Fix vqshlud_n_s64 implementation to be 64-bit.

The scalar vqshlud_n_s64 function was wrong, and there was no test case for it (or the scalar 32 bit form). Add test cases for the scalar functions (test vectors generated on GCE ARMv9 system), validate the fix for the 64-bit scalar form.
simd-everywhere · Jan 4, 2025 · 3527e86 · 3527e86
1 parent 858b005
commit 3527e86
Show file tree

Hide file tree

Showing 2 changed files with 1,210 additions and 86 deletions.
diff --git a/simde/arm/neon/qshlu_n.h b/simde/arm/neon/qshlu_n.h
@@ -94,8 +94,8 @@ SIMDE_FUNCTION_ATTRIBUTES
 uint64_t
 simde_vqshlud_n_s64(int64_t a, const int n)
     SIMDE_REQUIRE_CONSTANT_RANGE(n, 0, 63) {
-  uint32_t r = HEDLEY_STATIC_CAST(uint32_t, a << n);
-  r |= (((r >> n) != HEDLEY_STATIC_CAST(uint32_t, a)) ? UINT32_MAX : 0);
+  uint64_t r = HEDLEY_STATIC_CAST(uint64_t, a << n);
+  r |= (((r >> n) != HEDLEY_STATIC_CAST(uint64_t, a)) ? UINT64_MAX : 0);
   return (a < 0) ? 0 : r;
 }
 #if defined(SIMDE_ARM_NEON_A64V8_NATIVE)