PC Graphics

[mdclemen]

According to this datasheet: https://www.amd.com/system/files/documents/firepro-s-series-datasheet.pdf, the FirePro S7150 should be OpenCL 2.0 capable, but I've learned that often AMD doesn't necessarily support a lot of features with their Linux drivers (for example the FirePro W5100 is not OpenCL 2.0 capable on Linux, even though AMD advertises that it is: https://community.amd.com/t5/firepro-development/opencl-2-0-on-amd-firepro-w5100/m-p/683011#M747). Can anyone confirm that this GPU is OpenCL 2.0 capable on Linux? Thanks.

[fsadough]

The question is why you need OpenCL 2.0. Which specific OpenCL extension are you looking for? Anything beyond OpenCL 1.2 is made optional.

[mdclemen]

I would like to use Shared Virtual Memory with atomic support, which is only available in OpenCL 2.0 and after.

[mdclemen]

From digging around the OpenCL 2.0 spec, and this directory of extensions, there should be an extension called something like cl_amd_shared_virtual_memory though it doesn't exist in the /amd/ directory (something similar exists in the /arm/ directory: cl_arm_shared_virtual_memory). I was able to add the extension cl_amd_bus_addressable_memory by adding amdgpu.direct_gma_size=96 as a kernel boot parameter. Is there something similar I can do here to enable SVM?

[fsadough]

It's an extension to enable SVM on early versions of OpenCL. AMD doesn't need this extension, because we support OpenCL2.0:

registry.khronos.org/OpenCL/extensions/arm/cl_arm_shared_virtual_memory.txt

This extension enables the shared virtual memory features in versions prior to 2.0 as described in OpenCL version 2.0 section 3.

[mdclemen]

I'll be super explicit about what it is I am looking for. I currently have a FirePro W5100 on a linux machine (AlmaLinux 8.10, kernel 4.18.0-240.el8.x86_64) and here are some pertinent fields from clinfo:

$ clinfo
  Platform Version:				 OpenCL 2.1 AMD-APP (3224.4)
  Board name:					 AMD FirePro W5100
  SVM capabilities:				 
    Coarse grain buffer:			 No
    Fine grain buffer:				 No
    Fine grain system:				 No
    Atomics:					 No
  Device OpenCL C version:			 OpenCL C 1.2 
  Driver version:				 3224.4
  Version:					 OpenCL 1.2 AMD-APP (3224.4)

 Elsewhere on the forums @fsadough informed me that OpenCL 2.0 is not fully implemented in linux, even though AMD's own page for this GPU says it supports OpenCL 2.0. I'd like to figure out how to change the "No"s under SVM capablities to "Yes"s. Since you "support OpenCL2.0", how do I enable SVM capabilities on linux for GPUs which AMD says OpenCL 2.0 is supported (this would include the FirePro S7150)?

[fsadough]

Where did you obtain the driver from? Can you please share the link?

[mdclemen]

I am using this driver: https://drivers.amd.com/drivers/linux/amdgpu-pro-21.10-1263777-rhel-8.3.tar.xz

Which I got from this page: https://www.amd.com/en/support/downloads/drivers.html/graphics/firepro/firepro-wx100-series/firepro-...

This the most up-to-date 64-bit linux driver that AMD offers for the FirePro series of GPUs.

I also dug around online and found that there is an extension called cl_amd_svm. Do you have any information on that extension? I found it referenced and used in this version of clinfo (used in Arch Linux): https://github.com/Oblomov/clinfo.

[fsadough]

I checked with our OpenCL team. Both W5100 and S7150X2 are EOL'ed and we have no more support on those products. I am afraid, you are on your own.

[mdclemen]

It's completely ridiculous to me that AMD would advertise these GPUs as OpenCL 2.0 capable, and then not offer that capability for the entirety of these products' life. Even though I wish CUDA was open source like OpenCL, I guess it's back to NVIDIA, they are at least true to their word and their products work as advertised.

[fsadough · Accepted solution]

It works with older driver (amdgpu-install_22.10.2.50102-1_all.deb). I tested it on FirePro W7100 (same ASIC as S7150X2).

ryzentr@X399:~$ lsb_release -a

No LSB modules are available.

Distributor ID:    Ubuntu

Description:        Ubuntu 18.04.6 LTS

Release:               18.04

Codename:         bionic

ryzentr@X399:~$ uname -r

5.4.0-150-generic

$ sudo apt-get update
$ sudo apt-get dist-upgrade

$ sudo apt-get install ./amdgpu-install_22.10.2.50102-1_all.deb
$ sudo apt-get update

$ amdgpu-install --usecase=workstation -y --opencl=rocr,legacy

ryzentr@X399:~$ clinfo

Number of platforms                               1

  Platform Name                                   AMD Accelerated Parallel Processing

  Platform Vendor                                 Advanced Micro Devices, Inc.

  Platform Version                                OpenCL 2.1 AMD-APP (3380.4)

  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                                 1

  Device Name                                     Tonga

  Device Vendor                                   Advanced Micro Devices, Inc.

  Device Vendor ID                                0x1002

  Device Version                                  OpenCL 2.0 AMD-APP (3380.4)

  Driver Version                                  3380.4 (PAL,HSAIL)

  Device OpenCL C Version                         OpenCL C 2.0

  Device Type                                     GPU

  Device Board Name (AMD)                         AMD FirePro W7100

  Device Topology (AMD)                           PCI-E, 41:00.0

  Device Profile                                  FULL_PROFILE

  Device Available                                Yes

  Compiler Available                              Yes

  Linker Available                                Yes

  Max compute units                               28

  SIMD per compute unit (AMD)                     4

  SIMD width (AMD)                                16

  SIMD instruction width (AMD)                    1

  Max clock frequency                             920MHz

  Graphics IP (AMD)                               8.0

  Device Partition                                (core)

    Max number of sub-devices                     28

    Supported partition types                     None

  Max work item dimensions                        3

  Max work item sizes                             1024x1024x1024

  Max work group size                             256

  Preferred work group size (AMD)                 256

  Max work group size (AMD)                       1024

  Preferred work group size multiple              64

  Wavefront width (AMD)                           64

  Preferred / native vector sizes                

    char                                                 4 / 4      

    short                                                2 / 2      

    int                                                  1 / 1      

    long                                                 1 / 1      

    half                                                 1 / 1        (cl_khr_fp16)

    float                                                1 / 1      

    double                                               1 / 1        (cl_khr_fp64)

  Half-precision Floating-point support           (cl_khr_fp16)

    Denormals                                     No

    Infinity and NANs                             No

    Round to nearest                              No

    Round to zero                                 No

    Round to infinity                             No

    IEEE754-2008 fused multiply-add               No

    Support is emulated in software               No

  Single-precision Floating-point support         (core)

    Denormals                                     No

    Infinity and NANs                             Yes

    Round to nearest                              Yes

    Round to zero                                 Yes

    Round to infinity                             Yes

    IEEE754-2008 fused multiply-add               Yes

    Support is emulated in software               No

    Correctly-rounded divide and sqrt operations  Yes

  Double-precision Floating-point support         (cl_khr_fp64)

    Denormals                                     Yes

    Infinity and NANs                             Yes

    Round to nearest                              Yes

    Round to zero                                 Yes

    Round to infinity                             Yes

    IEEE754-2008 fused multiply-add               Yes

    Support is emulated in software               No

  Address bits                                    64, Little-Endian

  Global memory size                              8589934592 (8GiB)

  Global free memory (AMD)                        8323072 (7.938GiB)

  Global memory channels (AMD)                    8

  Global memory banks per channel (AMD)           4

  Global memory bank width (AMD)                  256 bytes

  Error Correction support                        No

  Max memory allocation                           7073274265 (6.587GiB)

  Unified memory for Host and Device              No

  Shared Virtual Memory (SVM) capabilities        (core)

    Coarse-grained buffer sharing                 Yes

    Fine-grained buffer sharing                   Yes

    Fine-grained system sharing                   No

    Atomics                                       No

  Minimum alignment for any data type             128 bytes

  Alignment of base address                       2048 bits (256 bytes)

  Preferred alignment for atomics                

    SVM                                           0 bytes

    Global                                        0 bytes

    Local                                         0 bytes

  Max size for global variable                    6365946624 (5.929GiB)

  Preferred total size of global vars             8589934592 (8GiB)

  Global Memory cache type                        Read/Write

  Global Memory cache size                        16384 (16KiB)

  Global Memory cache line size                   64 bytes

  Image support                                   Yes

    Max number of samplers per kernel             16

    Max size for 1D images from buffer            442079641 pixels

    Max 1D or 2D image array size                 2048 images

    Base address alignment for 2D image buffers   256 bytes

    Pitch alignment for 2D image buffers          256 pixels

    Max 2D image size                             16384x16384 pixels

    Max 3D image size                             2048x2048x2048 pixels

    Max number of read image args                 128

    Max number of write image args                64

    Max number of read/write image args           64

  Max number of pipe args                         16

  Max active pipe reservations                    16

  Max pipe packet size                            2778306969 (2.587GiB)

