mirror of
https://github.com/paboyle/Grid.git
synced 2024-11-09 23:45:36 +00:00
Improved the BLAS benchmark
This commit is contained in:
Peter Boyle
parent
2c9878fc3a
commit
f568c07bbd
@ -12,8 +12,8 @@
|
||||
#include <iostream>
|
||||
#include <sys/time.h>
|
||||
|
||||
#define GRID_SYCL
|
||||
#undef GRID_HIP
|
||||
#undef GRID_SYCL
|
||||
#define GRID_HIP
|
||||
#undef GRID_CUDA
|
||||
|
||||
#ifdef GRID_HIP
|
||||
@ -21,6 +21,7 @@
|
||||
#endif
|
||||
#ifdef GRID_CUDA
|
||||
#include <cublas_v2.h>
|
||||
|
||||
#endif
|
||||
#ifdef GRID_SYCL
|
||||
#include <oneapi/mkl.hpp>
|
||||
@ -45,6 +46,90 @@ inline void acceleratorFreeDevice(void *ptr,size_t bytes){free(ptr,*theAccelerat
|
||||
inline void acceleratorMemSet(void *base,int value,size_t bytes) { theAccelerator->memset(base,value,bytes); theAccelerator->wait();}
|
||||
inline void acceleratorCopyToDevice(void *from,void *to,size_t bytes) { theAccelerator->memcpy(to,from,bytes); theAccelerator->wait();}
|
||||
inline void acceleratorCopyFromDevice(void *from,void *to,size_t bytes){ theAccelerator->memcpy(to,from,bytes); theAccelerator->wait();}
|
||||
#define accelerator_barrier(dummy) { theAccelerator->wait(); }
|
||||
#endif
|
||||
|
||||
#ifdef GRID_HIP
|
||||
hipStream_t copyStream;
|
||||
hipStream_t computeStream;
|
||||
void acceleratorInit(void)
|
||||
{
|
||||
int device = 0;
|
||||
auto discard = hipSetDevice(device);
|
||||
discard = hipStreamCreate(©Stream);
|
||||
discard = hipStreamCreate(&computeStream);
|
||||
printf("AcceleratorHIPInit\n");
|
||||
}
|
||||
inline void *acceleratorAllocDevice(size_t bytes)
|
||||
{
|
||||
void *ptr=NULL;
|
||||
auto err = hipMalloc((void **)&ptr,bytes);
|
||||
if( err != hipSuccess ) {
|
||||
ptr = (void *) NULL;
|
||||
fprintf(stderr," hipMalloc failed for %ld %s \n",bytes,hipGetErrorString(err)); fflush(stderr);
|
||||
}
|
||||
return ptr;
|
||||
};
|
||||
inline void acceleratorFreeDevice(void *ptr,size_t bytes){ auto discard=hipFree(ptr);};
|
||||
inline void acceleratorFreeDevice(void *ptr){ auto discard=hipFree(ptr);};
|
||||
inline void acceleratorMemSet(void *base,int value,size_t bytes) { auto discard=hipMemset(base,value,bytes);}
|
||||
inline void acceleratorCopyToDevice(void *from,void *to,size_t bytes) { auto discard=hipMemcpy(to,from,bytes, hipMemcpyHostToDevice);}
|
||||
inline void acceleratorCopyFromDevice(void *from,void *to,size_t bytes){ auto discard=hipMemcpy(to,from,bytes, hipMemcpyDeviceToHost);}
|
||||
#define accelerator_barrier(dummy) \
|
||||
{ \
|
||||
auto tmp=hipStreamSynchronize(computeStream); \
|
||||
auto err = hipGetLastError(); \
|
||||
if ( err != hipSuccess ) { \
|
||||
printf("After hipDeviceSynchronize() : HIP error %s \n", hipGetErrorString( err )); \
|
||||
puts(__FILE__); \
|
||||
printf("Line %d\n",__LINE__); \
|
||||
exit(0); \
|
||||
} \
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
#ifdef GRID_CUDA
|
||||
cudaStream_t copyStream;
|
||||
cudaStream_t computeStream;
|
||||
void acceleratorInit(void)
|
||||
{
|
||||
int device = 0;
|
||||
cudaSetDevice(device);
|
||||
cudaStreamCreate(©Stream);
|
||||
cudaStreamCreate(&computeStream);
|
||||
}
|
||||
inline void *acceleratorAllocDevice(size_t bytes)
|
||||
{
|
||||
void *ptr=NULL;
|
||||
auto err = cudaMalloc((void **)&ptr,bytes);
|
||||
if( err != cudaSuccess ) {
|
||||
ptr = (void *) NULL;
|
||||
printf(" cudaMalloc failed for %d %s \n",bytes,cudaGetErrorString(err));
|
||||
}
|
||||
return ptr;
|
||||
};
|
||||
inline void acceleratorFreeShared(void *ptr){ cudaFree(ptr);};
|
||||
inline void acceleratorFreeDevice(void *ptr){ cudaFree(ptr);};
|
||||
inline void acceleratorCopyToDevice(void *from,void *to,size_t bytes) { cudaMemcpy(to,from,bytes, cudaMemcpyHostToDevice);}
|
||||
inline void acceleratorCopyFromDevice(void *from,void *to,size_t bytes){ cudaMemcpy(to,from,bytes, cudaMemcpyDeviceToHost);}
|
||||
inline void acceleratorMemSet(void *base,int value,size_t bytes) { cudaMemset(base,value,bytes);}
|
||||
#define accelerator_barrier(dummy) \
|
||||
{ \
|
||||
cudaStreamSynchronize(computeStream); \
|
||||
cudaError err = cudaGetLastError(); \
|
||||
if ( cudaSuccess != err ) { \
|
||||
printf("accelerator_barrier(): Cuda error %s \n", \
|
||||
cudaGetErrorString( err )); \
|
||||
printf("File %s Line %d\n",__FILE__,__LINE__); \
|
||||
fflush(stdout); \
|
||||
if (acceleratorAbortOnGpuError) assert(err==cudaSuccess); \
|
||||
} \
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
|
||||
template<class T> void acceleratorPut(T& dev,T&host)
|
||||
{
|
||||
acceleratorCopyToDevice(&host,&dev,sizeof(T));
|
||||
@ -55,9 +140,6 @@ template<class T> T acceleratorGet(T& dev)
|
||||
acceleratorCopyFromDevice(&dev,&host,sizeof(T));
|
||||
return host;
|
||||
}
|
||||
#define accelerator_barrier(dummy) { theAccelerator->wait(); }
|
||||
#endif
|
||||
|
||||
|
||||
/**************************************************************
|
||||
* Allocator
|
||||
@ -210,7 +292,270 @@ public:
|
||||
gridblasHandle->wait();
|
||||
#endif
|
||||
}
|
||||
|
||||
|
||||
/////////////////////////////////////////////////////////////
|
||||
// Single matrix GEMM -- fp64 and fp32
|
||||
/////////////////////////////////////////////////////////////
|
||||
void gemm(GridBLASOperation_t OpA,
|
||||
GridBLASOperation_t OpB,
|
||||
int m,int n, int k,
|
||||
ComplexD alpha,
|
||||
ComplexD* Amk, // Device pointer
|
||||
ComplexD* Bkn,
|
||||
ComplexD beta,
|
||||
ComplexD* Cmn)
|
||||
{
|
||||
RealD t2=usecond();
|
||||
|
||||
assert(OpA!=GridBLAS_OP_T); // Complex case expect no transpose
|
||||
assert(OpB!=GridBLAS_OP_T);
|
||||
|
||||
int lda = m; // m x k column major
|
||||
int ldb = k; // k x n column major
|
||||
int ldc = m; // m x b column major
|
||||
if(OpA!=GridBLAS_OP_N)
|
||||
lda = k;
|
||||
if(OpB!=GridBLAS_OP_N)
|
||||
ldb = n;
|
||||
|
||||
static deviceVector<ComplexD> alpha_p(1);
|
||||
static deviceVector<ComplexD> beta_p(1);
|
||||
// can prestore the 1 and the zero on device
|
||||
acceleratorCopyToDevice((void *)&alpha,(void *)&alpha_p[0],sizeof(ComplexD));
|
||||
acceleratorCopyToDevice((void *)&beta ,(void *)&beta_p[0],sizeof(ComplexD));
|
||||
RealD t0=usecond();
|
||||
|
||||
#ifdef GRID_HIP
|
||||
hipblasOperation_t hOpA;
|
||||
hipblasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = HIPBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = HIPBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = HIPBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = HIPBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = HIPBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = HIPBLAS_OP_C;
|
||||
auto err = hipblasZgemm(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(hipblasDoubleComplex *) &alpha_p[0],
|
||||
(hipblasDoubleComplex *) Amk, lda,
|
||||
(hipblasDoubleComplex *) Bkn, ldb,
|
||||
(hipblasDoubleComplex *) &beta_p[0],
|
||||
(hipblasDoubleComplex *) Cmn, ldc);
|
||||
assert(err==HIPBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_CUDA
|
||||
cublasOperation_t hOpA;
|
||||
cublasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = CUBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = CUBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = CUBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = CUBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = CUBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = CUBLAS_OP_C;
|
||||
auto err = cublasZgemm(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(cuDoubleComplex *) &alpha_p[0],
|
||||
(cuDoubleComplex *) Amk, lda,
|
||||
(cuDoubleComplex *) Bkn, ldb,
|
||||
(cuDoubleComplex *) &beta_p[0],
|
||||
(cuDoubleComplex *) Cmn, ldc);
|
||||
assert(err==CUBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_SYCL
|
||||
int64_t m64=m;
|
||||
int64_t n64=n;
|
||||
int64_t k64=k;
|
||||
int64_t lda64=lda;
|
||||
int64_t ldb64=ldb;
|
||||
int64_t ldc64=ldc;
|
||||
|
||||
oneapi::mkl::transpose iOpA;
|
||||
oneapi::mkl::transpose iOpB;
|
||||
|
||||
if ( OpA == GridBLAS_OP_N ) iOpA = oneapi::mkl::transpose::N;
|
||||
if ( OpA == GridBLAS_OP_T ) iOpA = oneapi::mkl::transpose::T;
|
||||
if ( OpA == GridBLAS_OP_C ) iOpA = oneapi::mkl::transpose::C;
|
||||
if ( OpB == GridBLAS_OP_N ) iOpB = oneapi::mkl::transpose::N;
|
||||
if ( OpB == GridBLAS_OP_T ) iOpB = oneapi::mkl::transpose::T;
|
||||
if ( OpB == GridBLAS_OP_C ) iOpB = oneapi::mkl::transpose::C;
|
||||
|
||||
oneapi::mkl::blas::column_major::gemm(*gridblasHandle,
|
||||
&iOpA,
|
||||
&iOpB,
|
||||
&m64,&n64,&k64,
|
||||
(ComplexD *) &alpha_p[0],
|
||||
(const ComplexD *)Amk, (const int64_t *)&lda64,
|
||||
(const ComplexD *)Bkn, (const int64_t *)&ldb64,
|
||||
(ComplexD *) &beta_p[0],
|
||||
(ComplexD *)Cmn, (const int64_t *)&ldc64);
|
||||
synchronise();
|
||||
#endif
|
||||
#if !defined(GRID_SYCL) && !defined(GRID_CUDA) && !defined(GRID_HIP)
|
||||
// Need a default/reference implementation; use Eigen
|
||||
if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_N) ) {
|
||||
Eigen::Map<Eigen::MatrixXcd> eAmk(Amk,m,k);
|
||||
Eigen::Map<Eigen::MatrixXcd> eBkn(Bkn,k,n);
|
||||
Eigen::Map<Eigen::MatrixXcd> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn ;
|
||||
} else if ( (OpA == GridBLAS_OP_C ) && (OpB == GridBLAS_OP_N) ) {
|
||||
Eigen::Map<Eigen::MatrixXcd> eAmk(Amk,k,m);
|
||||
Eigen::Map<Eigen::MatrixXcd> eBkn(Bkn,k,n);
|
||||
Eigen::Map<Eigen::MatrixXcd> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.adjoint() * eBkn ;
|
||||
} else if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_C) ) {
|
||||
Eigen::Map<Eigen::MatrixXcd> eAmk(Amk,m,k);
|
||||
Eigen::Map<Eigen::MatrixXcd> eBkn(Bkn,n,k);
|
||||
Eigen::Map<Eigen::MatrixXcd> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn.adjoint() ;
|
||||
} else if ( (OpA == GridBLAS_OP_C ) && (OpB == GridBLAS_OP_C) ) {
|
||||
Eigen::Map<Eigen::MatrixXcd> eAmk(Amk,k,m);
|
||||
Eigen::Map<Eigen::MatrixXcd> eBkn(Bkn,n,k);
|
||||
Eigen::Map<Eigen::MatrixXcd> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.adjoint() * eBkn.adjoint() ;
|
||||
} else {
|
||||
assert(0);
|
||||
}
|
||||
#endif
|
||||
RealD t1=usecond();
|
||||
RealD flops = 8.0*m*n*k;
|
||||
RealD bytes = 1.0*sizeof(ComplexD)*(m*k+k*n+m*n);
|
||||
}
|
||||
void gemm(GridBLASOperation_t OpA,
|
||||
GridBLASOperation_t OpB,
|
||||
int m,int n, int k,
|
||||
ComplexF alpha,
|
||||
ComplexF* Amk, // Device pointer
|
||||
ComplexF* Bkn,
|
||||
ComplexF beta,
|
||||
ComplexF* Cmn)
|
||||
{
|
||||
RealD t2=usecond();
|
||||
|
||||
assert(OpA!=GridBLAS_OP_T); // Complex case expect no transpose
|
||||
assert(OpB!=GridBLAS_OP_T);
|
||||
|
||||
int lda = m; // m x k column major
|
||||
int ldb = k; // k x n column major
|
||||
int ldc = m; // m x b column major
|
||||
if(OpA!=GridBLAS_OP_N)
|
||||
lda = k;
|
||||
if(OpB!=GridBLAS_OP_N)
|
||||
ldb = n;
|
||||
|
||||
static deviceVector<ComplexF> alpha_p(1);
|
||||
static deviceVector<ComplexF> beta_p(1);
|
