Skip to content

Instantly share code, notes, and snippets.

View jrmuizel's full-sized avatar

Jeff Muizelaar jrmuizel

View GitHub Profile
@jrmuizel
jrmuizel / gist:fff750455ba89f4fa2128f1a56cee234
Created February 3, 2021 01:32
brush_mix_blend.glsl gen9 mac output (unoptimized)
kernel
BB0 Preds: Succs: BB5 BB140
main:
pln (8) r123.0<1>:f r5.0<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//1 //$1:&1:%18446744073709551615
pln (8) r122.0<1>:f r5.4<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//2 //$2:&2:%18446744073709551615
sends (8) r14:f r123 r122 0x42:ud 0x2420008:ud{Q1, Align1} //$5:&3:%18446744073709551615 // sampler, resLen=4, msgLen=1, extMsgLen=1
pln (8) r121.0<1>:f r13.0<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//6 //$6:&4:%18446744073709551615
pln (8) r120.0<1>:f r13.4<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//7 //$7:&5:%18446744073709551615
sends (8) r110:f r121 r120 0x42:ud 0x2420109:ud{Q1, Align1} //$10:&6:%18446744073709551615 // sampler, resLen=4, msgLen=1, extMsgLen=1
@jrmuizel
jrmuizel / gist:0886486931901015ba0a3b4316cfb2bc
Created February 3, 2021 01:21
brush_mix_blend.glsl gen9 mac output
kernel
BB0 Preds: Succs: BB5 BB6
main:
pln (8) r123.0<1>:f r5.0<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//1 //$1:&1:%18446744073709551615
pln (8) r122.0<1>:f r5.4<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//2 //$2:&2:%18446744073709551615
sends (8) r14:f r123 r122 0x42:ud 0x2420008:ud{Q1, Align1} //$5:&3:%18446744073709551615 // sampler, resLen=4, msgLen=1, extMsgLen=1
pln (8) r121.0<1>:f r13.0<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//6 //$6:&4:%18446744073709551615
pln (8) r120.0<1>:f r13.4<0;1,0>:f r2.0<8;8,1>:f {Q1, Align1, NoMask}//7 //$7:&5:%18446744073709551615
sends (8) r110:f r121 r120 0x42:ud 0x2420109:ud{Q1, Align1} //$10:&6:%18446744073709551615 // sampler, resLen=4, msgLen=1, extMsgLen=1
kernel
BB0 Preds: Succs:
main:
mul (4) r22.0<1>:f r2.0<4;4,1>:f r4.0<0;1,0>:f {N1, Align1, NoMask}//2 //$2:&2:%18446744073709551615
mul (4) r97.0<1>:f r2.0<4;4,1>:f r4.4<0;1,0>:f {N1, Align1, NoMask}//18 //$18:&15:%18446744073709551615
mad (1) r114.0<1>:f r22.0<0;1,0>:f r2.4<0;1,0>:f r4.1<0;1,0>:f {Q1, Align1, NoMask}//6 //$6:&3:%18446744073709551615 {0=EL, 1=EL, 2=EL, BC=BAD}
mad (1) r11.0<1>:f r22.1<0;1,0>:f r2.5<0;1,0>:f r4.1<0;1,0>:f {Q1, Align1, NoMask}//7 //$7:&4:%18446744073709551615 {0=EL, 1=EL, 2=EL, BC=BAD}
mad (1) r114.4<1>:f r22.2<0;1,0>:f r2.6<0;1,0>:f r4.1<0;1,0>:f {Q1, Align1, NoMask}//8 //$8:&5:%18446744073709551615 {0=EL, 1=EL, 2=EL, BC=BAD}
mad (1) r11.4<1>:f r22.3<0;1,0>:f r2.7<0;1,0>:f r4.1<0;1,0>:f {Q1, Align1, NoMask}//9 //$9:&6:%18446744073709551615 {0=EL, 1=EL, 2=EL, BC=BAD}
kernel
BB0 Preds: Succs:
main:
mul (4) r21.0<1>:f r2.4<4;4,1>:f r4.4<0;1,0>:f {N1, Align1, NoMask}//1 //$1:&1:%18446744073709551615
mul (4) r97.0<1>:f r2.4<4;4,1>:f r5.0<0;1,0>:f {N1, Align1, NoMask}//17 //$17:&14:%18446744073709551615
mad (1) r0.0<1>:f r21.0<0;1,0>:f r3.0<0;1,0>:f r4.5<0;1,0>:f {Q1, Align1, NoMask}//5 //$5:&2:%18446744073709551615 {0=OL, 1=OL, 2=EL, BC=GOOD}
mad (1) r114.0<1>:f r21.1<0;1,0>:f r3.1<0;1,0>:f r4.5<0;1,0>:f {Q1, Align1, NoMask}//6 //$6:&3:%18446744073709551615 {0=OL, 1=OL, 2=EL, BC=GOOD}
mad (1) r0.4<1>:f r21.2<0;1,0>:f r3.2<0;1,0>:f r4.5<0;1,0>:f {Q1, Align1, NoMask}//7 //$7:&4:%18446744073709551615 {0=OL, 1=OL, 2=EL, BC=GOOD}
mad (1) r114.4<1>:f r21.3<0;1,0>:f r3.3<0;1,0>:f r4.5<0;1,0>:f {Q1, Align1, NoMask}//8 //$8:&5:%18446744073709551615 {0=OL, 1=OL, 2=EL, BC=GOOD}
@jrmuizel
jrmuizel / copy-region-scatter.cpp
Created January 28, 2021 16:33
copy region scatter bench
#pragma comment(lib, "d3d11.lib")
#include <d3d11.h>
#include <stdio.h>
#include <windows.h>
#include <assert.h>
#define WIDTH 8192
#define HEIGHT 8192
int buf[WIDTH * HEIGHT];
int buf2[WIDTH * HEIGHT];
library "/Users/jrmuizel/src/backtrace-rs/target/debug/deps/macos-6e484d5a3cd96e2e"
segment 0 100000000
segment 100000000 164000
segment 100164000 c000
segment 100170000 4000
segment 100174000 1bc000
bias ea04000
library "/System/Library/Frameworks/CoreGraphics.framework/Versions/A/CoreGraphics"
segment 7fff2d54c000 676000
segment 7fff84affb40 4a0f8
---- macos stdout ----
library "/Users/jrmuizel/src/backtrace-rs/target/debug/deps/macos-6e484d5a3cd96e2e"
segment 0 100000000
segment 100000000 164000
segment 100164000 c000
segment 100170000 4000
segment 100174000 1bc000
bias 266e000
library "/System/Library/Frameworks/CoreGraphics.framework/Versions/A/CoreGraphics"
segment 7fff2d54c000 676000
@jrmuizel
jrmuizel / 2017PopularCrates.md
Created April 21, 2017 20:08
2017 Popular crates
@jrmuizel
jrmuizel / PopularCrates.md
Created April 21, 2017 19:59
Popular crates
diff --git a/image/decoders/nsJPEGDecoder.cpp b/image/decoders/nsJPEGDecoder.cpp
index 30590f0..6bccb1a 100644
--- a/image/decoders/nsJPEGDecoder.cpp
+++ b/image/decoders/nsJPEGDecoder.cpp
@@ -422,6 +422,8 @@ nsJPEGDecoder::WriteInternal(const char *aBuffer, PRUint32 aCount)
*/
mInfo.dct_method = JDCT_ISLOW;
mInfo.dither_mode = JDITHER_FS;
+ mInfo.scale_denom = 4;
+ mInfo.scale_num = 1;