  Local memory type                               Local

  Local memory size                               65536 (64KiB)

  Local memory syze per CU (AMD)                  65536 (64KiB)

  Local memory banks (AMD)                        32

  Max number of constant args                     8

  Max constant buffer size                        7073274265 (6.587GiB)

  Preferred constant buffer size (AMD)            16384 (16KiB)

  Max size of kernel argument                     1024

  Queue properties (on host)                     

    Out-of-order execution                        No

    Profiling                                     Yes

  Queue properties (on device)                   

    Out-of-order execution                        Yes

    Profiling                                     Yes

    Preferred size                                262144 (256KiB)

    Max size                                      8388608 (8MiB)

  Max queues on device                            1

  Max events on device                            1024

  Prefer user sync for interop                    Yes

  Number of P2P devices (AMD)                     0

  P2P devices (AMD)                               <printDeviceInfo:144: get number of CL_DEVICE_P2P_DEVICES_AMD : error -30>

  Profiling timer resolution                      1ns

  Profiling timer offset since Epoch (AMD)        1721393667151371184ns (Fri Jul 19 14:54:27 2024)

  Execution capabilities                         

    Run OpenCL kernels                            Yes

    Run native kernels                            No

    Thread trace supported (AMD)                  Yes

    Number of async queues (AMD)                  4

    Max real-time compute queues (AMD)            1

    Max real-time compute units (AMD)             0

  printf() buffer size                            4194304 (4MiB)

  Built-in kernels                               

  Device 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_image2d_from_buffer cl_khr_subgroups cl_khr_gl_event cl_khr_depth_images cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_amd_copy_buffer_p2p

NULL platform behavior

  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  No platform

  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   No platform

  clCreateContext(NULL, ...) [default]            No platform

  clCreateContext(NULL, ...) [other]              Success [AMD]

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (1)

    Platform Name                                 AMD Accelerated Parallel Processing

    Device Name                                   Tonga

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform

  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  Success (1)

    Platform Name                                 AMD Accelerated Parallel Processing

    Device Name                                   Tonga

  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)  Success (1)

    Platform Name                                 AMD Accelerated Parallel Processing

    Device Name                                   Tonga

ryzentr@X399:~$

[mdclemen]

I think what you meant to say is that the FirePro W7100 has the same ASIC (TONGA) as the FirePro S7150 and not the FirePro W5100 (BONAIRE).

[fsadough]

Correct. My mistake Same as S7150X2. W5100 has a different ASIC but same driver.

[mdclemen]

I am on a RHEL 8 clone, but this driver (amdgpu-install_22.10.2.50102-1) has a RHEL 8 version.

$ cat /etc/redhat-release
AlmaLinux release 8.10 (Cerulean Leopard)
$ uname -r
4.18.0-348.el8.x86_64 # downgraded the kernel to 4.18.0-348.el8.x86_64 (RHEL 8.5) so it works with the driver 

I installed it:

$ sudo dnf localinstall ./amdgpu-install-22.10.2.50102-1.el8.noarch.rpm
$ sudo dnf update
$ sudo amdgpu-install --usecase=workstation -y --opencl=rocr,legacy
$ clinfo
Number of platforms: 1
Platform Profile: FULL_PROFILE
Platform Version: OpenCL 2.1 AMD-APP (3380.4)
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
Number of devices: 0

The driver doesn't seem to support the W5100 for OpenCL, even though it is there and the driver is loaded in the kernel:

$ lspci | grep -i amd
03:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Bonaire [FirePro W5100]
$ lsmod | grep amd
amdgpu               7892992  0
amddrm_ttm_helper      16384  1 amdgpu
amdttm                 73728  2 amdgpu,amddrm_ttm_helper
amdkcl                 28672  2 amdttm,amdgpu
iommu_v2               20480  1 amdgpu
i2c_algo_bit           16384  1 amdgpu
amd_sched              40960  1 amdgpu
drm_kms_helper        253952  3 amdgpu
drm                   573440  6 drm_kms_helper,amd_sched,amdttm,amdgpu,amddrm_ttm_helper
$ dmesg | grep "amdgpu version"
[    3.777434] [drm] amdgpu version: 5.13.20.22.10

 So this didn't work for me.

