Clock speed is too low for Vega56

Sneaky66 · 2022-10-08 12:19:40

Hello All!

I tried with 5.15 and 5.19 kernels, but it seems that my Vega56 can't reach it's full potential. It can't switch up to higher clock speeds.
All I can see is this 991Mhz is the max. On Windows it's usually uses the maximum ~1500Mhz while gaming. I use the Amdgpu kernel module + MESA.

@archlinux ~]$ uname -a
Linux archlinux 5.19.13-arch1-1 #1 SMP PREEMPT_DYNAMIC Tue, 04 Oct 2022 14:36:58 +0000 x86_64 GNU/Linux

cxcx@archlinux ~]$ cat /sys/class/drm/card0/device/pp_dpm_sclk
0: 852Mhz
1: 991Mhz *
2: 1138Mhz
3: 1269Mhz
4: 1312Mhz
5: 1474Mhz
6: 1538Mhz
7: 1590Mhz

Is it a common problem with these newer kernels?

Lone_Wolf · 2022-10-08 15:05:53

Normally amdgpu kernel module dynamically adjusts clockspeeds based on demand.

please post the output of

$ systool -v -m amdgpu
$ glxinfo -B #comes with mesa-utils

Incase you're running wayland, post the first 100 lines of the output of eglinfo (also part of mesa-utils) .

Welcome to archlinux forums.

Sneaky66 · 2022-10-09 10:04:32

hlinux ~]$ systool -v -m amdgpu
Module = "amdgpu"

Attributes:
coresize = "9379840"
initsize = "0"
initstate = "live"
refcnt = "6"
srcversion = "04098E6EC763122701D8D10"
taint = ""
uevent = <store method only>

Parameters:
abmlevel = "0"
aspm = "0"
async_gfx_ring = "1"
audio = "-1"
backlight = "-1"
bad_page_threshold = "-1"
bapm = "-1"
cg_mask = "18446744073709551615"
cik_support = "0"
compute_multipipe = "-1"
cwsr_enable = "1"
dc = "-1"
dcdebugmask = "0"
dcfeaturemask = "2"
debug_evictions = "N"
debug_largebar = "0"
deep_color = "0"
disable_cu = "(null)"
discovery = "-1"
disp_priority = "0"
dpm = "-1"
emu_mode = "0"
exp_hw_support = "0"
force_asic_type = "-1"
forcelongtraining = "0"
fw_load_type = "-1"
gpu_recovery = "-1"
halt_if_hws_hang = "0"
hw_i2c = "0"
hws_gws_support = "N"
hws_max_conc_proc = "-1"
ignore_crat = "0"
ip_block_mask = "4294967295"
job_hang_limit = "0"
lbpw = "-1"
lockup_timeout = ""
max_num_of_queues_per_device= "4096"
mcbp = "0"
mes_kiq = "0"
mes = "0"
msi = "-1"
no_queue_eviction_on_vm_fault= "0"
no_system_mem_limit = "N"
noretry = "-1"
num_kcq = "-1"
pcie_gen2 = "-1"
pcie_gen_cap = "0"
pcie_lane_cap = "0"
pg_mask = "4294967295"
ppfeaturemask = "0xfff7ffff"
queue_preemption_timeout_ms= "9000"
ras_enable = "-1"
ras_mask = "4294967295"
reset_method = "-1"
runpm = "-1"
sched_hw_submission = "2"
sched_jobs = "32"
sched_policy = "0"
sdma_phase_quantum = "32"
send_sigterm = "0"
si_support = "0"
smu_memory_pool_size= "0"
smu_pptable_id = "-1"
timeout_fatal_disable= "N"
timeout_period = "0"
tmz = "-1"
use_xgmi_p2p = "1"
vcnfw_log = "0"
virtual_display = "(null)"
vis_vramlimit = "0"
vm_block_size = "-1"
vm_debug = "0"
vm_fault_stop = "0"
vm_fragment_size = "-1"
vm_size = "-1"
vm_update_mode = "-1"

Sections:

--------------

chlinux ~]$ glxinfo -B
name of display: :0
display: :0 screen: 0
direct rendering: Yes
Extended renderer info (GLX_MESA_query_renderer):
Vendor: AMD (0x1002)
Device: AMD Radeon RX Vega (vega10, LLVM 14.0.6, DRM 3.47, 5.19.13-arch1-1) (0x687f)
Version: 22.1.7
Accelerated: yes
Video memory: 8192MB
Unified memory: no
Preferred profile: core (0x1)
Max core profile version: 4.6
Max compat profile version: 4.6
Max GLES1 profile version: 1.1
Max GLES[23] profile version: 3.2
Memory info (GL_ATI_meminfo):
VBO free memory - total: 7793 MB, largest block: 7793 MB
VBO free aux. memory - total: 7897 MB, largest block: 7897 MB
Texture free memory - total: 7793 MB, largest block: 7793 MB
Texture free aux. memory - total: 7897 MB, largest block: 7897 MB
Renderbuffer free memory - total: 7793 MB, largest block: 7793 MB
Renderbuffer free aux. memory - total: 7897 MB, largest block: 7897 MB
Memory info (GL_NVX_gpu_memory_info):
Dedicated video memory: 8192 MB
Total available memory: 16146 MB
Currently available dedicated video memory: 7793 MB
OpenGL vendor string: AMD
OpenGL renderer string: AMD Radeon RX Vega (vega10, LLVM 14.0.6, DRM 3.47, 5.19.13-arch1-1)
OpenGL core profile version string: 4.6 (Core Profile) Mesa 22.1.7
OpenGL core profile shading language version string: 4.60
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile

OpenGL version string: 4.6 (Compatibility Profile) Mesa 22.1.7
OpenGL shading language version string: 4.60
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile

OpenGL ES profile version string: OpenGL ES 3.2 Mesa 22.1.7
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

-----------------------

eglinfo
EGL client extensions string:
EGL_EXT_device_base EGL_EXT_device_enumeration EGL_EXT_device_query
EGL_EXT_platform_base EGL_KHR_client_get_all_proc_addresses
EGL_EXT_client_extensions EGL_KHR_debug EGL_EXT_platform_device
EGL_EXT_platform_wayland EGL_KHR_platform_wayland
EGL_EXT_platform_x11 EGL_KHR_platform_x11 EGL_MESA_platform_xcb
EGL_MESA_platform_gbm EGL_KHR_platform_gbm
EGL_MESA_platform_surfaceless

