diff options
| -rw-r--r-- | target/arm/vec_helper.c | 2 | ||||
| -rw-r--r-- | target/arm/vec_internal.h | 13 |
2 files changed, 14 insertions, 1 deletions
diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index 26c373e..9a9c034 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -2557,7 +2557,7 @@ DO_MMLA_B(gvec_usmmla_b, do_usmmla_b) * BFloat16 Dot Product */ -static float32 bfdotadd(float32 sum, uint32_t e1, uint32_t e2) +float32 bfdotadd(float32 sum, uint32_t e1, uint32_t e2) { /* FPCR is ignored for BFDOT and BFMMLA. */ float_status bf_status = { diff --git a/target/arm/vec_internal.h b/target/arm/vec_internal.h index 1d527fa..1f4ed80 100644 --- a/target/arm/vec_internal.h +++ b/target/arm/vec_internal.h @@ -230,4 +230,17 @@ uint64_t pmull_h(uint64_t op1, uint64_t op2); */ uint64_t pmull_w(uint64_t op1, uint64_t op2); +/** + * bfdotadd: + * @sum: addend + * @e1, @e2: multiplicand vectors + * + * BFloat16 2-way dot product of @e1 & @e2, accumulating with @sum. + * The @e1 and @e2 operands correspond to the 32-bit source vector + * slots and contain two Bfloat16 values each. + * + * Corresponds to the ARM pseudocode function BFDotAdd. + */ +float32 bfdotadd(float32 sum, uint32_t e1, uint32_t e2); + #endif /* TARGET_ARM_VEC_INTERNAL_H */ |