[fsadough]

Can you try this?
Radeon™ Software for Linux® 22.40.6 Release Notes (amd.com)

[mdclemen]

That didn't work either (same clinfo output as before). My system is normally headless, but I tried seeing if this driver would even drive a display and that doesn't even work. Lightdm (display manager) wont even start because the X server crashes. Output of Xorg.0.log:

[    22.344] 
X.Org X Server 1.20.11
X Protocol Version 11, Revision 0
[    22.344] Build Operating System:  4.18.0-513.18.1.el8_9.x86_64 
[    22.344] Current Operating System: Linux hero 4.18.0-425.3.1.el8.x86_64 #1 SMP Tue Nov 8 14:08:25 EST 2022 x86_64
[    22.344] Kernel command line: BOOT_IMAGE=(hd1,gpt2)/vmlinuz-4.18.0-425.3.1.el8.x86_64 root=/dev/mapper/kaiju-root8 ro crashkernel=auto resume=/dev/mapper/kaiju-swap rd.md.uuid=5601f3cb:d031299b:790c6837:62fa15c6 rd.lvm.lv=kaiju/root8 rd.md.uuid=282cfe99:1b8397bc:34a099ff:e2f249f5 rd.lvm.lv=kaiju/swap radeon.cik_support=0 amdgpu.cik_support=1 irqaffinity=0 console=tty0, console=ttyS0,115200n8 amdgpu.virtual_display=0000:03:00.0,1 amdgpu.direct_gma_size=96 nohz_full=1 rcu_nocbs=1 mitigations=off
[    22.344] Build Date: 02 July 2024  02:54:42PM
[    22.344] Build ID: xorg-x11-server 1.20.11-24.el8_10 
[    22.344] Current version of pixman: 0.38.4
[    22.344] 	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
[    22.344] Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[    22.344] (==) Log file: "/var/log/Xorg.0.log", Time: Mon Jul 22 17:49:07 2024
[    22.344] (==) Using config directory: "/etc/X11/xorg.conf.d"
[    22.344] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[    22.344] (==) No Layout section.  Using the first Screen section.
[    22.344] (==) No screen section available. Using defaults.
[    22.344] (**) |-->Screen "Default Screen Section" (0)
[    22.344] (**) |   |-->Monitor "<default monitor>"
[    22.344] (==) No monitor specified for screen "Default Screen Section".
	Using a default monitor configuration.
[    22.344] (==) Automatically adding devices
[    22.344] (==) Automatically enabling devices
[    22.344] (==) Automatically adding GPU devices
[    22.344] (==) Automatically binding GPU devices
[    22.344] (==) Max clients allowed: 256, resource mask: 0x1fffff
[    22.344] (==) FontPath set to:
	catalogue:/etc/X11/fontpath.d,
	built-ins
[    22.344] (**) ModulePath set to "/opt/amdgpu-pro/lib/xorg/modules,/opt/amdgpu/lib/xorg/modules,/usr/lib64/xorg/modules"
[    22.344] (II) The server relies on udev to provide the list of input devices.
	If no devices become available, reconfigure udev or disable AutoAddDevices.
[    22.344] (II) Loader magic: 0x557eb6b8c020
[    22.344] (II) Module ABI versions:
[    22.344] 	X.Org ANSI C Emulation: 0.4
[    22.344] 	X.Org Video Driver: 24.1
[    22.344] 	X.Org XInput driver : 24.1
[    22.344] 	X.Org Server Extension : 10.0
[    22.345] (++) using VT number 1