||||
// can prestore the 1 and the zero on device
|
||||
acceleratorCopyToDevice((void *)&alpha,(void *)&alpha_p[0],sizeof(ComplexF));
|
||||
acceleratorCopyToDevice((void *)&beta ,(void *)&beta_p[0],sizeof(ComplexF));
|
||||
RealD t0=usecond();
|
||||
|
||||
#ifdef GRID_HIP
|
||||
hipblasOperation_t hOpA;
|
||||
hipblasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = HIPBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = HIPBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = HIPBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = HIPBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = HIPBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = HIPBLAS_OP_C;
|
||||
auto err = hipblasCgemm(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(hipblasComplex *) &alpha_p[0],
|
||||
(hipblasComplex *) Amk, lda,
|
||||
(hipblasComplex *) Bkn, ldb,
|
||||
(hipblasComplex *) &beta_p[0],
|
||||
(hipblasComplex *) Cmn, ldc);
|
||||
assert(err==HIPBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_CUDA
|
||||
cublasOperation_t hOpA;
|
||||
cublasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = CUBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = CUBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = CUBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = CUBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = CUBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = CUBLAS_OP_C;
|
||||
auto err = cublasCgemm(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(cuComplex *) &alpha_p[0],
|
||||
(cuComplex *) Amk, lda,
|
||||
(cuComplex *) Bkn, ldb,
|
||||
(cuComplex *) &beta_p[0],
|
||||
(cuComplex *) Cmn, ldc);
|
||||
assert(err==CUBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_SYCL
|
||||
int64_t m64=m;
|
||||
int64_t n64=n;
|
||||
int64_t k64=k;
|
||||
int64_t lda64=lda;
|
||||
int64_t ldb64=ldb;
|
||||
int64_t ldc64=ldc;
|
||||
|
||||
oneapi::mkl::transpose iOpA;
|
||||
oneapi::mkl::transpose iOpB;
|
||||
|
||||
if ( OpA == GridBLAS_OP_N ) iOpA = oneapi::mkl::transpose::N;
|
||||
if ( OpA == GridBLAS_OP_T ) iOpA = oneapi::mkl::transpose::T;
|
||||
if ( OpA == GridBLAS_OP_C ) iOpA = oneapi::mkl::transpose::C;
|
||||
if ( OpB == GridBLAS_OP_N ) iOpB = oneapi::mkl::transpose::N;
|
||||
if ( OpB == GridBLAS_OP_T ) iOpB = oneapi::mkl::transpose::T;
|
||||
if ( OpB == GridBLAS_OP_C ) iOpB = oneapi::mkl::transpose::C;
|
||||
|
||||
oneapi::mkl::blas::column_major::gemm(*gridblasHandle,
|
||||
&iOpA,
|
||||
&iOpB,
|
||||
&m64,&n64,&k64,
|
||||
(ComplexF *) &alpha_p[0],
|
||||
(const ComplexF *)Amk, (const int64_t *)&lda64,
|
||||
(const ComplexF *)Bkn, (const int64_t *)&ldb64,
|
||||
(ComplexF *) &beta_p[0],
|
||||
(ComplexF *)Cmn, (const int64_t *)&ldc64);
|
||||
synchronise();
|
||||
#endif
|
||||
#if !defined(GRID_SYCL) && !defined(GRID_CUDA) && !defined(GRID_HIP)
|
||||
// Need a default/reference implementation; use Eigen
|
||||
if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_N) ) {
|
||||
Eigen::Map<Eigen::MatrixXcf> eAmk(Amk,m,k);
|
||||
Eigen::Map<Eigen::MatrixXcf> eBkn(Bkn,k,n);
|
||||
Eigen::Map<Eigen::MatrixXcf> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn ;
|
||||
} else if ( (OpA == GridBLAS_OP_C ) && (OpB == GridBLAS_OP_N) ) {
|
||||
Eigen::Map<Eigen::MatrixXcf> eAmk(Amk,k,m);
|
||||
Eigen::Map<Eigen::MatrixXcf> eBkn(Bkn,k,n);
|
||||
Eigen::Map<Eigen::MatrixXcf> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.adjoint() * eBkn ;
|
||||
} else if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_C) ) {
|
||||
Eigen::Map<Eigen::MatrixXcf> eAmk(Amk,m,k);
|
||||
Eigen::Map<Eigen::MatrixXcf> eBkn(Bkn,n,k);
|
