SoftGPU: Make arm64 and x86 render the GE probes identically

Found by the probe regression run (ppsspp-re geprobe/regress.py), which
gave different counts on the two:

- Vec4<int> >> was a logical shift on SSE, arithmetic elsewhere. A negative
  spot factor lit pixels white on x86 (gpu/probe exp100).
- Vec4<float>::ToRGBA rounded on SSE and truncated on NEON and the generic
  path. Spline colors were one lower on ARM (exp48-51, the PSP rounds).
- Texel coordinates past the int range: the C++ sampler now gives INT_MIN
  like the x86 JIT's conversion, as the PSP does with a repeating texture (exp5).
- The viewer direction is normalized with the GE's normalization instead of
  the SIMD float one (exp2, exp3).

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Henrik RydgårdandClaude Opus 5.5 committed 2026-10-02 15:09:43 -06:00
1 parent df5b55d5b8
commit c8819d5087
3 files changed
+26 -11

No files matched your search

+5 -1
View File
@@ -377,7 +377,11 @@ void ComputeTransformState(TransformState *state, const VertexReader &vreader) {
canSkipWorldPos = false;
}
}
state->lightingState.viewDir = PSPViewDirection(gstate.viewMatrix);
// The viewer direction (see PSPViewDirection), normalized like the GE does.
Vec3f viewDir(gstate.viewMatrix[2], gstate.viewMatrix[5], gstate.viewMatrix[8]);
if (GENormalize(viewDir) == 0.0f)
viewDir = Vec3f(0.0f, 0.0f, 1.0f);
state->lightingState.viewDir = viewDir;
float world[16];
float view[16];