GBM platform:
amdgpu_device_initialize: amdgpu_query_info(ACCEL_WORKING) failed (-13)
amdgpu: amdgpu_device_initialize failed.
amdgpu_device_initialize: amdgpu_query_info(ACCEL_WORKING) failed (-13)
amdgpu: amdgpu_device_initialize failed.
EGL API version: 1.4
EGL vendor string: Mesa Project
EGL version string: 1.4
EGL client APIs: OpenGL OpenGL_ES
EGL driver name: kms_swrast
EGL extensions string:
EGL_ANDROID_blob_cache EGL_EXT_buffer_age
EGL_EXT_image_dma_buf_import EGL_EXT_image_dma_buf_import_modifiers
EGL_KHR_cl_event2 EGL_KHR_config_attribs EGL_KHR_create_context
EGL_KHR_create_context_no_error EGL_KHR_fence_sync
EGL_KHR_get_all_proc_addresses EGL_KHR_gl_colorspace
EGL_KHR_gl_renderbuffer_image EGL_KHR_gl_texture_2D_image
EGL_KHR_gl_texture_3D_image EGL_KHR_gl_texture_cubemap_image
EGL_KHR_image EGL_KHR_image_base EGL_KHR_image_pixmap
EGL_KHR_no_config_context EGL_KHR_reusable_sync
EGL_KHR_surfaceless_context EGL_EXT_pixel_format_float
EGL_KHR_wait_sync EGL_MESA_configless_context
EGL_MESA_image_dma_buf_export EGL_MESA_query_driver
Configurations:
bf lv colorbuffer dp st ms vis cav bi renderable supported
id sz l r g b a th cl ns b id eat nd gl es es2 vg surfaces
---------------------------------------------------------------------
0x01 32 0 10 10 10 2 0 0 0 0 0x30335241-- y y y win
0x02 32 0 10 10 10 2 16 0 0 0 0x30335241-- y y y win
0x03 32 0 10 10 10 2 24 0 0 0 0x30335241-- y y y win
0x04 32 0 10 10 10 2 24 8 0 0 0x30335241-- y y y win
0x05 32 0 10 10 10 2 32 0 0 0 0x30335241-- y y y win
0x06 32 0 10 10 10 2 0 0 4 1 0x30335241-- y y y win
0x07 32 0 10 10 10 2 16 0 4 1 0x30335241-- y y y win
0x08 32 0 10 10 10 2 24 0 4 1 0x30335241-- y y y win
0x09 32 0 10 10 10 2 24 8 4 1 0x30335241-- y y y win
0x0a 32 0 10 10 10 2 32 0 4 1 0x30335241-- y y y win
0x0b 30 0 10 10 10 0 0 0 0 0 0x30335258-- y y y win
0x0c 30 0 10 10 10 0 16 0 0 0 0x30335258-- y y y win
0x0d 30 0 10 10 10 0 24 0 0 0 0x30335258-- y y y win
0x0e 30 0 10 10 10 0 24 8 0 0 0x30335258-- y y y win
0x0f 30 0 10 10 10 0 32 0 0 0 0x30335258-- y y y win
0x10 30 0 10 10 10 0 0 0 4 1 0x30335258-- y y y win
0x11 30 0 10 10 10 0 16 0 4 1 0x30335258-- y y y win
0x12 30 0 10 10 10 0 24 0 4 1 0x30335258-- y y y win
0x13 30 0 10 10 10 0 24 8 4 1 0x30335258-- y y y win
0x14 30 0 10 10 10 0 32 0 4 1 0x30335258-- y y y win
0x15 32 0 8 8 8 8 0 0 0 0 0x34325241-- y y y win
0x16 32 0 8 8 8 8 16 0 0 0 0x34325241-- y y y win
0x17 32 0 8 8 8 8 24 0 0 0 0x34325241-- y y y win
0x18 32 0 8 8 8 8 24 8 0 0 0x34325241-- y y y win
0x19 32 0 8 8 8 8 32 0 0 0 0x34325241-- y y y win

Sneaky66 · 2022-10-09 10:07:55

I run gpuowl and this is the maximum:
https://paste.pics/J29QV
It should be around 180W and 1500Mhz.

Last edited by Sneaky66 (2022-10-09 10:08:49)

Lone_Wolf · 2022-10-09 11:12:39

Please use [ code ] [ /code] tags for command output, see https://bbs.archlinux.org/help.php#bbcode (or clikc the link at left bottom of this window) .

Outputs look normal.

gpuowl appears to use OpenCL which performance heavily depends on the opencl implementation used.
opencl-mesa relies on libclc which is unlikely to support vega cards. You'd need AUR opencl-amd , AUR rocm or amd proprietary drivers.
Not a good choice to benchmark / stresstest a videocard.

https://wiki.archlinux.org/title/Benchmarking#Graphics lists several options.

My personal favorites are unigine-superposition & unigine-valley .
They are old and focus on OpenGL, but look good and still manage to stress modern videocards.

Basemark fails to run OpenGL tests on my system, its vulkan tests do look good and stress the videocard.

Sneaky66 · 2022-10-09 13:38:22

Unigine-valley is running fine on my machine, but the clock is still max .960 Mhz.

Sneaky66 · 2022-10-14 18:55:34

It looks like this is only an issue with cold boot to Linux. If I restart from Windows then it's working normally. Strange.

Arch Linux

#1 2022-10-08 12:19:40

Clock speed is too low for Vega56

#2 2022-10-08 15:05:53

Re: Clock speed is too low for Vega56

#3 2022-10-09 10:04:32

Re: Clock speed is too low for Vega56

#4 2022-10-09 10:07:55

Re: Clock speed is too low for Vega56

#5 2022-10-09 11:12:39

Re: Clock speed is too low for Vega56

#6 2022-10-09 13:38:22

Re: Clock speed is too low for Vega56

#7 2022-10-14 18:55:34

Re: Clock speed is too low for Vega56

Board footer