||||
Eigen::Map<Eigen::MatrixXcf> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn.adjoint() ;
|
||||
} else if ( (OpA == GridBLAS_OP_C ) && (OpB == GridBLAS_OP_C) ) {
|
||||
Eigen::Map<Eigen::MatrixXcf> eAmk(Amk,k,m);
|
||||
Eigen::Map<Eigen::MatrixXcf> eBkn(Bkn,n,k);
|
||||
Eigen::Map<Eigen::MatrixXcf> eCmn(Cmn,m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.adjoint() * eBkn.adjoint() ;
|
||||
} else {
|
||||
assert(0);
|
||||
}
|
||||
#endif
|
||||
RealD t1=usecond();
|
||||
RealD flops = 8.0*m*n*k;
|
||||
RealD bytes = 1.0*sizeof(ComplexF)*(m*k+k*n+m*n);
|
||||
}
|
||||
|
||||
|
||||
/////////////////////////////////////////////////////////////
|
||||
void gemmBatched(int m,int n, int k,
|
||||
ComplexD alpha,
|
||||
deviceVector<ComplexD*> &Amk, // pointer list to matrices
|
||||
@ -623,301 +968,6 @@ public:
|
||||
RealD flops = 8.0*m*n*k*batchCount;
|
||||
RealD bytes = 1.0*sizeof(ComplexF)*(m*k+k*n+m*n)*batchCount;
|
||||
}
|
||||
|
||||
///////////////////////////////////////////////////////////////////////////
|
||||
// Single precision real GEMM
|
||||
///////////////////////////////////////////////////////////////////////////
|
||||
|
||||
void gemmBatched(GridBLASOperation_t OpA,
|
||||
GridBLASOperation_t OpB,
|
||||
int m,int n, int k,
|
||||
RealF alpha,
|
||||
deviceVector<RealF*> &Amk, // pointer list to matrices
|
||||
deviceVector<RealF*> &Bkn,
|
||||
RealF beta,
|
||||
deviceVector<RealF*> &Cmn)
|
||||
{
|
||||
RealD t2=usecond();
|
||||
int32_t batchCount = Amk.size();
|
||||
|
||||
assert(OpA!=GridBLAS_OP_C); // Real case no conjugate
|
||||
assert(OpB!=GridBLAS_OP_C);
|
||||
|
||||
int lda = m; // m x k column major
|
||||
int ldb = k; // k x n column major
|
||||
int ldc = m; // m x b column major
|
||||
if(OpA!=GridBLAS_OP_N)
|
||||
lda = k;
|
||||
if(OpB!=GridBLAS_OP_N)
|
||||
ldb = n;
|
||||
static deviceVector<RealF> alpha_p(1);
|
||||
static deviceVector<RealF> beta_p(1);
|
||||
// can prestore the 1 and the zero on device
|
||||
acceleratorCopyToDevice((void *)&alpha,(void *)&alpha_p[0],sizeof(RealF));
|
||||
acceleratorCopyToDevice((void *)&beta ,(void *)&beta_p[0],sizeof(RealF));
|
||||
RealD t0=usecond();
|
||||
|
||||
assert(Bkn.size()==batchCount);
|
||||
assert(Cmn.size()==batchCount);
|
||||
#ifdef GRID_HIP
|
||||
hipblasOperation_t hOpA;
|
||||
hipblasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = HIPBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = HIPBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = HIPBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = HIPBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = HIPBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = HIPBLAS_OP_C;
|
||||
auto err = hipblasSgemmBatched(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(float *) &alpha_p[0],
|
||||
(float **)&Amk[0], lda,
|
||||
(float **)&Bkn[0], ldb,
|
||||
(float *) &beta_p[0],
|
||||
(float **)&Cmn[0], ldc,
|
||||
batchCount);
|
||||
assert(err==HIPBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_CUDA
|
||||
cublasOperation_t hOpA;
|
||||
cublasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = CUBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = CUBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = CUBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = CUBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = CUBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = CUBLAS_OP_C;
|
||||
auto err = cublasSgemmBatched(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(float *) &alpha_p[0],
|
||||
(float **)&Amk[0], lda,
|
||||
(float **)&Bkn[0], ldb,
|
||||
(float *) &beta_p[0],
