicker
Joined: 1 May 13
Posts: 8
Credit: 12,035,093
RAC: 0

Two proper ATI tasks, run on

Message 79831 in response to message 79830

Two proper ATI tasks, run on my ati:

https://albertathome.org/workunit/528234
https://albertathome.org/workunit/528233

Two supposed Intel tasks (by task name) also on the ATI (this was reported in the manager also, so at least resource scheduling wasn't interfered with):
https://albertathome.org/workunit/528182
https://albertathome.org/workunit/528179

My ati OpenCL runtime is environment 0 according to an OpenCl thing i ran, so is the intel app just running in the 0'th environment, and 0th device (ati card) by default? (I had a similar issue on PrimeGrid where the tasks were all running on the first device)

The iGPU runs Einstein short BRP tasks fine - two at a time even (using older drivers).

zombie67 [MM]
zombie67 [MM]
Joined: 10 Oct 06
Posts: 73
Credit: 30,924,459
RAC: 0

Yeah, I have the same

Yeah, I have the same problem. Getting tired of constantly aborting intel GPU tasks on my machines that have no Intel GPUs.

On the other hand, I cannot get any tasks for my other machine that has an intel GPU.

I notice that there is no checkbox in the project preferences for intel GPU.

So how is it that there are tasks for intel GPU?

Dublin, California
Team: SETI.USA

Claggy
Claggy
Joined: 29 Dec 06
Posts: 122
Credit: 4,040,969
RAC: 0

RE: Yeah, I have the same

Message 79833 in response to message 79832

Quote:

Yeah, I have the same problem. Getting tired of constantly aborting intel GPU tasks on my machines that have no Intel GPUs.

On the other hand, I cannot get any tasks for my other machine that has an intel GPU.

I notice that there is no checkbox in the project preferences for intel GPU.

So how is it that there are tasks for intel GPU?


I don't think the project has updated it's Boinc server software to a new enough level to even display hosts that have Intel GPUs, let alone send work out to intel GPUs.

My i5-3210M/GT 650M/HD Graphics 4000 host at Albert:

NVIDIA GeForce GT 650M (2048MB) driver: 331.93

My i5-3210M/GT 650M/HD Graphics 4000 host at Einstein:

NVIDIA GeForce GT 650M (2048MB) driver: 33193, INTEL Intel(R) HD Graphics 4000 (1624MB)

Claggy

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.