Henrik RydgårdandClaude Opus 5.5 bc7afa8845 ARM32 IR JIT: Call a NEON dot routine for the accurate dot products ...

After clang's ARMv7 output for vfpu_dot_neon(), in Q8-Q15, which the register
cache leaves alone. The end is done in 32-bit integers instead of 64-bit ones,
checked bit-exact against vfpu_dot_finish(). The call site gathers the lanes into
Q8 and Q9 and flushes whatever guest state is in R2, R3, R12 and LR, usually
nothing. Inf and NaN go through a routine that calls vfpu_dot_reference().
Without NEON, Vec4VfpuDot still goes through the IR interpreter.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-10-14 22:30:37 +02:00

..

ArmIRAsm.cpp

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00

ArmIRCompALU.cpp

ARM32 IR JIT: Integer arithmetic, logic, shifts, compares, mult and div

2026-10-14 17:32:53 +02:00

ArmIRCompBranch.cpp

ARM32: Start an IR-based JIT

2026-10-14 17:32:53 +02:00

ArmIRCompFPU.cpp

ARM32 IR JIT: Pass float helper arguments the softfp way where needed

2026-10-14 18:23:00 +02:00

ArmIRCompLoadStore.cpp

ARM32 IR JIT: Loads and stores

2026-10-14 17:32:53 +02:00

ArmIRCompSystem.cpp

ARM32: Start an IR-based JIT

2026-10-14 17:32:53 +02:00

ArmIRCompVec.cpp

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00

ArmIRJit.cpp

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00

ArmIRJit.h

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00

ArmIRRegCache.cpp

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00

ArmIRRegCache.h

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00

ArmVfpuDot.h

ARM32 IR JIT: Call a NEON dot routine for the accurate dot products

2026-10-14 22:30:37 +02:00