|
||||
(float **)&Cmn[0], ldc,
|
||||
batchCount);
|
||||
assert(err==CUBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_SYCL
|
||||
int64_t m64=m;
|
||||
int64_t n64=n;
|
||||
int64_t k64=k;
|
||||
int64_t lda64=lda;
|
||||
int64_t ldb64=ldb;
|
||||
int64_t ldc64=ldc;
|
||||
int64_t batchCount64=batchCount;
|
||||
|
||||
oneapi::mkl::transpose iOpA;
|
||||
oneapi::mkl::transpose iOpB;
|
||||
|
||||
if ( OpA == GridBLAS_OP_N ) iOpA = oneapi::mkl::transpose::N;
|
||||
if ( OpA == GridBLAS_OP_T ) iOpA = oneapi::mkl::transpose::T;
|
||||
if ( OpA == GridBLAS_OP_C ) iOpA = oneapi::mkl::transpose::C;
|
||||
if ( OpB == GridBLAS_OP_N ) iOpB = oneapi::mkl::transpose::N;
|
||||
if ( OpB == GridBLAS_OP_T ) iOpB = oneapi::mkl::transpose::T;
|
||||
if ( OpB == GridBLAS_OP_C ) iOpB = oneapi::mkl::transpose::C;
|
||||
|
||||
oneapi::mkl::blas::column_major::gemm_batch(*gridblasHandle,
|
||||
&iOpA,
|
||||
&iOpB,
|
||||
&m64,&n64,&k64,
|
||||
(float *) &alpha_p[0],
|
||||
(const float **)&Amk[0], (const int64_t *)&lda64,
|
||||
(const float **)&Bkn[0], (const int64_t *)&ldb64,
|
||||
(float *) &beta_p[0],
|
||||
(float **)&Cmn[0], (const int64_t *)&ldc64,
|
||||
(int64_t)1,&batchCount64,std::vector<sycl::event>());
|
||||
synchronise();
|
||||
#endif
|
||||
#if !defined(GRID_SYCL) && !defined(GRID_CUDA) && !defined(GRID_HIP)
|
||||
// Need a default/reference implementation; use Eigen
|
||||
if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_N) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXf> eAmk(Amk[p],m,k);
|
||||
Eigen::Map<Eigen::MatrixXf> eBkn(Bkn[p],k,n);
|
||||
Eigen::Map<Eigen::MatrixXf> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn ;
|
||||
});
|
||||
} else if ( (OpA == GridBLAS_OP_T ) && (OpB == GridBLAS_OP_N) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXf> eAmk(Amk[p],k,m);
|
||||
Eigen::Map<Eigen::MatrixXf> eBkn(Bkn[p],k,n);
|
||||
Eigen::Map<Eigen::MatrixXf> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.transpose() * eBkn ;
|
||||
});
|
||||
} else if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_T) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXf> eAmk(Amk[p],m,k);
|
||||
Eigen::Map<Eigen::MatrixXf> eBkn(Bkn[p],n,k);
|
||||
Eigen::Map<Eigen::MatrixXf> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn.transpose() ;
|
||||
});
|
||||
} else if ( (OpA == GridBLAS_OP_T ) && (OpB == GridBLAS_OP_T) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXf> eAmk(Amk[p],k,m);
|
||||
Eigen::Map<Eigen::MatrixXf> eBkn(Bkn[p],n,k);
|
||||
Eigen::Map<Eigen::MatrixXf> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.transpose() * eBkn.transpose() ;
|
||||
} );
|
||||
} else {
|
||||
assert(0);
|
||||
}
|
||||
#endif
|
||||
RealD t1=usecond();
|
||||
RealD flops = 2.0*m*n*k*batchCount;
|
||||
RealD bytes = 1.0*sizeof(RealF)*(m*k+k*n+m*n)*batchCount;
|
||||
}
|
||||
|
||||
|
||||
///////////////////////////////////////////////////////////////////////////
|
||||
// Double precision real GEMM
|
||||
///////////////////////////////////////////////////////////////////////////
|
||||
void gemmBatched(GridBLASOperation_t OpA,
|
||||
GridBLASOperation_t OpB,
|
||||
int m,int n, int k,
|
||||
RealD alpha,
|
||||
deviceVector<RealD*> &Amk, // pointer list to matrices
|
||||
deviceVector<RealD*> &Bkn,
|
||||
RealD beta,
|
||||
deviceVector<RealD*> &Cmn)
|
||||
{
|
||||
RealD t2=usecond();
|
||||
int32_t batchCount = Amk.size();
|
||||
|
||||
assert(OpA!=GridBLAS_OP_C); // Real case no conjugate
|
||||
assert(OpB!=GridBLAS_OP_C);
|
||||
|
||||
int lda = m; // m x k column major
|
||||
int ldb = k; // k x n column major
|
||||
