Message boards :
Number crunching :
OpenCL and Mint 20.2
Message board moderation
Author | Message |
---|---|
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Hello all I have installed Linux Mint 20.2 and also Rocm drivers for OpenCL but Boinc reports no usable GPU found. OpenCL is detected by CLinfo. I know this has been discuesed before but Mint 20 uses other drivers than previous versions. I remember some libs has to be copied to the boinc folder but can`t remember which. I hope some of you Linux guys can point me to it. regards Mike With each crime and every kindness we birth our future. |
Kiska Send message Joined: 31 Mar 12 Posts: 302 Credit: 3,067,762 RAC: 0 |
Hello all You'd probably want the opencl-icd package installed. mesa-opencl-icd for amd devices and libnvidia-compute-<version> for nvidia |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Hello all Mesa is already installed. Like i said CLinfo reports my GPU correctly. With each crime and every kindness we birth our future. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Hello all what GPU do you have? AMD? Nvidia? Intel? What model exactly What version of BOINC? Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
The GPU is a AMD R9 380 OpenCL 2.0. Boinc 7.16.6 Hello all With each crime and every kindness we birth our future. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
The GPU is a AMD R9 380 OpenCL 2.0. what drivers did you install? I've always heard that mesa drivers do not work for most projects. you'll need either the ROCm drivers or the amdgpu-pro what project are you trying to contribute to? can you post your complete clinfo output? but I have come across several instances where under Linux the AMD GPU is not detected by BOINC, even when installed. the detection seemed to work OK under newer versions of BOINC however. can you install a newer version? Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Like i posted already Rocm drivers are installed. Even a project would not run the GPU has to be detected by Boinc. Here my CLinfo Platform Name: AMD Accelerated Parallel Processing Number of devices: 1 Device Type: CL_DEVICE_TYPE_GPU Vendor ID: 1002h Board name: Tonga PRO [Radeon R9 285/380] Device Topology: PCI[ B#41, D#0, F#0 ] Max compute units: 28 Max work items dimensions: 3 Max work items[0]: 1024 Max work items[1]: 1024 Max work items[2]: 1024 Max work group size: 256 Preferred vector width char: 4 Preferred vector width short: 2 Preferred vector width int: 1 Preferred vector width long: 1 Preferred vector width float: 1 Preferred vector width double: 1 Native vector width char: 4 Native vector width short: 2 Native vector width int: 1 Native vector width long: 1 Native vector width float: 1 Native vector width double: 1 Max clock frequency: 985Mhz Address bits: 64 Max memory allocation: 3650722200 Image support: Yes Max number of images read arguments: 128 Max number of images write arguments: 8 Max image 2D width: 16384 Max image 2D height: 16384 Max image 3D width: 16384 Max image 3D height: 16384 Max image 3D depth: 8192 Max samplers within kernel: 26937 Max size of kernel argument: 1024 Alignment (bits) of base address: 1024 Minimum alignment (bytes) for any datatype: 128 Single precision floating point capability Denorms: No Quiet NaNs: Yes Round to nearest even: Yes Round to zero: Yes Round to +ve and infinity: Yes IEEE754-2008 fused multiply-add: Yes Cache type: Read/Write Cache line size: 64 Cache size: 16384 Global memory size: 4294967296 Constant buffer size: 3650722200 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 65536 Max pipe arguments: 16 Max pipe active reservations: 16 Max pipe packet size: 3650722200 Max global variable size: 3650722200 Max global variable preferred total size: 4294967296 Max read/write image args: 64 Max on device events: 1024 Queue on device max size: 8388608 Max on device queues: 1 Queue on device preferred size: 262144 SVM capabilities: Coarse grain buffer: Yes Fine grain buffer: Yes Fine grain system: No Atomics: No Preferred platform atomic alignment: 0 Preferred global atomic alignment: 0 Preferred local atomic alignment: 0 Kernel Preferred work group size multiple: 64 Error correction support: 0 Unified memory for Host and Device: 0 Profiling timer resolution: 1 Device endianess: Little Available: Yes Compiler available: Yes Execution capabilities: Execute OpenCL kernels: Yes Execute native function: No Queue on Host properties: Out-of-Order: No Profiling : Yes Queue on Device properties: Out-of-Order: Yes Profiling : Yes Platform ID: 0x7f205c302e10 Name: gfx802 Vendor: Advanced Micro Devices, Inc. Device OpenCL C version: OpenCL C 2.0 Driver version: 3305.0 (HSA1.1,LC) Profile: FULL_PROFILE Version: OpenCL 1.2 Extensions: cl_khr_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_amd_device_attribute_query cl_amd_media_ops cl_amd_media_ops2 cl_khr_image2d_from_buffer cl_khr_subgroups cl_khr_depth_images cl_amd_copy_buffer_p2p cl_amd_assembly_program With each crime and every kindness we birth our future. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
Like i posted already Rocm drivers are installed. You actually posted previously that Mesa drivers were installed. ROCm and Mesa are not the same thing. but if you have ROCm, then that leads me to believe it's the same issue that I experienced previously. what version of ROCm did you install? the latest 4.3.1? have you tried installing a more recent version of BOINC? this was the easiest way to fix the issue. I have a workaround that should work if the newer version of BOINC doesnt work. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Like i posted already Rocm drivers are installed. I have installed Linux Mint 20.2 and also ---> Rocm drivers for OpenCL ]but Boinc reports no usable GPU found. OpenCL is detected by CLinfo. I know this has been discuesed before but Mint 20 uses other drivers than previous versions. I remember some libs has to be copied to the boinc folder but can`t remember which. I can`t find a newer boinc version. With each crime and every kindness we birth our future. |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
