Skip to content

Performance lower than amdvlk in CS2

Description

I used mangohud to check the performance in the maps: ancient, vertigo, and inferno. In practice mode without bots and with and withouh smoke grenades being active, the without being taken after the grenade dissipated.

A table with values as it's easier to read than screenshots:

stage radv amdvlk
ancient/smoke 128/117 146/133
vertigo/smoke 210/182 239/210
inferno/smoke 202/165 237/186

#Screenshots

Counter-Strike_2_20240115_184802

Counter-Strike_2_20240115_184813

Counter-Strike_2_20240115_184911

Counter-Strike_2_20240115_184921

Counter-Strike_2_20240115_185046

Counter-Strike_2_20240115_185056

Counter-Strike_2_20240115_185326

Counter-Strike_2_20240115_185335

Counter-Strike_2_20240115_185558

Counter-Strike_2_20240115_185608

Counter-Strike_2_20240115_185704

Counter-Strike_2_20240115_185714

Steps to reproduce

Just stand in position and compare the performance

System information

System:
  Host: DeepBlue Kernel: 6.7.0-arch3-1 arch: x86_64 bits: 64 compiler: gcc
    v: 13.2.1 Desktop: KDE Plasma v: 5.27.10 tk: Qt v: 5.15.12 wm: kwin_x11 dm:
    1: LightDM note: stopped 2: SDDM Distro: Arch Linux
CPU:
  Info: 6-core model: AMD Ryzen 5 5600G with Radeon Graphics bits: 64
    type: MT MCP arch: Zen 3 rev: 0 cache: L1: 384 KiB L2: 3 MiB L3: 16 MiB
  Speed (MHz): avg: 3436 high: 4443 min/max: 400/4464 cores: 1: 3551 2: 4442
    3: 3551 4: 3552 5: 3543 6: 3543 7: 3553 8: 4443 9: 3553 10: 3552 11: 3552
    12: 400 bogomips: 93456
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm
Graphics:
  Device-1: AMD Navi 23 [Radeon RX 6600/6600 XT/6600M] vendor: ASRock
    driver: amdgpu v: kernel arch: RDNA-2 pcie: speed: 16 GT/s lanes: 16 ports:
    active: HDMI-A-1 empty: DP-1,DP-2,DP-3 bus-ID: 03:00.0 chip-ID: 1002:73ff
  Display: x11 server: X.Org v: 21.1.10 with: Xwayland v: 23.2.3
    compositor: kwin_x11 driver: X: loaded: amdgpu
    unloaded: fbdev,modesetting,vesa dri: radeonsi gpu: amdgpu display-ID: :0
    screens: 1
  Screen-1: 0 s-res: 1680x1050 s-dpi: 96
  Monitor-1: HDMI-A-1 mapped: HDMI-A-0 model: LG (GoldStar) TV
    res: 1680x1050 dpi: 90 diag: 558mm (22")
  API: EGL v: 1.5 platforms: device: 0 drv: radeonsi device: 1 drv: swrast
    surfaceless: drv: radeonsi x11: drv: radeonsi inactive: gbm,wayland
  API: OpenGL v: 4.6 compat-v: 4.5 vendor: amd v: N/A glx-v: 1.4
    direct-render: yes renderer: AMD Radeon RX 6600 (radeonsi navi23 LLVM
    16.0.6 DRM 3.56 6.7.0-arch3-1) device-ID: 1002:73ff
  API: Vulkan v: 1.3.274 surfaces: xcb,xlib device: 0 type: discrete-gpu
    driver: mesa radv device-ID: 1002:73ff

If applicable

  • Xserver version: 1.21.1.10

Regression

Tried with 23.3.3 and I see equivalent performance to that of cfba24cc

Further information (optional)

RADV_DEBUG=llvm has slightly lower performance Despite the results here, it's still better to play with radv because shader compilation causes amdvlk to stutter, which doesn't happen with radv/aco.

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information