int ldc = m; // m x b column major
|
||||
if(OpA!=GridBLAS_OP_N)
|
||||
lda = k;
|
||||
if(OpB!=GridBLAS_OP_N)
|
||||
ldb = n;
|
||||
|
||||
static deviceVector<RealD> alpha_p(1);
|
||||
static deviceVector<RealD> beta_p(1);
|
||||
// can prestore the 1 and the zero on device
|
||||
acceleratorCopyToDevice((void *)&alpha,(void *)&alpha_p[0],sizeof(RealD));
|
||||
acceleratorCopyToDevice((void *)&beta ,(void *)&beta_p[0],sizeof(RealD));
|
||||
RealD t0=usecond();
|
||||
|
||||
assert(Bkn.size()==batchCount);
|
||||
assert(Cmn.size()==batchCount);
|
||||
#ifdef GRID_HIP
|
||||
hipblasOperation_t hOpA;
|
||||
hipblasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = HIPBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = HIPBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = HIPBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = HIPBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = HIPBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = HIPBLAS_OP_C;
|
||||
auto err = hipblasDgemmBatched(gridblasHandle,
|
||||
HIPBLAS_OP_N,
|
||||
HIPBLAS_OP_N,
|
||||
m,n,k,
|
||||
(double *) &alpha_p[0],
|
||||
(double **)&Amk[0], lda,
|
||||
(double **)&Bkn[0], ldb,
|
||||
(double *) &beta_p[0],
|
||||
(double **)&Cmn[0], ldc,
|
||||
batchCount);
|
||||
assert(err==HIPBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_CUDA
|
||||
cublasOperation_t hOpA;
|
||||
cublasOperation_t hOpB;
|
||||
if ( OpA == GridBLAS_OP_N ) hOpA = CUBLAS_OP_N;
|
||||
if ( OpA == GridBLAS_OP_T ) hOpA = CUBLAS_OP_T;
|
||||
if ( OpA == GridBLAS_OP_C ) hOpA = CUBLAS_OP_C;
|
||||
if ( OpB == GridBLAS_OP_N ) hOpB = CUBLAS_OP_N;
|
||||
if ( OpB == GridBLAS_OP_T ) hOpB = CUBLAS_OP_T;
|
||||
if ( OpB == GridBLAS_OP_C ) hOpB = CUBLAS_OP_C;
|
||||
auto err = cublasDgemmBatched(gridblasHandle,
|
||||
hOpA,
|
||||
hOpB,
|
||||
m,n,k,
|
||||
(double *) &alpha_p[0],
|
||||
(double **)&Amk[0], lda,
|
||||
(double **)&Bkn[0], ldb,
|
||||
(double *) &beta_p[0],
|
||||
(double **)&Cmn[0], ldc,
|
||||
batchCount);
|
||||
assert(err==CUBLAS_STATUS_SUCCESS);
|
||||
#endif
|
||||
#ifdef GRID_SYCL
|
||||
int64_t m64=m;
|
||||
int64_t n64=n;
|
||||
int64_t k64=k;
|
||||
int64_t lda64=lda;
|
||||
int64_t ldb64=ldb;
|
||||
int64_t ldc64=ldc;
|
||||
int64_t batchCount64=batchCount;
|
||||
|
||||
oneapi::mkl::transpose iOpA;
|
||||
oneapi::mkl::transpose iOpB;
|
||||
|
||||
if ( OpA == GridBLAS_OP_N ) iOpA = oneapi::mkl::transpose::N;
|
||||
if ( OpA == GridBLAS_OP_T ) iOpA = oneapi::mkl::transpose::T;
|
||||
if ( OpA == GridBLAS_OP_C ) iOpA = oneapi::mkl::transpose::C;
|
||||
if ( OpB == GridBLAS_OP_N ) iOpB = oneapi::mkl::transpose::N;
|
||||
if ( OpB == GridBLAS_OP_T ) iOpB = oneapi::mkl::transpose::T;
|
||||
if ( OpB == GridBLAS_OP_C ) iOpB = oneapi::mkl::transpose::C;
|
||||
|
||||
oneapi::mkl::blas::column_major::gemm_batch(*gridblasHandle,
|
||||
&iOpA,
|
||||
&iOpB,
|
||||
&m64,&n64,&k64,
|
||||
(double *) &alpha_p[0],
|
||||
(const double **)&Amk[0], (const int64_t *)&lda64,
|
||||
(const double **)&Bkn[0], (const int64_t *)&ldb64,
|
||||
(double *) &beta_p[0],
|
||||
(double **)&Cmn[0], (const int64_t *)&ldc64,
|
||||
(int64_t)1,&batchCount64,std::vector<sycl::event>());
|
||||
synchronise();
|
||||
#endif
|
||||
#if !defined(GRID_SYCL) && !defined(GRID_CUDA) && !defined(GRID_HIP)
|
||||
// Need a default/reference implementation; use Eigen
|
||||
if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_N) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXd> eAmk(Amk[p],m,k);
|
||||
Eigen::Map<Eigen::MatrixXd> eBkn(Bkn[p],k,n);
|
||||
Eigen::Map<Eigen::MatrixXd> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn ;
