News:

Members can see all forum boards and posts. Non members can only see a few boards.  If you have forgotten your password use this link to change it.
https://forum.boinc-australia.net/index.php?action=reminder

Main Menu

MW@Home only uses1 GPU

Started by TouchuvGrey, November 11, 2020, 05:23:41 PM

Previous topic - Next topic

TouchuvGrey

i have two video cards, a Radeon RX580 and a NVidia GTX680.
Device Manager shows both of the functioning properly, yet MW @Home
is only using the Radeon card even though i have both ticked in preferences.
Can someone tell me how to get MW@Home to use both ?

Daniel

#1
Hi TouchuvGrey,

Here are some things I can think of ...

1) Can you see entries for both cards in the BOINC Manager event log ?

For example, in one of mine with 2 NVidia cards I see ...

12/11/2020 5:40:00 PM |  | CUDA: NVIDIA GPU 0: GeForce GTX 1060 6GB (driver version 457.09, CUDA version 11.1, compute capability 6.1, 4096MB, 3556MB available, 4374 GFLOPS peak)
12/11/2020 5:40:00 PM |  | CUDA: NVIDIA GPU 1: GeForce GTX 960 (driver version 457.09, CUDA version 11.1, compute capability 5.2, 2048MB, 1669MB available, 2748 GFLOPS peak)
12/11/2020 5:40:00 PM |  | OpenCL: NVIDIA GPU 0: GeForce GTX 1060 6GB (driver version 457.09, device version OpenCL 1.2 CUDA, 6144MB, 3556MB available, 4374 GFLOPS peak)
12/11/2020 5:40:00 PM |  | OpenCL: NVIDIA GPU 1: GeForce GTX 960 (driver version 457.09, device version OpenCL 1.2 CUDA, 2048MB, 1669MB available, 2748 GFLOPS peak)

I expect you shouldn't see the CUDA lines for the AMD GPU though.
If you don't see entries for the NVidia card, you could try installing the latest driver from the NVidia website.
I think if you are just using the driver that Windows provides, the NVidia card might not be usable for crunching BOINC - I always use the driver from the NVidia website.


2) Check that you have the following line in the <options> section of your cc_config.xml file (located in C:\ProgramData\BOINC\) ...
<use_all_gpus>1</use_all_gpus>


3) If you are using an app_config.xml for Milkyway (located in C:\ProgramData\BOINC\projects\milkyway.cs.rpi.edu_milkyway\), check that you are not limiting the number of concurrent tasks. For example <max_concurrent>1</max_concurrent> would limit the number of concurrent tasks to 1. You may not have an app_config.xml though, in which case you shouldn't have a problem.


4) After making any of the driver/config changes mentioned above restart BOINC. Or in the case of a driver change you could also try restarting Windows, although I don't usually have to do that.

TouchuvGrey

not sure just what happened. i have neither app config xml or cc config xml.

However i just stopped and restarted Boinc and now i have

12/11/2020 6:03:05 PM |  | cc_config.xml not found - using defaults
12/11/2020 6:03:05 PM |  | Starting BOINC client version 7.16.7 for windows_x86_64
12/11/2020 6:03:05 PM |  | Libraries: libcurl/7.47.1 OpenSSL/1.0.2s zlib/1.2.8
12/11/2020 6:03:05 PM |  | Data directory: C:\ProgramData\BOINC
12/11/2020 6:03:05 PM |  | Running under account Owner
12/11/2020 6:03:06 PM |  | CUDA: NVIDIA GPU 0: GeForce GTX 680 (driver version 457.30, CUDA version 11.1, compute capability 3.0, 2048MB, 1671MB available, 3533 GFLOPS peak)
12/11/2020 6:03:06 PM |  | OpenCL: NVIDIA GPU 0: GeForce GTX 680 (driver version 457.30, device version OpenCL 1.2 CUDA, 2048MB, 1671MB available, 3533 GFLOPS peak)
12/11/2020 6:03:06 PM |  | OpenCL: AMD/ATI GPU 0: Radeon RX 580 Series (driver version 3110.7, device version OpenCL 2.0 AMD-APP (3110.7), 8192MB, 8192MB available, 6267 GFLOPS peak)

TouchuvGrey

Had some issues here and ended up reinstalling windows.

    Now MW@Hom is only using NVidia card.

Event log

16/11/2020 3:10:05 PM |  | cc_config.xml not found - using defaults
16/11/2020 3:10:05 PM |  | Starting BOINC client version 7.16.11 for windows_x86_64
16/11/2020 3:10:05 PM |  | log flags: file_xfer, sched_ops, task
16/11/2020 3:10:05 PM |  | Libraries: libcurl/7.47.1 OpenSSL/1.0.2s zlib/1.2.8
16/11/2020 3:10:05 PM |  | Data directory: C:\ProgramData\BOINC
16/11/2020 3:10:05 PM |  | Running under account micha
16/11/2020 3:10:05 PM |  | CUDA: NVIDIA GPU 0: GeForce GTX 680 (driver version 432.00, CUDA version 10.1, compute capability 3.0, 2048MB, 1671MB available, 3533 GFLOPS peak)
16/11/2020 3:10:05 PM |  | OpenCL: NVIDIA GPU 0: GeForce GTX 680 (driver version 432.00, device version OpenCL 1.2 CUDA, 2048MB, 1671MB available, 3533 GFLOPS peak)
16/11/2020 3:10:05 PM |  | OpenCL: AMD/ATI GPU 0: Radeon RX 580 Series (driver version 3110.7, device version OpenCL 2.0 AMD-APP (3110.7), 8192MB, 8192MB available, 6267 GFLOPS peak)
16/11/2020 3:10:06 PM |  | Windows processor group 0: 6 processors
16/11/2020 3:10:06 PM |  | Host name: DESKTOP-67L00OK
16/11/2020 3:10:06 PM |  | Processor: 6 GenuineIntel Intel(R) Core(TM) i5-8600K CPU @ 3.60GHz [Family 6 Model 158 Stepping 10]
16/11/2020 3:10:06 PM |  | 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 fma cx16 sse4_1 sse4_2 movebe popcnt aes f16c rdrandsyscall nx lm avx avx2 vmx smx tm2 pbe fsgsbase bmi1 hle smep bmi2
16/11/2020 3:10:06 PM |  | OS: Microsoft Windows 10: Core x64 Edition, (10.00.19042.00)
16/11/2020 3:10:06 PM |  | Memory: 15.95 GB physical, 18.82 GB virtual
16/11/2020 3:10:06 PM |  | Disk: 464.25 GB total, 429.04 GB free
16/11/2020 3:10:06 PM |  | Local time is UTC +10 hours

coproc Info

<?xml version="1.0"?>

-<coprocs>

<have_cuda>1</have_cuda>

<cuda_version>10010</cuda_version>


-<coproc_cuda>

<count>1</count>

<name>GeForce GTX 680</name>

<available_ram>1751836263.000000</available_ram>

<have_cuda>1</have_cuda>

<have_opencl>0</have_opencl>

<peak_flops>3532800000000.000000</peak_flops>

<cudaVersion>10010</cudaVersion>

<drvVersion>43200</drvVersion>

<totalGlobalMem>2147483648.000000</totalGlobalMem>

<sharedMemPerBlock>49152.000000</sharedMemPerBlock>

<regsPerBlock>65536</regsPerBlock>

<warpSize>32</warpSize>

<memPitch>2147483647.000000</memPitch>

<maxThreadsPerBlock>1024</maxThreadsPerBlock>

<maxThreadsDim>1024 1024 64</maxThreadsDim>

<maxGridSize>2147483647 65535 65535</maxGridSize>

<clockRate>1150000</clockRate>

<totalConstMem>65536.000000</totalConstMem>

<major>3</major>

<minor>0</minor>

<textureAlignment>512.000000</textureAlignment>

<deviceOverlap>1</deviceOverlap>

<multiProcessorCount>8</multiProcessorCount>


-<pci_info>

<bus_id>1</bus_id>

<device_id>0</device_id>

<domain_id>0</domain_id>

</pci_info>

</coproc_cuda>


-<ati_opencl>

<name>Radeon RX 580 Series</name>

<vendor>Advanced Micro Devices, Inc.</vendor>

<vendor_id>4098</vendor_id>

<available>1</available>

<half_fp_config>0</half_fp_config>

<single_fp_config>190</single_fp_config>

<double_fp_config>63</double_fp_config>

<endian_little>1</endian_little>

<execution_capabilities>1</execution_capabilities>

<extensions>cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_gl_sharing cl_khr_gl_depth_images cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_khr_image2d_from_buffer cl_khr_spir cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_liquid_flash cl_amd_planar_yuv</extensions>

