08:18 <heeen> Available names: fps frametime cpu cpu0 cpu1 cpu2 cpu3 samples-passed prims-generated draw-calls rs-operations hi-total-cycles hi-idle-cycles hi-axi-cycles-read-request-stalled hi-axi-cycles-write-request-stalled hi-axi-cycles-write-data-stalled pe-pixel-count-killed-by-color-pipe pe-pixel-count-killed-by-depth-pipe pe-pixel-count-drawn-by-color-pipe pe-pixel-count-drawn-by-depth-pipe

08:18 <heeen> sh-shader-cycles sh-ps-inst-counter sh-rendered-pixel-counter sh-vs-inst-counter sh-rendered-vertice-counter sh-vtx-branch-inst-counter sh-vtx-texld-inst-counter sh-plx-branch-inst-counter sh-plx-texld-inst-counter pa-input-vtx-counter pa-input-prim-counter pa-output-prim-counter pa-depth-clipped-counter pa-trivial-rejected-counter pa-culled-counter se-culled-triangle-count

08:18 <heeen> se-culled-lines-count ra-valid-pixel-count ra-total-quad-count ra-valid-quad-count-after-early-z ra-total-primitive-count ra-pipe-cache-miss-counter ra-prefetch-cache-miss-counter ra-pculled-quad-count tx-total-bilinear-requests tx-total-trilinear-requests tx-total-discarded-texture-requests tx-total-texture-requests tx-mem-read-count tx-mem-read-in-8b-count tx-cache-miss-count

08:18 <heeen> tx-cache-hit-texel-count tx-cache-miss-texel-count mc-total-read-req-8b-from-pipeline mc-total-read-req-8b-from-ip mc-total-write-req-8b-from-pipeline

08:22 <lynxeye> heeen: https://lists.freedesktop.org/archives/etnaviv/2022-September/003887.html

08:23 <lynxeye> Use together with: https://cgit.freedesktop.org/~tursulin/intel-gpu-tools/?h=gputop

09:54 <heeen> dos anyone have performance numbers of wayland on imx6 quad

10:23 <lynxeye> heeen: glmark2-es2-wayland -s 1280x800 should have a overall score of around 120-130. glmark2 overall score is not really a good benchmark, but it should give you an idea if your system performance is in the right ballpark.

10:25 <heeen> the image I got is rather limited, I do have an sdk sysroot but I fear it misses a lot of the headers and libs

10:25 <heeen> not sure if I can build mesa

10:40 <heeen> qt.scenegraph.general: R/G/B/A Buffers: 8 8 8 0

10:40 <heeen> qt.scenegraph.general: Depth Buffer: 24

10:40 <heeen> qt.scenegraph.general: Stencil Buffer: 8

10:40 <heeen> qt.scenegraph.general: Samples: 0

10:40 <heeen> qt.scenegraph.general: GL_VENDOR: etnaviv

10:40 <heeen> qt.scenegraph.general: GL_RENDERER: Vivante GC2000 rev 5108

10:40 <heeen> qt.scenegraph.general: GL_VERSION: OpenGL ES 2.0 Mesa 21.2.4.20211020-1

10:43 <heeen> hmm

10:43 <heeen> qt.scenegraph.time.renderloop: Frame rendered with 'basic' renderloop in 26ms, polish=0, sync=0, render=16, swap=0, frameDelta=33

10:43 <heeen> qt.scenegraph.time.renderer: time in renderer: total=16ms, preprocess=0, updates=0, binding=0, rendering=16

10:43 <heeen> qt.scenegraph.time.renderloop: Frame rendered with 'basic' renderloop in 26ms, polish=0, sync=0, render=16, swap=0, frameDelta=33

10:43 <heeen> qt.scenegraph.time.renderer: time in renderer: total=15ms, preprocess=0, updates=0, binding=0, rendering=15

10:48 <heeen> is there a way to disable VSYNC from an env variable

13:37 <heeen> so I can see that when I reduce the wayland window size I can reach 60fps, which means its not some wayland buffer sync contention but rather an actual fill rate limitation

14:53 <lynxeye> heeen: The problem with real Wayland composition on i.MX6Q are the number of incompatible formats. Your client application renders as multi-supertiled, for sampling in the compositor this needs to be transformed into single-tiled, which is essentially a full buffer blit. Your compositor then again renders as multi-supertiled, which needs to be transformed into linear for display scanout, again a full blit. Combined with the low availab

14:53 <lynxeye> e memory bandwidth on those chips, this is going to kill your performance.