|
||||
});
|
||||
} else if ( (OpA == GridBLAS_OP_T ) && (OpB == GridBLAS_OP_N) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXd> eAmk(Amk[p],k,m);
|
||||
Eigen::Map<Eigen::MatrixXd> eBkn(Bkn[p],k,n);
|
||||
Eigen::Map<Eigen::MatrixXd> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.transpose() * eBkn ;
|
||||
});
|
||||
} else if ( (OpA == GridBLAS_OP_N ) && (OpB == GridBLAS_OP_T) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXd> eAmk(Amk[p],m,k);
|
||||
Eigen::Map<Eigen::MatrixXd> eBkn(Bkn[p],n,k);
|
||||
Eigen::Map<Eigen::MatrixXd> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk * eBkn.transpose() ;
|
||||
});
|
||||
} else if ( (OpA == GridBLAS_OP_T ) && (OpB == GridBLAS_OP_T) ) {
|
||||
thread_for (p, batchCount, {
|
||||
Eigen::Map<Eigen::MatrixXd> eAmk(Amk[p],k,m);
|
||||
Eigen::Map<Eigen::MatrixXd> eBkn(Bkn[p],n,k);
|
||||
Eigen::Map<Eigen::MatrixXd> eCmn(Cmn[p],m,n);
|
||||
eCmn = beta * eCmn + alpha * eAmk.transpose() * eBkn.transpose() ;
|
||||
});
|
||||
} else {
|
||||
assert(0);
|
||||
}
|
||||
#endif
|
||||
RealD t1=usecond();
|
||||
RealD flops = 2.0*m*n*k*batchCount;
|
||||
RealD bytes = 1.0*sizeof(RealD)*(m*k+k*n+m*n)*batchCount;
|
||||
}
|
||||
|
||||
template<class CComplex>
|
||||
double benchmark(int M, int N, int K, int BATCH)
|
||||
@ -967,6 +1017,47 @@ public:
|
||||
return flops; // Returns gigaflops
|
||||
}
|
||||
|
||||
template<class CComplex>
|
||||
double benchmark(int M, int N, int K)
|
||||
{
|
||||
int32_t N_A = M*K;
|
||||
int32_t N_B = K*N;
|
||||
int32_t N_C = M*N;
|
||||
deviceVector<CComplex> A(N_A); acceleratorMemSet(&A[0],0,N_A*sizeof(CComplex));
|
||||
deviceVector<CComplex> B(N_B); acceleratorMemSet(&B[0],0,N_B*sizeof(CComplex));
|
||||
deviceVector<CComplex> C(N_C); acceleratorMemSet(&C[0],0,N_C*sizeof(CComplex));
|
||||
CComplex alpha(1.0);
|
||||
CComplex beta (1.0);
|
||||
RealD flops = 8.0*M*N*K;
|
||||
int ncall=10;
|
||||
|
||||
gemm(GridBLAS_OP_C,GridBLAS_OP_N,
|
||||
M,N,K,
|
||||
alpha,
|
||||
&A[0], // m x k
|
||||
&B[0], // k x n
|
||||
beta,
|
||||
&C[0]);
|
||||
synchronise();
|
||||
|
||||
RealD t0 = usecond();
|
||||
for(int i=0;i<ncall;i++){
|
||||
gemm(GridBLAS_OP_N,GridBLAS_OP_N,
|
||||
M,N,K,
|
||||
alpha,
|
||||
&A[0], // m x k
|
||||
&B[0], // k x n
|
||||
beta,
|
||||
&C[0]);
|
||||
synchronise();
|
||||
}
|
||||
RealD t1 = usecond();
|
||||
RealD bytes = 1.0*sizeof(CComplex)*(M*N*2+N*K+M*K);
|
||||
flops = 8.0*M*N*K*ncall;
|
||||
flops = flops/(t1-t0)/1.e3;
|
||||
return flops; // Returns gigaflops
|
||||
}
|
||||
|
||||
};
|
||||
|
||||
|
||||
@ -1035,6 +1126,21 @@ static void BLAS(void)
|
||||
std::cout<< M<<"\t\t"<<N<<"\t\t"<<K<<"\t\t"<<BATCH<<"\t\t"<<p<<std::endl;
|
||||
}}
|
||||
fprintf(FP,"\n\n\n");
|
||||
|
||||
std::cout << "----------------------------------------------------------"<<std::endl;
|
||||
std::cout << " M "<<"\t\t"<<"N"<<"\t\t\t"<<"K"<<"\t\t"<<"Gflop/s / rank (inner product matrix)"<<std::endl;
|
||||
std::cout << "----------------------------------------------------------"<<std::endl;
|
||||
{
|
||||
int M=12;
|
||||
int N=12;
|
||||
std::vector<int> ks({4*1024*1024, 2*1024*1024, 1024*1024, 256*1024, 1024 });
|
||||
for( int kk=0;kk<ks.size();kk++ ) {
|
||||
int K = ks[kk];
|
||||
double p=blas.benchmark<CComplex>(M,N,K);
|
||||
fprintf(FP,"%d, %d, %d, %d, %f\n", M, N, K, 1, p);
|
||||
std::cout<< M<<"\t\t"<<N<<"\t\t"<<K<<"\t\t"<<1<<"\t\t"<<p<<std::endl;
|
||||
}
|
||||
}
|
||||
std::cout << "=================================================================================="<<std::endl;
|
||||
};
|
||||
|
||||
|
||||
Loading…
Reference in New Issue
Block a user