COMPMID-1271: New system for GEMM heuristics
This patch implements a system for separating the "validity" from
"preferred" aspect of the current heuristics in gemm_*.cpp.
Now, each gemm_*.cpp defines a list of candidate implementations,
each of which supplies an is_valid() function (to check for
validity), an is_preferred() function (the "heuristic" part), and an
instantiate() function which actually produces the GemmCommon object
pointer.
The actual gemm() function is now templated and uses this list to
select an implementation. This patch also implements a mechanism to
identify the preferred implementation, and override it via the
GemmConfig structure.
Change-Id: Id49ab7af8bf2e3e9fd951a9698883ade234d40e1
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/139120
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: Jenkins <bsgcomp@arm.com>
diff --git a/src/core/NEON/kernels/arm_gemm/gemm_int16.cpp b/src/core/NEON/kernels/arm_gemm/gemm_int16.cpp
index 57cd15f..f61cc13 100644
--- a/src/core/NEON/kernels/arm_gemm/gemm_int16.cpp
+++ b/src/core/NEON/kernels/arm_gemm/gemm_int16.cpp
@@ -25,20 +25,36 @@
#include "arm_gemm.hpp"
#include "gemm_common.hpp"
+#include "gemm_implementation.hpp"
#include "gemm_interleaved.hpp"
#include "kernels/a64_gemm_s16_12x8.hpp"
namespace arm_gemm {
+class GemmImpl_gemm_s16_interleaved : public GemmImplementation<int16_t, int32_t> {
+public:
+ UniqueGemmCommon<int16_t, int32_t> instantiate(const GemmArgs<int32_t> &args) override {
+ return UniqueGemmCommon<int16_t, int32_t>(new GemmInterleaved<gemm_s16_12x8, int16_t, int32_t>(args));
+ }
+
+ GemmImpl_gemm_s16_interleaved() : GemmImplementation<int16_t, int32_t>(GemmMethod::GEMM_INTERLEAVED) { }
+};
+
+static std::vector<GemmImplementation<int16_t, int32_t> *> gemm_s16_methods = {
+ new GemmImpl_gemm_s16_interleaved()
+};
+
template<>
-UniqueGemmCommon<int16_t, int32_t> gemm<int16_t, int32_t>(const CPUInfo &ci, const unsigned int M, const unsigned int N, const unsigned int K,
- const unsigned int nbatches, const unsigned int nmulti,
- const bool trA, const bool trB, const int32_t alpha, const int32_t beta,
- const int maxthreads, const bool pretransposed_hint) {
- return UniqueGemmCommon<int16_t, int32_t>(new GemmInterleaved<gemm_s16_12x8, int16_t, int32_t>(&ci, M, N, K, nbatches, nmulti, trA, trB, alpha, beta, maxthreads, pretransposed_hint));
+std::vector<GemmImplementation<int16_t, int32_t> *> &gemm_implementation_list<int16_t, int32_t>() {
+ return gemm_s16_methods;
}
+/* Explicitly instantiate the external functions for these types. */
+template UniqueGemmCommon<int16_t, int32_t> gemm<int16_t, int32_t>(GemmArgs<int32_t> &args, GemmConfig *cfg);
+template GemmMethod get_gemm_method<int16_t, int32_t>(GemmArgs<int32_t> &args);
+template bool method_is_compatible<int16_t, int32_t>(GemmMethod method, GemmArgs<int32_t> &args);
+
} // namespace arm_gemm
#endif // __aarch64__