ThreadwiseGemmDlops_km_kn_mn_v3< FloatA, FloatB, FloatC, AThreadDesc_E1_K_E2, BThreadDesc_E1_N_Ho_Wo_E2, CThreadDesc_K_N_Ho_Wo, type > Struct Template Reference

ThreadwiseGemmDlops_km_kn_mn_v3&lt; FloatA, FloatB, FloatC, AThreadDesc_E1_K_E2, BThreadDesc_E1_N_Ho_Wo_E2, CThreadDesc_K_N_Ho_Wo, type &gt; Struct Template Reference#

Composable Kernel: ck::ThreadwiseGemmDlops_km_kn_mn_v3< FloatA, FloatB, FloatC, AThreadDesc_E1_K_E2, BThreadDesc_E1_N_Ho_Wo_E2, CThreadDesc_K_N_Ho_Wo, type > Struct Template Reference
ck::ThreadwiseGemmDlops_km_kn_mn_v3< FloatA, FloatB, FloatC, AThreadDesc_E1_K_E2, BThreadDesc_E1_N_Ho_Wo_E2, CThreadDesc_K_N_Ho_Wo, type > Struct Template Reference

#include <threadwise_gemm_dlops_v3.hpp>

Static Public Member Functions

template<typename ABuffer, typename AOriginIdx, typename BBuffer, typename BOriginIdx, typename CBuffer, typename COriginIdx>
static __device__ void Run (const ABuffer &a_buf, AOriginIdx, const BBuffer &b_buf, BOriginIdx, CBuffer &c_buf, COriginIdx)

Member Function Documentation

◆ Run()

template<typename FloatA, typename FloatB, typename FloatC, typename AThreadDesc_E1_K_E2, typename BThreadDesc_E1_N_Ho_Wo_E2, typename CThreadDesc_K_N_Ho_Wo, typename enable_if< AThreadDesc_E1_K_E2::IsKnownAtCompileTime() &&BThreadDesc_E1_N_Ho_Wo_E2::IsKnownAtCompileTime() &&CThreadDesc_K_N_Ho_Wo::IsKnownAtCompileTime(), bool >::type = false>
template<typename ABuffer, typename AOriginIdx, typename BBuffer, typename BOriginIdx, typename CBuffer, typename COriginIdx>
__device__ void ck::ThreadwiseGemmDlops_km_kn_mn_v3< FloatA, FloatB, FloatC, AThreadDesc_E1_K_E2, BThreadDesc_E1_N_Ho_Wo_E2, CThreadDesc_K_N_Ho_Wo, type >::Run ( const ABuffer & a_buf,
AOriginIdx ,
const BBuffer & b_buf,
BOriginIdx ,
CBuffer & c_buf,
COriginIdx  )
inlinestatic

The documentation for this struct was generated from the following file: