I manually built powertop from the tip of the tree then ran it on my Ubuntu 20.04 with the objective to get the total watts drawn by my PC.
Latest version didn't show watts at all, but ms/s etc in Overview section.
How do we interpret the output to come up with total watts? Is it possible at all?
I'm on Ubuntu 20.04.
I was on 5.8.16-050816-generic, this kernel version was installed by downloading from Ubuntu's mainline.
Just recently, I manually build kernel 5.14.0-next-20210913+ grabbed from upstream linux-next tree. I did update initrd, and grub. But one thing I notice and curious. Why initrd.img still points to the old kernel version, but vmlinuz points to the newest one. See below.
$ ls -lah /boot
total 396M
drwxr-xr-x 3 root root 4.0K Sep 14 22:05 .
drwxr-xr-x 20 root root 4.0K Dec 18 2020 ..
-rw-r--r-- 1 root root 253K Sep 14 21:18 config-5.14.0-next-20210913+
-rw-r--r-- 1 root root 243K Apr 12 20:47 config-5.8.0-50-generic
-rw-r--r-- 1 root root 242K Oct 17 2020 config-5.8.16-050816-generic
drwxr-xr-x 4 root root 4.0K Sep 14 21:43 grub
lrwxrwxrwx 1 root root 32 Jun 6 19:41 initrd.img -> initrd.img-5.8.16-050816-generic
-rw-r--r-- 1 root root 71M Sep 14 22:05 initrd.img-5.14.0-next-20210913+
-rw-r--r-- 1 root root 75M Jul 27 06:48 initrd.img-5.8.0-50-generic
-rw-r--r-- 1 root root 51M Jun 6 17:26 initrd.img-5.8.0-50-generic.old-dkms
-rw-r--r-- 1 root root 74M Sep 4 10:00 initrd.img-5.8.16-050816-generic
-rw-r--r-- 1 root root 74M Jul 27 06:48 initrd.img-5.8.16-050816-generic.old-dkms
lrwxrwxrwx 1 root root 27 Jun 6 19:16 initrd.img.old -> initrd.img-5.8.0-50-generic
-rw-r--r-- 1 root root 179K Aug 18 2020 memtest86+.bin
-rw-r--r-- 1 root root 181K Aug 18 2020 memtest86+.elf
-rw-r--r-- 1 root root 181K Aug 18 2020 memtest86+_multiboot.bin
-rw-r--r-- 1 root root 5.6M Sep 14 21:18 System.map-5.14.0-next-20210913+
-rw------- 1 root root 5.3M Apr 12 20:47 System.map-5.8.0-50-generic
-rw------- 1 root root 5.4M Oct 17 2020 System.map-5.8.16-050816-generic
lrwxrwxrwx 1 root root 29 Sep 14 21:18 vmlinuz -> vmlinuz-5.14.0-next-20210913+
-rw-r--r-- 1 root root 14M Sep 14 21:18 vmlinuz-5.14.0-next-20210913+
-rw------- 1 root root 9.4M Apr 12 22:17 vmlinuz-5.8.0-50-generic
-rw------- 1 root root 14M Oct 17 2020 vmlinuz-5.8.16-050816-generic
lrwxrwxrwx 1 root root 29 Jun 6 19:41 vmlinuz.old -> vmlinuz-5.8.16-050816-generic
What might be the issue?
Is it safe for me to manually re-link symlink of initrd.img myself to point to the correct one?
My main goal is to make Blender 2.82a able to detect my system's GPU then use as OpenCL compute to fasten the time in rendering.
The problem seems to be my Ubuntu distribution version is too high as I have 19.10 (with kernel 5.5.11-050511-lowlatency) + AMDGPU-PRO's OpenCL component can only officially support up to Ubuntu 18.04, same goes to ROCm.
I did try install AMDGPU-PRO's OpenCL via its installation script targeting only OpenCL component (via --headless), and even manually install one by one package; all with no success. It is clear as stated, not support shown up in error message. I would like to try again with ROCm but its package seems to be large. With no pre-built package available for my distro, I guess it will fail the same.
