<div dir="ltr"><div>I'm pretty sure I merged this series in February.</div><div><br></div><div>Marek<br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Apr 8, 2019 at 6:10 PM Dieter Nützel <<a href="mailto:Dieter@nuetzel-hh.de">Dieter@nuetzel-hh.de</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Maybe someone working on this, too.<br>
<br>
I'm feeling fine again after a short 'trip' into the hospital...;-)<br>
<br>
Dieter<br>
<br>
Am 26.02.2019 07:36, schrieb Marek Olšák:<br>
> We need to extend the CS ioctl to allow submitting 2 command buffers<br>
> at the same time.<br>
> <br>
> Marek<br>
> <br>
> On Mon, Feb 25, 2019, 10:06 PM Dieter Nützel <<a href="mailto:Dieter@nuetzel-hh.de" target="_blank">Dieter@nuetzel-hh.de</a>><br>
> wrote:<br>
> <br>
>> Hello Marek,<br>
>> <br>
>> you wrote with your series sent:<br>
>> <br>
>> [-]<br>
>> Trivial benchmarks such as glxgears can expect 20% decrease<br>
>> in performance due to the added cost of the SDMA CS ioctl that<br>
>> wasn't<br>
>> there before.<br>
>> [-]<br>
>> <br>
>> Any ideas to speed this up, again?<br>
>> glmark2 went from 9766 (best ever) down to 7455 (all with NIR).<br>
>> Or are micro benchmarks not worth more effort?<br>
>> <br>
>> Dieter<br>
>> <br>
>> SDMA<br>
>> =======================================================<br>
>> glmark2 2017.07<br>
>> =======================================================<br>
>> OpenGL Information<br>
>> GL_VENDOR: X.Org<br>
>> GL_RENDERER: Radeon RX 580 Series (POLARIS10, DRM 3.30.0,<br>
>> 5.0.0-rc1-1.g7262353-default+, LLVM 9.0.0)<br>
>> GL_VERSION: 4.5 (Compatibility Profile) Mesa 19.1.0-devel<br>
>> (git-a9b32aaa16)<br>
>> =======================================================<br>
>> [build] use-vbo=false: FPS: 3694 FrameTime: 0.271 ms<br>
>> [build] use-vbo=true: FPS: 9341 FrameTime: 0.107 ms<br>
>> [texture] texture-filter=nearest: FPS: 9140 FrameTime: 0.109 ms<br>
>> [texture] texture-filter=linear: FPS: 9163 FrameTime: 0.109 ms<br>
>> [texture] texture-filter=mipmap: FPS: 9161 FrameTime: 0.109 ms<br>
>> [shading] shading=gouraud: FPS: 9234 FrameTime: 0.108 ms<br>
>> [shading] shading=blinn-phong-inf: FPS: 9255 FrameTime: 0.108 ms<br>
>> [shading] shading=phong: FPS: 9226 FrameTime: 0.108 ms<br>
>> [shading] shading=cel: FPS: 9310 FrameTime: 0.107 ms<br>
>> [bump] bump-render=high-poly: FPS: 9298 FrameTime: 0.108 ms<br>
>> [bump] bump-render=normals: FPS: 9121 FrameTime: 0.110 ms<br>
>> [bump] bump-render=height: FPS: 9120 FrameTime: 0.110 ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [effect2d] kernel=0,1,0;1,-4,1;0,1,0;: FPS: 9858 FrameTime: 0.101 ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [effect2d] kernel=1,1,1,1,1;1,1,1,1,1;1,1,1,1,1;: FPS: 9854<br>
>> FrameTime:<br>
>> 0.101 ms<br>
>> [pulsar] light=false:quads=5:texture=false: FPS: 8468 FrameTime:<br>
>> 0.118<br>
>> ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [desktop]<br>
>> blur-radius=5:effect=blur:passes=1:separable=true:windows=4:<br>
>> FPS: 5181 FrameTime: 0.193 ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [desktop] effect=shadow:windows=4: FPS: 5374 FrameTime: 0.186 ms<br>
>> [buffer]<br>
>> <br>
> columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=map:<br>
>> <br>
>> FPS: 824 FrameTime: 1.214 ms<br>
>> [buffer]<br>
>> <br>
> columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata:<br>
>> <br>
>> FPS: 1114 FrameTime: 0.898 ms<br>
>> [buffer]<br>
>> <br>
> columns=200:interleave=true:update-dispersion=0.9:update-fraction=0.5:update-method=map:<br>
>> <br>
>> FPS: 899 FrameTime: 1.112 ms<br>
>> [ideas] speed=duration: FPS: 3485 FrameTime: 0.287 ms<br>
>> [jellyfish] <default>: FPS: 7992 FrameTime: 0.125 ms<br>
>> [terrain] <default>: FPS: 1796 FrameTime: 0.557 ms<br>
>> [shadow] <default>: FPS: 7350 FrameTime: 0.136 ms<br>
>> [refract] <default>: FPS: 3595 FrameTime: 0.278 ms<br>
>> [conditionals] fragment-steps=0:vertex-steps=0: FPS: 9401 FrameTime:<br>
>> <br>
>> 0.106 ms<br>
>> [conditionals] fragment-steps=5:vertex-steps=0: FPS: 9413 FrameTime:<br>
>> <br>
>> 0.106 ms<br>
>> [conditionals] fragment-steps=0:vertex-steps=5: FPS: 9417 FrameTime:<br>
>> <br>
>> 0.106 ms<br>
>> [function] fragment-complexity=low:fragment-steps=5: FPS: 9365<br>
>> FrameTime: 0.107 ms<br>
>> [function] fragment-complexity=medium:fragment-steps=5: FPS: 9451<br>
>> FrameTime: 0.106 ms<br>
>> [loop] fragment-loop=false:fragment-steps=5:vertex-steps=5: FPS:<br>
>> 9300<br>
>> FrameTime: 0.108 ms<br>
>> [loop] fragment-steps=5:fragment-uniform=false:vertex-steps=5: FPS:<br>
>> 9440<br>
>> FrameTime: 0.106 ms<br>
>> [loop] fragment-steps=5:fragment-uniform=true:vertex-steps=5: FPS:<br>
>> 9392<br>
>> FrameTime: 0.106 ms<br>
>> =======================================================<br>
>> glmark2 Score: 7455<br>
>> =======================================================<br>
>> <br>
>> Before<br>
>> =======================================================<br>
>> glmark2 2017.07<br>
>> =======================================================<br>
>> OpenGL Information<br>
>> GL_VENDOR: X.Org<br>
>> GL_RENDERER: Radeon RX 580 Series (POLARIS10, DRM 3.27.0,<br>
>> 4.20.0-rc3-1.g7262353-default+, LLVM 8.0.0)<br>
>> GL_VERSION: 4.5 (Compatibility Profile) Mesa 19.0.0-devel<br>
>> (git-c49b3df3cb)<br>
>> =======================================================<br>
>> [build] use-vbo=false: FPS: 3373 FrameTime: 0.296 ms<br>
>> [build] use-vbo=true: FPS: 13121 FrameTime: 0.076 ms<br>
>> [texture] texture-filter=nearest: FPS: 12172 FrameTime: 0.082 ms<br>
>> [texture] texture-filter=linear: FPS: 12557 FrameTime: 0.080 ms<br>
>> [texture] texture-filter=mipmap: FPS: 12228 FrameTime: 0.082 ms<br>
>> [shading] shading=gouraud: FPS: 12536 FrameTime: 0.080 ms<br>
>> [shading] shading=blinn-phong-inf: FPS: 12782 FrameTime: 0.078 ms<br>
>> [shading] shading=phong: FPS: 12619 FrameTime: 0.079 ms<br>
>> [shading] shading=cel: FPS: 12735 FrameTime: 0.079 ms<br>
>> [bump] bump-render=high-poly: FPS: 11412 FrameTime: 0.088 ms<br>
>> [bump] bump-render=normals: FPS: 12467 FrameTime: 0.080 ms<br>
>> [bump] bump-render=height: FPS: 12422 FrameTime: 0.081 ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [effect2d] kernel=0,1,0;1,-4,1;0,1,0;: FPS: 13252 FrameTime: 0.075<br>
>> ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [effect2d] kernel=1,1,1,1,1;1,1,1,1,1;1,1,1,1,1;: FPS: 11468<br>
>> FrameTime:<br>
>> 0.087 ms<br>
>> [pulsar] light=false:quads=5:texture=false: FPS: 11385 FrameTime:<br>
>> 0.088<br>
>> ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [desktop]<br>
>> blur-radius=5:effect=blur:passes=1:separable=true:windows=4:<br>
>> FPS: 5609 FrameTime: 0.178 ms<br>
>> libpng warning: iCCP: known incorrect sRGB profile<br>
>> [desktop] effect=shadow:windows=4: FPS: 5944 FrameTime: 0.168 ms<br>
>> [buffer]<br>
>> <br>
> columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=map:<br>
>> <br>
>> FPS: 837 FrameTime: 1.195 ms<br>
>> [buffer]<br>
>> <br>
> columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata:<br>
>> <br>
>> FPS: 1153 FrameTime: 0.867 ms<br>
>> [buffer]<br>
>> <br>
> columns=200:interleave=true:update-dispersion=0.9:update-fraction=0.5:update-method=map:<br>
>> <br>
>> FPS: 925 FrameTime: 1.081 ms<br>
>> [ideas] speed=duration: FPS: 3042 FrameTime: 0.329 ms<br>
>> [jellyfish] <default>: FPS: 10209 FrameTime: 0.098 ms<br>
>> [terrain] <default>: FPS: 1791 FrameTime: 0.558 ms<br>
>> [shadow] <default>: FPS: 9248 FrameTime: 0.108 ms<br>
>> [refract] <default>: FPS: 3502 FrameTime: 0.286 ms<br>
>> [conditionals] fragment-steps=0:vertex-steps=0: FPS: 13061<br>
>> FrameTime:<br>
>> 0.077 ms<br>
>> [conditionals] fragment-steps=5:vertex-steps=0: FPS: 13225<br>
>> FrameTime:<br>
>> 0.076 ms<br>
>> [conditionals] fragment-steps=0:vertex-steps=5: FPS: 12882<br>
>> FrameTime:<br>
>> 0.078 ms<br>
>> [function] fragment-complexity=low:fragment-steps=5: FPS: 12806<br>
>> FrameTime: 0.078 ms<br>
>> [function] fragment-complexity=medium:fragment-steps=5: FPS: 12936<br>
>> FrameTime: 0.077 ms<br>
>> [loop] fragment-loop=false:fragment-steps=5:vertex-steps=5: FPS:<br>
>> 12954<br>
>> FrameTime: 0.077 ms<br>
>> [loop] fragment-steps=5:fragment-uniform=false:vertex-steps=5: FPS:<br>
>> 12788 FrameTime: 0.078 ms<br>
>> [loop] fragment-steps=5:fragment-uniform=true:vertex-steps=5: FPS:<br>
>> 12846<br>
>> FrameTime: 0.078 ms<br>
>> =======================================================<br>
>> glmark2 Score: 9766<br>
>> =======================================================<br>
</blockquote></div></div>