<global_mem_size>8589934592</global_mem_size>

<local_mem_size>32768</local_mem_size>

<max_clock_frequency>1360</max_clock_frequency>

<max_compute_units>36</max_compute_units>

<nv_compute_capability_major>0</nv_compute_capability_major>

<nv_compute_capability_minor>0</nv_compute_capability_minor>

<amd_simd_per_compute_unit>4</amd_simd_per_compute_unit>

<amd_simd_width>16</amd_simd_width>

<amd_simd_instruction_width>1</amd_simd_instruction_width>

<opencl_platform_version>OpenCL 2.1 AMD-APP (3110.7)</opencl_platform_version>

<opencl_device_version>OpenCL 2.0 AMD-APP (3110.7)</opencl_device_version>

<opencl_driver_version>3110.7</opencl_driver_version>

<device_num>0</device_num>

<peak_flops>6266880000000.000000</peak_flops>

<opencl_available_ram>8589934592.000000</opencl_available_ram>

<opencl_device_index>0</opencl_device_index>

<warn_bad_cuda>0</warn_bad_cuda>

</ati_opencl>


-<nvidia_opencl>

<name>GeForce GTX 680</name>

<vendor>NVIDIA Corporation</vendor>

<vendor_id>4318</vendor_id>

<available>1</available>

<half_fp_config>0</half_fp_config>

<single_fp_config>191</single_fp_config>

<double_fp_config>63</double_fp_config>

<endian_little>1</endian_little>

<execution_capabilities>1</execution_capabilities>

<extensions>cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_d3d10_sharing cl_khr_d3d10_sharing cl_nv_d3d11_sharing cl_nv_copy_opts cl_nv_create_buffer</extensions>

<global_mem_size>2147483648</global_mem_size>

<local_mem_size>49152</local_mem_size>

<max_clock_frequency>1150</max_clock_frequency>

<max_compute_units>8</max_compute_units>

<nv_compute_capability_major>3</nv_compute_capability_major>

<nv_compute_capability_minor>0</nv_compute_capability_minor>

<amd_simd_per_compute_unit>0</amd_simd_per_compute_unit>

<amd_simd_width>0</amd_simd_width>

<amd_simd_instruction_width>0</amd_simd_instruction_width>

<opencl_platform_version>OpenCL 1.2 CUDA 10.1.120</opencl_platform_version>

<opencl_device_version>OpenCL 1.2 CUDA</opencl_device_version>

<opencl_driver_version>432.00</opencl_driver_version>

<device_num>0</device_num>

<peak_flops>3532800000000.000000</peak_flops>

<opencl_available_ram>1751836263.000000</opencl_available_ram>

<opencl_device_index>0</opencl_device_index>

<warn_bad_cuda>0</warn_bad_cuda>

</nvidia_opencl>

<warning>NVIDIA library reports 1 GPU</warning>

<warning>No ATI library found.</warning>

</coprocs>





TouchuvGrey

And now it is back to using both.

chooka03

Problem solved!
Phew...that was easy  :jester:

TouchuvGrey

And yet another puzzle ( to me anyway )

i decided i would try running N-body tasks again and my computer seems
to be prioritizing those over Separation tasks so what i have now is

    Separations tasks "waiting to run"
AMD card task  Separation "waiting to run"
Nvidia card running
N-Body running ( 6 CPU's )

What have i done to myself this time ?