So what would be my options if going back to 18.04 would be a last resort in installing OpenCL driver for my GPU (APU from Ryzen 3 2200g to be exact, I have no discrete GPU) alongside already installed open-source OpenGL/Vulkan driver from Mesa ?
Edit:
In the mean time, I did what I can to test things out by installing OpenCL 1.1 via libclang-common-10-dev libclc-amdgcn libclc-dev libclc-r600 libllvm10 mesa-opencl-icd packages. It shows meaningful information via clinfo. But Blender still didn't detect any GPU as OpenCL device.
Update 1:
I tested with ROCm. With good reports from here, and here, it becomes clear Ubuntu release version of 19.10 should not be a problem, only just need a proper kernel version. Thus I reverted my kernel version back to 5.3.x (specifically use 5.3.18 generic as 5.3 is stated as a supported version for ROCm, and from the report).
Note that I minimally install packages as stated in 2nd report link as ROCm packages are huge then follow the less of installation instruction from README.md of its github page. Although I'm on eoan and the package is from xenial, I do take the risk for now. All passed. I can validate via clinfo or its /opt/rocm-3.1.0/opencl/bin/x86_64/clinfo both give meaningful result that I now have OpenCL 2.1 but with NO device found!!!.
Result from clinfo
Number of platforms 1
Platform Name AMD Accelerated Parallel Processing
Platform Vendor Advanced Micro Devices, Inc.
Platform Version OpenCL 2.1 AMD-APP (3084.0)
Platform Profile FULL_PROFILE
Platform Extensions cl_khr_icd cl_amd_event_callback cl_amd_offline_devices
Platform Host timer resolution 1ns
Platform Extensions function suffix AMD
Platform Name AMD Accelerated Parallel Processing
Number of devices 0
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) AMD Accelerated Parallel Processing
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)
clCreateContext(NULL, ...) [default] No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) No devices found in platform
ICD loader properties
ICD loader Name OpenCL ICD Loader
ICD loader Vendor OCL Icd free software
ICD loader Version 2.2.11
ICD loader Profile OpenCL 2.1
Result from /opt/rocm-3.1.0/opencl/bin/x86_64/clinfo
Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 2.1 AMD-APP (3084.0)
Platform Name: AMD Accelerated Parallel Processing
Platform Vendor: Advanced Micro Devices, Inc.
Platform Extensions: cl_khr_icd cl_amd_event_callback cl_amd_offline_devices
Platform Name: AMD Accelerated Parallel Processing
ERROR: clGetDeviceIDs(-1)
Asides testing with Blender still doesn't detect GPU, that's right as now I'm having a problem with device not found.
I'm further investigating ...
Update 2:
Problem solved although performance is not good. See my answer of this question below.
I did found this similar question but the problem still pops up now and then later after executing such command.
The problem came back after I open random applications. Sometimes after opening google chrome, sometimes OpenGL-based application, sometimes a certain applications which eat much of the RAM & CPU.
Just for the record, the problem is while I'm holding down the key, it keeps repeating the key in which I don't want this behavior to happen.
For reference, my system is 5.0.0-37-generic #40~18.04.1-Ubuntu SMP Thu Nov 14 12:06:39 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux on xfce.
Update
Repeat Delay to be highest and Repeat speed to be lowest. Anyway, it doesn't completely solve the problem but at least not let you face the full blown repeating key problem.
I have Nikon D3300 camera and want to feed video output into my computer as part of live streaming with OBS.
I successfully installed v4l2loopback as well as gphoto2 on my Ubuntu 18.04. I already had ffmpeg installed. (Their all version information can be found at the bottom)
Then I execute the command gphoto2 --stdout --capture-movie | ffmpeg -i - -vcodec rawvideo -pix_fmt yuv420p -threads 0 -f v4l2 /dev/video2. I tested to see video output via guvcview and OBS. It shows video output properly but with one issue. It looks greenish like this (captured from OBS).
I also asked in v4l2loopback on github and developer pin-points that the problem is with the parameters feeding the program i.e. gphoto2 (or need some adjustment of parameters in ffmpeg too or not I'm not sure).
So do you know whether I use the right parameters? or else what's the right ones to solve this greenish video output.