check the /etc/OpenCL/vendors/ directory. you should find an .icd file there named "amdocl64_40200.icd"(*). and is it the only file in this directory? if you're on a fresh install and have only tried the ROCm install, I imagine it's the only file there right now. next check your /opt/rocm/opencl/lib/ directory and verify that the libamdocl64.so file is in there. if not please let me know. open the amdocl64_40200.icd(*) file with nano to edit: sudo nano /etc/OpenCL/vendors/amdocl64_40200.icd contents is likely just "libamdocl64.so" change the contents to "/opt/rocm/opencl/lib/libamdocl64.so" (without the quotes) [Ctrl]+[x] to exit, you will be prompted to save, enter [y], and hit [Enter] to verify filename (don't change it) and it will save and close. Then reboot and retry. (*)note, the suffix of the libamdocl64.icd file (in the above instructions it’s “40200” referring to ROCm v4.2 which is what I have) might be different if ROCm has been updated in the repository. Be sure to check the file names that exist and make the necessary modifications to the given instructions. Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Works like a charm, thank you very much. Rocm version is 4.3 so filename has the ending 40300 but no big deal. With each crime and every kindness we birth our future. |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
I ran into another issue here. Since i am using win 10 i have an issue with Linux Mint. I have Mint installed on 2 different hard drives. Mint 20.2 on Nvme 970 Evo plus 2TB andMint 20.3 on Nvme 960 Evo. Everything is running fine but each time i am working on win and reboot into Mint it is hanging in initramfs and i have to repair the filesystem with gparted. Fast boot is deactivated already. I know we have some Linux geeks here, maybe someone has got an idea how to solve this. With each crime and every kindness we birth our future. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Check your /boot/grub/grub.cfg file and /etc/fstab file to make sure the grub entry for each Mint version is pointing to the correct M.2 drive UUID. Run sudo blkid to find the UUID of each M.2 drive. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Thanks Keith, i will try this on my next Linux session. With each crime and every kindness we birth our future. |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
It`s a little weird. What i can see is that the second nvme drive has different drive names. Yesterday it was sde5 but today it was sdf5. It depends on the USB stick with the live system. I just don`t understand why it corrupts the filesystem all the time. Would it help to disconnect the second nvme drive before starting Mint from the system drive? With each crime and every kindness we birth our future. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I think it is a bit much to have to open the computer and pull out the nvme drive from the slot every time you want to boot the other Mint version. The drive names shouldn't be changing for the nvme drives. You mention a USB drive though and I am confused. USB drive names will change if you plug them into a different port. The drive names shouldn't be used for the grub entry or the fstab entries. That is why I suggested you define a drive by its UUID. The UUID will not change. It is like a fingerprint identifying the drive no matter where it is installed. You didn't by chance clone the second nvme drive from the 1st did you. That would make both drives have the same UUID and I can see why you are having issues. That would be a big no-no. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
The second nvme drive is plugged via USB because it is the old one 500GB the new one has 2TB. My Crosshair VI only has one M2 slot. Yes, it is cloned but the UUID is different anyways. UUID is correct for all drives in grub.cfg. Maybe i should remove the Linux partition on the old Nvme. With each crime and every kindness we birth our future. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Oh, I get it. You have installed a nvme drive in a USB carrier. I thought it was installed in a motherboard M.2 slot. Have you looked in the USB hosted nvme drive via Disks to see if it has a FAT UEFI partition on it? You may be having conflicts with two UEFI partitions hosting two different grub configs. You should have only one in play. Whatever Mint installation is your normal default one probably should support the grub file that boots both versions of Mint. If you can get both Mint versions visible to your default OS, I would run os-prober and see if you can get it to write a correct grub file for both Mint entries. Barring getting that to work, I would do some editing in the default /boot/grub/grub.cfg file to add the second Mint version. You could just copy the block for the normal Mint OS entry behind the original one and just change the UUID entry for the second Mint version. Then go into your /etc/default/grub file and change the GRUB_DEFAULT= line to point at the preferred Mint OS entry. Remember to do a update-grub afterwards. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
Yep, i bought a Icy Box to carry the 960 Evo because i want to use it as a backup drive. I only use the grub from the GBT drive 970 Evo Plus which has win 10 and Mint 20.3 (upgraded) It contains all Osses and i can boot from it to win 10, Mint 20.3, Mint 18.3 (850 Evo) and Mint 20.3 from the 960 Evo. The new 970 Evo Plus is the only GBT drive in my system off course. So no conflicts there. But i got another issue since i upgraded to Mint 20.3. I did a update and it installed Kernel 5.13.28 before upgrading to Mint 20.3. All works fine so far except i am back to no usable GPU`s found. I reinstalled Rocm but it seems it doesn`t like the kernel. CLInfo says no device found in plattform. I made all the changes Ian suggested earlier in this thread but now i have 2 entries in /etc/OpenCL/vendors/ directory. amdocl64_40300.icd and amdocl64_40500.icd one for Rocm ver 4.3 and one for 4.5 So i deinstalled it and made a clean install again. With each crime and every kindness we birth our future. |
Mike Send message Joined: 17 Feb 01 Posts: 34380 Credit: 79,922,639 RAC: 80 |
I found this after Rocm was updated to 5.0 The installation process for ver 4.5 apparently does not support the default linux kernel 5.13.0-27-generic. (Apparently the previous version rocm 4.3 does not have this issue.) Yes, Ubuntu 20.04 updated HWE kernel to 5.13 which will be supported in the next ROCm release. So i will just wait. Do i have to remove the old files from 4.3 and 4.5 and if yes how ? With each crime and every kindness we birth our future. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.