17/11/2020 3:26:00 PM | Milkyway@Home | Sending scheduler request: To fetch work.
17/11/2020 3:26:00 PM | Milkyway@Home | Reporting 1 completed tasks
17/11/2020 3:26:00 PM | Milkyway@Home | Requesting new tasks for CPU and NVIDIA GPU and AMD/ATI GPU
17/11/2020 3:26:02 PM | Milkyway@Home | Scheduler request completed: got 0 new tasks
17/11/2020 3:26:02 PM | Milkyway@Home | No tasks sent
17/11/2020 3:26:02 PM | Milkyway@Home | This computer has reached a limit on tasks in progress
17/11/2020 3:26:02 PM | Milkyway@Home | Project requested delay of 91 seconds
17/11/2020 3:26:19 PM | Milkyway@Home | Computation for task de_modfit_86_bundle4_4s_south4s_bgset_3_1603804501_15758751_1 finished
17/11/2020 3:26:19 PM | Milkyway@Home | Starting task de_modfit_83_bundle4_4s_south4s_bgset_3_1603804501_15797758_0
17/11/2020 3:26:23 PM | Milkyway@Home | Computation for task de_modfit_86_bundle4_4s_south4s_bgset_3_1603804501_15022498_0 finished
17/11/2020 3:26:24 PM | Milkyway@Home | Starting task de_modfit_84_bundle4_4s_south4s_bgset_3_1603804501_14829183_1
17/11/2020 3:27:33 PM | Milkyway@Home | Sending scheduler request: To fetch work.
17/11/2020 3:27:33 PM | Milkyway@Home | Reporting 2 completed tasks
17/11/2020 3:27:33 PM | Milkyway@Home | Requesting new tasks for CPU and NVIDIA GPU and AMD/ATI GPU
17/11/2020 3:27:36 PM | Milkyway@Home | Scheduler request completed: got 0 new tasks
17/11/2020 3:27:36 PM | Milkyway@Home | No tasks sent
17/11/2020 3:27:36 PM | Milkyway@Home | This computer has reached a limit on tasks in progress
17/11/2020 3:27:36 PM | Milkyway@Home | Project requested delay of 91 seconds
17/11/2020 3:27:54 PM | Milkyway@Home | Computation for task de_modfit_83_bundle4_4s_south4s_bgset_3_1603804501_15797758_0 finished
17/11/2020 3:27:54 PM | Milkyway@Home | Starting task de_modfit_86_bundle4_4s_south4s_bgset_3_1603804501_15755378_1
17/11/2020 3:29:08 PM | Milkyway@Home | Sending scheduler request: To fetch work.
17/11/2020 3:29:08 PM | Milkyway@Home | Reporting 1 completed tasks
17/11/2020 3:29:08 PM | Milkyway@Home | Requesting new tasks for CPU and NVIDIA GPU and AMD/ATI GPU
17/11/2020 3:29:10 PM | Milkyway@Home | Scheduler request completed: got 0 new tasks
17/11/2020 3:29:10 PM | Milkyway@Home | No tasks sent
17/11/2020 3:29:10 PM | Milkyway@Home | This computer has reached a limit on tasks in progress
17/11/2020 3:29:10 PM | Milkyway@Home | Project requested delay of 91 seconds
17/11/2020 3:29:29 PM | Milkyway@Home | Computation for task de_modfit_86_bundle4_4s_south4s_bgset_3_1603804501_15755378_1 finished
17/11/2020 3:29:29 PM | Milkyway@Home | Starting task de_modfit_82_bundle4_4s_south4s_bgset_3_1603804501_15695002_1
17/11/2020 3:30:41 PM | Milkyway@Home | Sending scheduler request: To fetch work.
17/11/2020 3:30:41 PM | Milkyway@Home | Reporting 1 completed tasks
17/11/2020 3:30:41 PM | Milkyway@Home | Requesting new tasks for CPU and NVIDIA GPU and AMD/ATI GPU
17/11/2020 3:30:44 PM | Milkyway@Home | Scheduler request completed: got 0 new tasks
17/11/2020 3:30:44 PM | Milkyway@Home | No tasks sent
17/11/2020 3:30:44 PM | Milkyway@Home | This computer has reached a limit on tasks in progress
17/11/2020 3:30:44 PM | Milkyway@Home | Project requested delay of 91 seconds
17/11/2020 3:31:04 PM | Milkyway@Home | Computation for task de_modfit_82_bundle4_4s_south4s_bgset_3_1603804501_15695002_1 finished
17/11/2020 3:31:04 PM | Milkyway@Home | Starting task de_modfit_84_bundle4_4s_south4s_bgset_3_1603804501_15651708_1
17/11/2020 3:32:16 PM | Milkyway@Home | Sending scheduler request: To fetch work.
17/11/2020 3:32:16 PM | Milkyway@Home | Reporting 1 completed tasks
17/11/2020 3:32:16 PM | Milkyway@Home | Requesting new tasks for CPU and NVIDIA GPU and AMD/ATI GPU
17/11/2020 3:32:19 PM | Milkyway@Home | Scheduler request completed: got 0 new tasks
17/11/2020 3:32:19 PM | Milkyway@Home | No tasks sent
17/11/2020 3:32:19 PM | Milkyway@Home | This computer has reached a limit on tasks in progress
17/11/2020 3:32:19 PM | Milkyway@Home | Project requested delay of 91 seconds
17/11/2020 3:32:45 PM | Milkyway@Home | Computation for task de_modfit_84_bundle4_4s_south4s_bgset_3_1603804501_15651708_1 finished
17/11/2020 3:32:45 PM | Milkyway@Home | Starting task de_modfit_80_bundle4_4s_south4s_bgset_3_1603804501_15784330_1
17/11/2020 3:33:50 PM | Milkyway@Home | Sending scheduler request: To fetch work.
17/11/2020 3:33:50 PM | Milkyway@Home | Reporting 1 completed tasks

