skia2/resources/SkVMTest.expected
Mike Klein 6b43c30cd7 add mad_unorm8 instruction
At some point adding more and more complex instructions reduces
to the absurdity of SolveTheWholeProblem-The-Instruction, but
I think this one will come up often enough to still make sense.
mad() makes sense for unorm8 just about everywhere mad() makes
sense for f32.

This instruction won't matter to a JIT, but helps the interpreter.

Change-Id: Iace92296cffbb6fbc3acd1f853cb01c51792f796
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/218716
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
2019-06-05 20:46:02 +00:00

324 lines
6.2 KiB
Plaintext

A8 over A8
7 registers, 15 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = splat 437F0000 (255)
r3 = splat 3F000000 (0.5)
loop:
r4 = load8 arg(0)
r4 = to_f32 r4
r4 = mul_f32 r0 r4
r5 = load8 arg(1)
r5 = to_f32 r5
r5 = mul_f32 r0 r5
r6 = sub_f32 r1 r4
r6 = mad_f32 r5 r6 r4
r6 = mad_f32 r6 r2 r3
r6 = to_i32 r6
store8 arg(1) r6
A8 over G8
9 registers, 21 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = splat 3E59B3D0 (0.21259999)
r3 = splat 3F371759 (0.71520001)
r4 = splat 3D93DD98 (0.0722)
r5 = splat 437F0000 (255)
r6 = splat 3F000000 (0.5)
loop:
r7 = load8 arg(0)
r7 = to_f32 r7
r7 = mul_f32 r0 r7
r8 = load8 arg(1)
r8 = to_f32 r8
r8 = mul_f32 r0 r8
r7 = sub_f32 r1 r7
r7 = mul_f32 r8 r7
r8 = mul_f32 r7 r4
r8 = mad_f32 r7 r3 r8
r8 = mad_f32 r7 r2 r8
r8 = mad_f32 r8 r5 r6
r8 = to_i32 r8
store8 arg(1) r8
A8 over RGBA_8888
10 registers, 37 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = splat 437F0000 (255)
r3 = splat 3F000000 (0.5)
loop:
r4 = load8 arg(0)
r4 = to_f32 r4
r4 = mul_f32 r0 r4
r5 = load32 arg(1)
r6 = extract r5 FF
r6 = to_f32 r6
r6 = mul_f32 r0 r6
r7 = extract r5 FF00
r7 = to_f32 r7
r7 = mul_f32 r0 r7
r8 = extract r5 FF0000
r8 = to_f32 r8
r8 = mul_f32 r0 r8
r5 = shr r5 24
r5 = to_f32 r5
r5 = mul_f32 r0 r5
r9 = sub_f32 r1 r4
r6 = mul_f32 r6 r9
r7 = mul_f32 r7 r9
r8 = mul_f32 r8 r9
r9 = mad_f32 r5 r9 r4
r6 = mad_f32 r6 r2 r3
r6 = to_i32 r6
r7 = mad_f32 r7 r2 r3
r7 = to_i32 r7
r8 = mad_f32 r8 r2 r3
r8 = to_i32 r8
r9 = mad_f32 r9 r2 r3
r9 = to_i32 r9
r7 = pack r6 r7 8
r9 = pack r8 r9 8
r9 = pack r7 r9 16
store32 arg(1) r9
G8 over A8
6 registers, 12 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = sub_f32 r1 r1
r3 = splat 437F0000 (255)
r4 = splat 3F000000 (0.5)
loop:
r5 = load8 arg(1)
r5 = to_f32 r5
r5 = mul_f32 r0 r5
r5 = mad_f32 r5 r2 r1
r5 = mad_f32 r5 r3 r4
r5 = to_i32 r5
store8 arg(1) r5
G8 over G8
10 registers, 21 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = sub_f32 r1 r1
r3 = splat 3E59B3D0 (0.21259999)
r4 = splat 3F371759 (0.71520001)
r5 = splat 3D93DD98 (0.0722)
r6 = splat 437F0000 (255)
r7 = splat 3F000000 (0.5)
loop:
r8 = load8 arg(0)
r8 = to_f32 r8
r8 = mul_f32 r0 r8
r9 = load8 arg(1)
r9 = to_f32 r9
r9 = mul_f32 r0 r9
r9 = mad_f32 r9 r2 r8
r8 = mul_f32 r9 r5
r8 = mad_f32 r9 r4 r8
r8 = mad_f32 r9 r3 r8
r8 = mad_f32 r8 r6 r7
r8 = to_i32 r8
store8 arg(1) r8
G8 over RGBA_8888
10 registers, 37 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = sub_f32 r1 r1
r3 = splat 437F0000 (255)
r4 = splat 3F000000 (0.5)
loop:
r5 = load8 arg(0)
r5 = to_f32 r5
r5 = mul_f32 r0 r5
r6 = load32 arg(1)
r7 = extract r6 FF
r7 = to_f32 r7
r7 = mul_f32 r0 r7
r8 = extract r6 FF00
r8 = to_f32 r8
r8 = mul_f32 r0 r8
r9 = extract r6 FF0000
r9 = to_f32 r9
r9 = mul_f32 r0 r9
r6 = shr r6 24
r6 = to_f32 r6
r6 = mul_f32 r0 r6
r7 = mad_f32 r7 r2 r5
r8 = mad_f32 r8 r2 r5
r9 = mad_f32 r9 r2 r5
r6 = mad_f32 r6 r2 r1
r7 = mad_f32 r7 r3 r4
r7 = to_i32 r7
r8 = mad_f32 r8 r3 r4
r8 = to_i32 r8
r9 = mad_f32 r9 r3 r4
r9 = to_i32 r9
r6 = mad_f32 r6 r3 r4
r6 = to_i32 r6
r8 = pack r7 r8 8
r6 = pack r9 r6 8
r6 = pack r8 r6 16
store32 arg(1) r6
RGBA_8888 over A8
7 registers, 16 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = splat 437F0000 (255)
r3 = splat 3F000000 (0.5)
loop:
r4 = load32 arg(0)
r4 = shr r4 24
r4 = to_f32 r4
r4 = mul_f32 r0 r4
r5 = load8 arg(1)
r5 = to_f32 r5
r5 = mul_f32 r0 r5
r6 = sub_f32 r1 r4
r6 = mad_f32 r5 r6 r4
r6 = mad_f32 r6 r2 r3
r6 = to_i32 r6
store8 arg(1) r6
RGBA_8888 over G8
12 registers, 33 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = splat 3E59B3D0 (0.21259999)
r3 = splat 3F371759 (0.71520001)
r4 = splat 3D93DD98 (0.0722)
r5 = splat 437F0000 (255)
r6 = splat 3F000000 (0.5)
loop:
r7 = load32 arg(0)
r8 = extract r7 FF
r8 = to_f32 r8
r8 = mul_f32 r0 r8
r9 = extract r7 FF00
r9 = to_f32 r9
r9 = mul_f32 r0 r9
r10 = extract r7 FF0000
r10 = to_f32 r10
r10 = mul_f32 r0 r10
r7 = shr r7 24
r7 = to_f32 r7
r7 = mul_f32 r0 r7
r11 = load8 arg(1)
r11 = to_f32 r11
r11 = mul_f32 r0 r11
r7 = sub_f32 r1 r7
r8 = mad_f32 r11 r7 r8
r9 = mad_f32 r11 r7 r9
r7 = mad_f32 r11 r7 r10
r7 = mul_f32 r7 r4
r7 = mad_f32 r9 r3 r7
r7 = mad_f32 r8 r2 r7
r7 = mad_f32 r7 r5 r6
r7 = to_i32 r7
store8 arg(1) r7
RGBA_8888 over RGBA_8888
13 registers, 47 instructions:
r0 = splat 3B808081 (0.0039215689)
r1 = splat 3F800000 (1)
r2 = splat 437F0000 (255)
r3 = splat 3F000000 (0.5)
loop:
r4 = load32 arg(0)
r5 = extract r4 FF
r5 = to_f32 r5
r5 = mul_f32 r0 r5
r6 = extract r4 FF00
r6 = to_f32 r6
r6 = mul_f32 r0 r6
r7 = extract r4 FF0000
r7 = to_f32 r7
r7 = mul_f32 r0 r7
r4 = shr r4 24
r4 = to_f32 r4
r4 = mul_f32 r0 r4
r8 = load32 arg(1)
r9 = extract r8 FF
r9 = to_f32 r9
r9 = mul_f32 r0 r9
r10 = extract r8 FF00
r10 = to_f32 r10
r10 = mul_f32 r0 r10
r11 = extract r8 FF0000
r11 = to_f32 r11
r11 = mul_f32 r0 r11
r8 = shr r8 24
r8 = to_f32 r8
r8 = mul_f32 r0 r8
r12 = sub_f32 r1 r4
r9 = mad_f32 r9 r12 r5
r10 = mad_f32 r10 r12 r6
r11 = mad_f32 r11 r12 r7
r12 = mad_f32 r8 r12 r4
r9 = mad_f32 r9 r2 r3
r9 = to_i32 r9
r10 = mad_f32 r10 r2 r3
r10 = to_i32 r10
r11 = mad_f32 r11 r2 r3
r11 = to_i32 r11
r12 = mad_f32 r12 r2 r3
r12 = to_i32 r12
r10 = pack r9 r10 8
r12 = pack r11 r12 8
r12 = pack r10 r12 16
store32 arg(1) r12
I32 8888 over 8888
10 registers, 20 instructions:
r0 = splat FF (3.5733111e-43)
loop:
r1 = load32 arg(0)
r2 = extract r1 FF
r3 = extract r1 FF00
r4 = extract r1 FF0000
r1 = shr r1 24
r5 = load32 arg(1)
r6 = extract r5 FF
r7 = extract r5 FF00
r8 = extract r5 FF0000
r5 = shr r5 24
r9 = sub_i32 r0 r1
r6 = mad_unorm8 r6 r9 r2
r7 = mad_unorm8 r7 r9 r3
r8 = mad_unorm8 r8 r9 r4
r9 = mad_unorm8 r5 r9 r1
r7 = pack r6 r7 8
r9 = pack r8 r9 8
r9 = pack r7 r9 16
store32 arg(1) r9
I32 (SWAR) 8888 over 8888
7 registers, 20 instructions:
r0 = splat FF (3.5733111e-43)
r1 = splat FF00FF (2.3418409e-38)
loop:
r2 = load32 arg(0)
r3 = extract r2 FF00FF
r2 = extract r2 FF00FF00
r4 = load32 arg(1)
r5 = extract r4 FF00FF
r4 = extract r4 FF00FF00
r6 = shr r2 16
r6 = sub_i32 r0 r6
r5 = mul_i32 r5 r6
r5 = add_i32 r5 r1
r5 = extract r5 FF00FF00
r5 = add_i32 r3 r5
r6 = mul_i32 r4 r6
r6 = add_i32 r6 r1
r6 = extract r6 FF00FF00
r6 = add_i32 r2 r6
r6 = pack r5 r6 8
store32 arg(1) r6