Score:0

Poor video performance after amdgpu upgrade

ar flag

I have a Radeon RX 7900 XTX on Ubuntu 22.04. It previously was running fine after installing amdgpu-install_5.4.50406-1_all.deb, before I decided to update to the latest amdgpu-install_5.5.50502-1_all.deb (July 2023); now I get poor framerates in games.

I used sudo amdgpu-install --usecase=graphics,workstation,amf --vulkan=amdvlk,pro to do the install. Prior to the install, my glxinfo -B looked like this:

name of display: :0
display: :0  screen: 0
direct rendering: Yes
Extended renderer info (GLX_MESA_query_renderer):
    Vendor: AMD (0x1002)
    Device: Radeon RX 7900 XTX (gfx1100, LLVM 15.0.3, DRM 3.48, 5.19.0-46-generic) (0x744c)
    Version: 22.3.0
    Accelerated: yes
    Video memory: 24576MB
    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: 21960 MB, largest block: 21960 MB
    VBO free aux. memory - total: 63924 MB, largest block: 63924 MB
    Texture free memory - total: 21960 MB, largest block: 21960 MB
    Texture free aux. memory - total: 63924 MB, largest block: 63924 MB
    Renderbuffer free memory - total: 21960 MB, largest block: 21960 MB
    Renderbuffer free aux. memory - total: 63924 MB, largest block: 63924 MB
Memory info (GL_NVX_gpu_memory_info):
    Dedicated video memory: 24576 MB
    Total available memory: 88548 MB
    Currently available dedicated video memory: 21960 MB
OpenGL vendor string: AMD
OpenGL renderer string: Radeon RX 7900 XTX (gfx1100, LLVM 15.0.3, DRM 3.48, 5.19.0-46-generic)
OpenGL core profile version string: 4.6 (Core Profile) Mesa 22.3.0-devel
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 22.3.0-devel
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 22.3.0-devel
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

Afterwards it looks like this:

name of display: :0
display: :0  screen: 0
direct rendering: Yes
Memory info (GL_ATI_meminfo):
    VBO free memory - total: 24559 MB, largest block: 24559 MB
    VBO free aux. memory - total: 63972 MB, largest block: 63972 MB
    Texture free memory - total: 24559 MB, largest block: 24559 MB
    Texture free aux. memory - total: 63972 MB, largest block: 63972 MB
    Renderbuffer free memory - total: 24559 MB, largest block: 24559 MB
    Renderbuffer free aux. memory - total: 63972 MB, largest block: 63972 MB
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: Radeon RX 7900 XTX
OpenGL core profile version string: 4.6.0 Core Profile Context 23.10.230614
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: 4.6.0 Compatibility Profile Context 23.10.230614
OpenGL shading language version string: 4.60
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile

OpenGL ES profile version string: OpenGL ES 3.2.0 23.10.230614 
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20

Seems suspicious that the Extended renderer info (GLX_MESA_query_renderer) has disappeared from glxinfo -B?

I've tried:

  • Re-installing kisak Mesa drivers
  • Reverting to the older amdgpu package: performance is back, Extended renderer info reappears.
I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.