[    22.345] (II) systemd-logind: logind integration requires -keeptty and -keeptty was not provided, disabling logind integration
[    22.345] (II) xfree86: Adding drm device (/dev/dri/card0)
[    22.345] (II) Platform probe for /sys/devices/pci0000:00/0000:00:02.0/0000:03:00.0/drm/card0
[    22.387] (--) PCI:*(3@0:0:0) 1002:6649:1028:230c rev 0, Mem @ 0x30100000000/4294967296, 0x30080000000/8388608, 0xf7e00000/262144, I/O @ 0x0000e000/256, BIOS @ 0x????????/131072
[    22.387] (II) LoadModule: "glx"
[    22.387] (II) Loading /opt/amdgpu-pro/lib/xorg/modules/extensions/libglx.so
[    22.394] (II) Module glx: vendor="X.Org Foundation"
[    22.394] 	compiled for 1.20.11, module version = 1.0.0
[    22.394] 	ABI class: X.Org Server Extension, version 10.0
[    22.394] (II) Applying OutputClass "AMDgpu" to /dev/dri/card0
[    22.394] 	loading driver: amdgpu
[    22.394] (==) Matched amdgpu as autoconfigured driver 0
[    22.394] (==) Matched ati as autoconfigured driver 1
[    22.394] (==) Matched modesetting as autoconfigured driver 2
[    22.394] (==) Matched fbdev as autoconfigured driver 3
[    22.394] (==) Matched vesa as autoconfigured driver 4
[    22.394] (==) Assigned the driver to the xf86ConfigLayout
[    22.394] (II) LoadModule: "amdgpu"
[    22.394] (II) Loading /opt/amdgpu/lib/xorg/modules/drivers/amdgpu_drv.so
[    22.395] (II) Module amdgpu: vendor="X.Org Foundation"
[    22.395] 	compiled for 1.20.11, module version = 22.0.0
[    22.395] 	Module class: X.Org Video Driver
[    22.395] 	ABI class: X.Org Video Driver, version 24.1
[    22.395] (II) LoadModule: "ati"
[    22.396] (WW) Warning, couldn't open module ati
[    22.396] (EE) Failed to load module "ati" (module does not exist, 0)
[    22.396] (II) LoadModule: "modesetting"
[    22.396] (II) Loading /usr/lib64/xorg/modules/drivers/modesetting_drv.so
[    22.396] (II) Module modesetting: vendor="X.Org Foundation"
[    22.396] 	compiled for 1.20.11, module version = 1.20.11
[    22.396] 	Module class: X.Org Video Driver
[    22.396] 	ABI class: X.Org Video Driver, version 24.1
[    22.396] (II) LoadModule: "fbdev"
[    22.396] (II) Loading /usr/lib64/xorg/modules/drivers/fbdev_drv.so
[    22.396] (II) Module fbdev: vendor="X.Org Foundation"
[    22.396] 	compiled for 1.20.3, module version = 0.5.0
[    22.396] 	Module class: X.Org Video Driver
[    22.396] 	ABI class: X.Org Video Driver, version 24.0
[    22.396] (II) LoadModule: "vesa"
[    22.397] (II) Loading /usr/lib64/xorg/modules/drivers/vesa_drv.so
[    22.397] (II) Module vesa: vendor="X.Org Foundation"
[    22.397] 	compiled for 1.20.3, module version = 2.4.0
[    22.397] 	Module class: X.Org Video Driver
[    22.397] 	ABI class: X.Org Video Driver, version 24.0
[    22.397] (II) AMDGPU: Driver for AMD Radeon:
	All GPUs supported by the amdgpu kernel driver
[    22.397] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[    22.397] (II) FBDEV: driver for framebuffer: fbdev
[    22.397] (II) VESA: driver for VESA chipsets: vesa
[    22.397] (WW) Falling back to old probe method for modesetting
[    22.398] (WW) Falling back to old probe method for fbdev
[    22.398] (II) Loading sub module "fbdevhw"
[    22.398] (II) LoadModule: "fbdevhw"
[    22.398] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so
[    22.398] (II) Module fbdevhw: vendor="X.Org Foundation"
[    22.398] 	compiled for 1.20.11, module version = 0.0.2
[    22.398] 	ABI class: X.Org Video Driver, version 24.1
[    22.398] (II) AMDGPU(0): Creating default Display subsection in Screen section
	"Default Screen Section" for depth/fbbpp 24/32
[    22.398] (==) AMDGPU(0): Depth 24, (--) framebuffer bpp 32
[    22.398] (II) AMDGPU(0): Pixel depth = 24 bits stored in 4 bytes (32 bpp pixmaps)
[    22.398] (==) AMDGPU(0): Default visual is TrueColor
[    22.398] (==) AMDGPU(0): RGB weight 888
[    22.398] (II) AMDGPU(0): Using 8 bits per RGB (8 bit DAC)
[    22.398] (--) AMDGPU(0): Chipset: "AMD FirePro W5100" (ChipID = 0x6649)
[    22.398] (II) Loading sub module "fb"
[    22.398] (II) LoadModule: "fb"
[    22.398] (II) Loading /usr/lib64/xorg/modules/libfb.so
[    22.398] (II) Module fb: vendor="X.Org Foundation"
[    22.398] 	compiled for 1.20.11, module version = 1.0.0
[    22.398] 	ABI class: X.Org ANSI C Emulation, version 0.4
[    22.398] (II) Loading sub module "dri2"
[    22.398] (II) LoadModule: "dri2"
[    22.398] (II) Module "dri2" already built-in
[    22.450] (II) Loading sub module "glamoregl"
[    22.450] (II) LoadModule: "glamoregl"
[    22.450] (II) Loading /usr/lib64/xorg/modules/libglamoregl.so
[    22.453] (II) Module glamoregl: vendor="X.Org Foundation"
[    22.453] 	compiled for 1.20.11, module version = 1.0.1
[    22.453] 	ABI class: X.Org ANSI C Emulation, version 0.4
[    22.457] (EE) 
[    22.457] (EE) Backtrace:
[    22.457] (EE) 0: /usr/libexec/Xorg (xorg_backtrace+0x7d) [0x557eb68fefad]
[    22.457] (EE) 1: /usr/libexec/Xorg (0x557eb6744000+0x1bed6d) [0x557eb6902d6d]
[    22.457] (EE) 2: /lib64/libpthread.so.0 (0x7ff2ee393000+0x12d20) [0x7ff2ee3a5d20]
[    22.457] (EE) 3: /opt/amdgpu/lib64/dri/amdgpu_dri.so (0x7ff2e4fb8000+0x11e049d) [0x7ff2e619849d]
[    22.457] (EE) 4: /opt/amdgpu/lib64/dri/amdgpu_dri.so (eglInitialize+0x8c) [0x7ff2e619284c]
[    22.457] (EE) 5: /usr/lib64/xorg/modules/libglamoregl.so (glamor_egl_init+0xa3) [0x7ff2de023193]
[    22.457] (EE) 6: /opt/amdgpu/lib/xorg/modules/drivers/amdgpu_drv.so (0x7ff2f0a2d000+0x20889) [0x7ff2f0a4d889]
[    22.457] (EE) 7: /opt/amdgpu/lib/xorg/modules/drivers/amdgpu_drv.so (0x7ff2f0a2d000+0x12566) [0x7ff2f0a3f566]
[    22.457] (EE) 8: /opt/amdgpu/lib/xorg/modules/drivers/amdgpu_drv.so (0x7ff2f0a2d000+0x12e1b) [0x7ff2f0a3fe1b]
[    22.457] (EE) 9: /usr/libexec/Xorg (InitOutput+0x9ee) [0x557eb67e1e6e]
[    22.457] (EE) 10: /usr/libexec/Xorg (0x557eb6744000+0x60607) [0x557eb67a4607]
[    22.457] (EE) 11: /lib64/libc.so.6 (__libc_start_main+0xe5) [0x7ff2edff77e5]
[    22.457] (EE) 12: /usr/libexec/Xorg (_start+0x2e) [0x557eb678e03e]
[    22.457] (EE) 
[    22.457] (EE) Segmentation fault at address 0x18
[    22.457] (EE) 
Fatal server error:
[    22.457] (EE) Caught signal 11 (Segmentation fault). Server aborting
[    22.457] (EE) 
[    22.457] (EE) 
Please consult the The X.Org Foundation support 
	 at http://wiki.x.org
 for help. 
[    22.457] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[    22.457] (EE) 
[    22.459] (EE) Server terminated with error (1). Closing log file.

[mdclemen]

With the amdgpu-install scripts that are provided at repo.radeon.com, the --usecase=workstation option isn't the way to go, as this installs older, proprietary OpenGL libs, --usecase=graphics, is the better choice. After doing that, I was able to get some of the amdgpu drivers from repo.radeon.com to at least give graphics support, but still no OpenCL support.

The most recent amdgpu driver that works for me, before OpenCL breaks, is this one, which is .10 greater (21.20 vs 21.10) than the most recent one that linked on AMD own supported driver page for the W5100. My guess is sometime after the release of 21.10, AMD dropped OpenCL support for the BONAIRE chipset and GCN2.0 architecture GPUs.

The FirePro W5100 was released in 2014, and driver 21.20 was released in 2021, so by my calculations, in SEVEN years AMD couldn't even be bothered to give OpenCL 2.0 functionality on linux to a GPU, which they advertise as OpenCL 2.0 capable (without any caveats as to OS). This archived AMD page for the W5100 from 2014 says  "AMD FirePro™ W5100 supports OpenCL 1.2 and is expected to support OpenCL 2.0 (2)", where the footnote states: "AMD plans to release OpenCL 2.0 drivers for enabled AMD FirePro W5100, W7100, W8100 and W9100 graphics cards in Q4 2014", and this final archived page from 2018 is still promising OpenCL 2.0.

What kind of shady business practice is this? Do better AMD.

I'm giving up on this and going to the NVIDIA darkside.