7d987cd78b
Make fast_matrix_mul_4x4 inlineable
2026-01-30 14:10:32 +01:00
58eaf56c5e
libretro: build fix
...
gcc9 and the version of msvc19 that the libretro builders use don't
have _mm_storeu_si32. Using _m_cvtsi128_si32 followed by memcpy is
picked up by compilers that support _mm_storeu_si32 as being the same
and they generate the same assembly, so this is no harm to them.
2026-01-28 00:15:35 -05:00
47f6b9975e
Optimize the color alpha computation for color morph
2026-01-20 13:51:24 +01:00
3378faca5e
Optimize morph code (non-JIT vertex decoders)
2026-01-20 13:51:24 +01:00
a8f89e1544
Correct rounding in StoreConvertToU8, correct Vec4F32::Store3.
2026-01-20 13:51:24 +01:00
fc09669ba7
Add DualSense support. Remove trying to support bluetooth devices (not reliable without more work)
2025-07-03 11:11:34 +02:00
fd88f79d07
CrossSIMD: Fix more no-simd fallbacks. The depth rasterizer now works in TEST_FALLBACK mode.
2025-02-10 11:51:22 -06:00
8907a4c001
More fixes, improve unit test
2025-02-06 10:26:46 -06:00
fc0b1eef4c
Fix 4x4
2025-02-06 10:26:46 -06:00
07e892e1a2
Fix some type mismatches in CrossSIMD.h that not all compilers complain about.
...
Fixes #19905
2025-02-01 10:30:02 -06:00
c78fa60431
Add better way to check if CrossSIMD has been natively implemented
2025-01-28 10:56:52 +01:00
acd5b24924
Complete CrossSIMD non-simd fallback (although buggy, it seems). Minor ARM64 opt.
2025-01-28 10:54:43 +01:00
2aaa1e5379
CrossSIMD: Expand the no-simd path
2025-01-28 10:54:43 +01:00
74501b06b6
CrossSIMD: Add more no-simd fallback types
2025-01-28 10:54:43 +01:00
9d164b71fb
Some new CrossSIMD operations
2025-01-28 10:54:43 +01:00
9ccf47a6b3
Fix type error in minor ARM64 optimization, which some compilers don't like.
...
Fixes #19905
2025-01-22 09:54:20 +01:00
a5116e1590
Use vmaxvq_s32 to implement AnyZeroSignBit more efficiently on ARM64
2025-01-19 17:49:09 +01:00
206d4d1fea
Implement the low-quality depth raster mode, default to it on Android/iOS.
...
I really can't tell much of a difference in practice...
2024-12-31 11:19:38 +01:00
f85d7db5b1
Comment fixes, buildfix
2024-12-31 02:39:58 +01:00
dee5fe6990
Fix issue in Midnight Club where Z now wrapped around at a distance, after removing the clamp. Might as well cull.
2024-12-31 02:30:05 +01:00
f5cc41caab
More CrossSIMD (breaking change)
2024-12-31 01:59:11 +01:00
c3ac798545
More crosssimd
2024-12-31 01:59:08 +01:00
ef934df0f2
Compute and cull by triangle area early before writing 4-groups of triangles
2024-12-29 01:21:16 +01:00
e8786fc401
Cull 4-groups of triangles early
2024-12-29 01:12:53 +01:00
66926e28c1
Simplify / fix through mode triangles
2024-12-28 18:51:37 +01:00
eec7853efe
More SIMD: Add some matrix operations to CrossSIMD ( #19773 )
...
* More CrossSIMD functionality
* Use the new SIMD API for the matrix multiplies
2024-12-28 18:45:14 +01:00
8c069917b5
Depth raster optimizations: Merge viewport into projection matrix, prepare for further SIMD-ification ( #19769 )
...
* Remove some old redundant reports
* Fix scissor off by one
* More CrossSIMD
* Move the viewport scale out to the proj matrix
* DepthRaster: Also merge the viewport translation into the projection matrix.
* Depth raster: Do the triangle clipping in batches of 4 triangles
* Cleanup
2024-12-28 10:36:39 +01:00
cc040ab251
Depth raster: Bugfix, minor opt ( #19768 )
...
* Correct two errors in CrossSIMD.h, thanks hiroyuki177
Fixes #19767
* DepthRaster: Merge offset into viewport X/Y
2024-12-26 20:23:52 +01:00
0e59cd7641
DepthRaster: Fix typo breaking LESS depth comparison mode on x86(64)
2024-12-24 23:26:32 +01:00
8e747dc948
CrossSIMD: Add a multiply-as-16bit function to Vec4S32. This can be implemented quickly on SSE2.
2024-12-22 18:53:10 +01:00
3bd2441f4a
CrossSIMD: Add SwapLowerElements to Vec4F32
2024-12-22 10:32:16 +01:00
bd3a252104
CrossSIMD: Add a simple operation
2024-12-21 21:21:40 +01:00
8cd86b47b5
AnyZeroSignBit arm fix, more crosssimd fixes. Now works on ARM.
2024-12-21 14:28:01 +01:00
5df88fc1aa
Convert the rect implementation to CrossSIMD
2024-12-21 14:28:01 +01:00
399570e411
CrossSIMD: make the transpose function compatible with ARM32
2024-12-21 14:28:00 +01:00
2051d55c90
CrossSIMD: Add a bunch more functonality for use by the rasterizer
2024-12-21 14:28:00 +01:00
65692d036e
CrossSIMD: possible buildfix?
2024-12-21 14:28:00 +01:00
0b009c10be
CrossSIMD: Add reciprocal, clamp, swaplowerelements, etc
2024-12-21 14:28:00 +01:00
03b9f98055
Add more funcionality to CrossSIMD.h, like fast matrix mul and some conversion
2024-12-21 14:28:00 +01:00
de45960420
Reformat CrossSIMD.h for easier editing. Add some new methods.
2024-12-21 14:28:00 +01:00
c92b3b6521
Move prototype cross simd wrapper structs to CrossSIMD.h
2024-12-21 14:27:59 +01:00
c5ad81e3d5
Add DepthRaster.cpp/h. Rasterize depth rectangles, some triangles
2024-12-21 14:27:53 +01:00
5326d87f9c
Rename CrossSIMD to SIMDHeader, but also keep CrossSIMD.h (will have a future use)
2024-12-19 15:15:43 +01:00
45cf6d423e
Buildfixes
2024-12-18 13:57:29 +01:00
a6a189c24d
An old version of android had strict JNI validation of UTF-8 strings enabled. Let's pass it.
2024-10-22 11:02:26 +02:00
531f145e0b
Fix NEON compilation error on strict compilers. Fixes #18688
2024-01-18 10:38:30 +01:00
aee8f62ed5
Buildfix (included from c file)
2024-01-11 14:59:54 +01:00
7e85d3d10a
Disable the new culling on RISC-V for now.
2023-12-09 16:49:02 +01:00
4e2a1bf81c
NEON: vcvtq can scale directly, no need for a mul by const.
2023-12-09 16:48:59 +01:00
99548be8a3
NEON culling: Use mla operations to shave off some more cycles. ARM32 compat.
2023-12-09 16:36:01 +01:00
Henrik Rydgård