commit c22e1263ba3a6945ceb1fdccb33eac512fd156fb
author David Mansell <David.Mansell@arm.com> Fri May 03 13:24:48 2024 +0100
committer Pablo Marquez Tello <pablo.tello@arm.com> Fri May 03 13:33:02 2024 +0000
tree 0240a38789ef97e5f1ccbc2d049acbdec25379ef
parent 0c5ba9ee5ad36aeb6d684c20d473a27f482e83fb

arm_gemm: fix SVE check on fast mode kernels.

SVE BF16 kernels need to check for svebf16(), not just bf16().

Change-Id: I89494aac40166eba59719bed9822194a48ac282d
Signed-off-by: David Mansell <David.Mansell@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11520
Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>

src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp

diff --git a/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp b/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp
index f223dea..0c1d3a3 100644
--- a/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp
+++ b/src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp
@@ -199,14 +199,14 @@
 GemmImplementation<float, float>::with_estimate(
     GemmMethod::GEMM_HYBRID,
     "sve_hybrid_fp32bf16fp32_mmla_6x4VL",
-    [](const GemmArgs &args) { return args._fast_mode && args._ci->has_bf16(); },
+    [](const GemmArgs &args) { return args._fast_mode && args._ci->has_svebf16(); },
     [](const GemmArgs &args) { return GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_6x4VL, float, float>::estimate_cycles<float>(args); },
     [](const GemmArgs &args) { return new GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_6x4VL, float, float>(args); }
 ),
 GemmImplementation<float, float>::with_estimate(
     GemmMethod::GEMM_HYBRID,
     "sve_hybrid_fp32bf16fp32_mmla_4x6VL",
-    [](const GemmArgs &args) { return args._fast_mode && args._ci->has_bf16(); },
+    [](const GemmArgs &args) { return args._fast_mode && args._ci->has_svebf16(); },
     [](const GemmArgs &args) { return GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_4x6VL, float, float>::estimate_cycles<float>(args); },
     [](const GemmArgs &args) { return new GemmHybridIndirect<cls_sve_hybrid_fp32bf16fp32_mmla_4x6VL, float, float>(args); }
 ),