Dingo

#7
You must have the BOINC Manager "Computing Preferences" set to have a lot of work under "Store at least * Days of work" as you have over 758 tasks on this machine alone.  BOINC will process the work so that it can complete the tasks by the time the deadline runs out.  N_Body tasks take about three times longer to run than Separator so it is running those first I suspect.

I would suggest that you decrease the the values you have in the preferences under  "Store at least * Days of work" and if you have anything in "Store up to * additional days work". 


Radioactive@home graph
Have a look at the BOINC@AUSTRALIA Facebook Page and join and also the Twitter Page.

Proud Founder and member of BOINC@AUSTRALIA

My Luck Prime 1,056,356 digits.
Have a look at my  Web Cam of Parliament House Ottawa, CANADA

TouchuvGrey

Yes, i did have it set to store a lot of work under "Store at least" and
"store up to" additional days. That was my  ( i thought ) way of getting around
any server outages or work shortages. It would seem that once again i
thought i was being clever but stuffed up. i have adjusted them.

chooka03

No, there is merit in doing that. Milkyway goes down each day which is frustrating for GPU work with a fast card because you run out of work but that's not the same for CPU tasks.
FWIW, most people don't bother with Milkyway CPU work. The credits are woeful. If you have a preferred CPU project, Id crunch that and leave the GPU on M@H.

:cheers:

TouchuvGrey

i crunched for SETI for 16 years but they are now shut down

i've been a space nut since i read my first SF book, sometime
around  1963 or 1964 so it would likely be either Cosmology or
Einstein. How are they for CPU credits ?

chooka03

Cosmology is a low credit project but I like it. Its stable with consistent work.  :thumbsup:

Honestly...I'm not sure how Eistein@Home credit goes for CPU tasks. I've thought about trying some one day but 100% of my E@H credits is from GPU crunching. I'd say 99% of users on crunch GPU for E@H.

Universe@Home is another good CPU project. Give that a look  :thumbsup:

tazzduke

You could also give try Asteroids, but chooka03 beat me with the Universe@home suggestion.

Problem with Cosmology, not to sure they are actually doing any constructive work, not much has been presented on the Website lately.

But we all have our own favourites to crunch though.

Happy Crunching.



 AA 24 - 53 participant

TouchuvGrey

Ridiculous though it may be after 16 ( or more, things get blurry back about then ) years
of crunching for SETI@Home i was pretty emotionally involved. i really believe we are not alone
and hope to see proof while i'm still here. i was very upset when SETI shut down.

      i will probably go with universe for my CPU work if i change at all.

Dataman

Of the projects discussed, Universe has the best CPU credits, almost always has work and supports ARM. It is a "set it and forget it" project.

Asteroids has good GPU credits but they have server capacity problems which they control by stopping new wu generation until the work in progress drops below a certain level. I frequently run it together with Milkyway or Einstein to keep the GPU's busy.

Running a large wu cache is counterproductive (imo). A setting of 1.0 is sufficient unless you do not look at your machines on a daily basis.

:cheers: