Henrik RydgårdandClaude Opus 5.5 16a5046124 SoftGPU: Read bilinear CLUT and 8888 texels in vector lanes ...

The four-pixel bilinear sampler read its sixteen texels one at a time,
each with its CLUT lookup and index transform, and loaded them as
vectors afterwards. For CLUT4 and CLUT8 with a 32-bit palette and no
index shift (by far the most common), and for 8888, the texels are now
gathered in lanes: the index bytes, the nibble, the mask and start, then
the palette entries. CrossSIMD gets Vec4S32::Gather and GatherU8 for it,
one load per lane. LocoRoco executes 5.5% fewer instructions, Burnout 3%.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

2026-10-10 15:01:49 +02:00

..

fast

Address LSX CrossSIMD review feedback

2026-05-25 15:40:22 +02:00

lin

Remove a confusing operator overload from Lin::Vec3

2026-05-19 16:33:45 +02:00

CrossSIMD.h

SoftGPU: Read bilinear CLUT and 8888 texels in vector lanes

2026-10-10 15:01:49 +02:00

curves.cpp

Fix scaling issues in cocoa text drawer

2024-06-01 13:44:11 +02:00

curves.h

Move math and some file and data conversion files out from native to Common.

2020-10-04 09:12:46 +02:00

expression_parser.cpp

expression_parser: bound token length to fix stack buffer overflow

2026-08-10 01:00:28 +02:00

expression_parser.h

Expose PSPThread in the same manner

2025-03-31 10:24:03 +02:00

geom2d.h

Rename FocusDirection to FocusMove, make into enum class

2026-04-14 18:17:54 -06:00

math_util.cpp

Experimentally revert "Delete some misguided FPU control code, that we definitely don't want"

2022-10-10 18:35:29 +02:00

math_util.h

FPU: saturate float-to-int in the interpreters

2026-09-22 14:25:41 -06:00

SIMDHeaders.h

GEMath: GERowSum4 takes its left operands as one vector

2026-10-09 09:07:12 +02:00

Statistics.cpp

Track min and max GPU time, init time, render CPU time.

2022-10-10 10:53:27 +02:00

Statistics.h

Vulkan profiler: Show CPU command buffer recording timing even if GPU timing is unavailable

2023-06-14 09:15:29 +02:00