Jit64: Avoid unnecessary MOVAPS instructions #8992
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
AVX has a four-operand VBLENDVPD instruction, which allows for the first input and the destination to be different. By taking advantage of this, we no longer need to copy one of the inputs around and we can just reference it directly, provided it's already in a register (I have yet to see this not be the case). Before: 66 0F 57 C0 xorpd xmm0,xmm0 F2 41 0F C2 C6 06 cmpnlesd xmm0,xmm14 41 0F 28 CE movaps xmm1,xmm14 66 41 0F 38 15 CA blendvpd xmm1,xmm10,xmm0 F2 44 0F 10 F1 movsd xmm14,xmm1 After: 66 0F 57 C0 xorpd xmm0,xmm0 F2 41 0F C2 C6 06 cmpnlesd xmm0,xmm14 C4 C3 09 4B CA 00 vblendvpd xmm1,xmm14,xmm10,xmm0 F2 44 0F 10 F1 movsd xmm14,xmm1
For the packed variant, we can skip the final MOVAPS and write the result directly into the destination register. Before: 66 0F 57 C0 xorpd xmm0,xmm0 66 41 0F C2 C1 06 cmpnlepd xmm0,xmm9 C4 C3 09 4B CC 00 vblendvpd xmm1,xmm14,xmm12,xmm0 44 0F 28 F1 movaps xmm14,xmm1 After: 66 0F 57 C0 xorpd xmm0,xmm0 66 41 0F C2 C1 06 cmpnlepd xmm0,xmm9 C4 43 09 4B F4 00 vblendvpd xmm14,xmm14,xmm12,xmm0
Pretty much the same optimization we did for AVX, although slightly more constrained because we're stuck with the two-operand instruction where destination and source have to match. We could also specialize the case where registers b, c, and d are all distinct, but I decided against it since I couldn't find any game that does this. Before: 66 0F 57 C0 xorpd xmm0,xmm0 66 41 0F C2 C1 06 cmpnlepd xmm0,xmm9 41 0F 28 CE movaps xmm1,xmm14 66 41 0F 38 15 CC blendvpd xmm1,xmm12,xmm0 44 0F 28 F1 movaps xmm14,xmm1 After: 66 0F 57 C0 xorpd xmm0,xmm0 66 41 0F C2 C1 06 cmpnlepd xmm0,xmm9 66 45 0F 38 15 F4 blendvpd xmm14,xmm12,xmm0
|
Have you tried some GameCube games? https://github.com/devkitPro/libogc/blob/8d525d412737b20e70ca931120d7cb4d0b15b529/libogc/gu.c#L951 Edit: Oh, I guess |
Using AVX we can eliminate another MOVAPS instruction here. Before: 0F 28 C8 movaps xmm1,xmm0 66 0F DB 0D CF 2C 00 00 pand xmm1,xmmword ptr [1F8D283B220h] After: C5 F9 DB 0D D2 2C 00 00 vpand xmm1,xmm0,xmmword ptr [271835FB220h]
Sintendo
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
We can avoid MOVAPS in a few places by taking advantage of AVX or by realizing when we can write the result directly into the target register.
I've only been able to find one game that uses ps_sel at all (Mario Kart Wii). If anyone knows of any others, please let me know.
fsel with AVX
Before:
After:
ps_sel with AVX
Before:
After:
ps_sel with SSE4.1
Before:
After:
ConvertDoubleToSingle with AVX
Before:
After: