log_interval=100

Rocket League | first minute of match replay | no additional workload | Ryzen 5700X | RTX 3060Ti | kernel 6.15rc1

Marvel Rivals Krakoa sched-ext comparison

Testing performance between no sched-ext, bpfland gaming, and lavd gaming. Launch chain used: env RADV_PERFTEST=gpl DXVK_ASYNC=1 DXVK_FRAME_RATE=0 LD_PRELOAD="" mangohud game-performance %command%

Marvel Rivals Krakoa 4090 vs 9070 XT

Different processors, 4090 setup has launch chain commands and running sched-ext bpfland gaming, 9070 XT setup is running launch chain commands and default scheduler. 4090 launch chain: env __GL_THREADED_OPTIMIZATIONS=1 __GL_YIELD="USLEEP" __GL_MaxFramesAllowed=1 DXVK_ASYNC=1 DXVK_FRAME_RATE=0 LD_PRELOAD="" 9070 XT launch chain: env RADV_PERFTEST=gpl DXVK_ASYNC=1 DXVK_FRAME_RATE=0 LD_PRELOAD="" mangohud game-performance %command%

Marvel Rivals CachyOS vs Windows 10 debloated (KRAKOA MAP)

CachyOS with launch chain commands, LACT locked clocks and raised power limits, BORE scheduler + sched-ext bpfland gaming vs Windows 10 debloated, profile inspector tweaks, with regedits tweaks, and manual GPU OC with locked clocks Launch Chain: env __GL_THREADED_OPTIMIZATIONS=1 __GL_YIELD="USLEEP" __GL_MaxFramesAllowed=1 DXVK_ASYNC=1 DXVK_FRAME_RATE=0 LD_PRELOAD="" mangohud game-performance %command%

Marvel Rivals CachyOS vs Windows 10 debloated

CachyOS with launch chain commands, LACT locked clocks and raised power limits vs Windows 10 debloated, profile inspector tweaks, with regedits tweaks, and manual GPU OC with locked clocks

Outer Wilds | BORE vs SCX

Static GPU limited scene with no background load from the start of the game; scene should be close to 100% consistent between runs; Game was run at 450p, GPU clocks were above driver limits but wattage wasn't fully maxed out. 60 second frame recording with 0ms interval to get the most accurate results; single run only | Wayland Native + NTSync. Performance Governor/EPP | BORE default settings | BPFLAND --performance | FLASH | LAVD --performance | LAVD --autopilot

EEVDF SMT FIX in Rocket League | kernel 6.15rc1 #2

Rocket League | first two minutes of match replay | no additional workload | Ryzen 5700X | RTX 3060Ti | power profile performance

EEVDF SMT FIX in Rocket League | kernel 6.15rc1

Rocket League | first two minutes of match replay | no additional workload | Ryzen 5700X | RTX 3060Ti | power profile balanced

Elden Ring | BORE vs SCX

Static fairly GPU limited scene (not fully saturated) with some background load composed of a Discord call + 1080p60fps CPU encoded streaming. 30 second frame recording with 0ms interval to get the most accurate results; single run only. Performance Governor/EPP | BORE default settings | BPFLAND --performance | FLASH | LAVD --performance | LAVD --autopilot Since the scene isn't fully stable I retested BORE at the end; order: BORE, BPFLAND, FLASH, LAVD --performance, LAVD --autopilot, BORE.

Bpfland in Rocket League | kernel 6.15rc1

Rocket League | first two minutes of match replay | no additional workload | Ryzen 5700X | RTX 3060Ti | power profile balanced