| /libCEED/backends/sycl-shared/ |
| H A D | ceed-sycl-shared-basis.sycl.cpp | 86 … const CeedScalar *d_grad_1d = (impl->d_collo_grad_1d) ? impl->d_collo_grad_1d : impl->d_grad_1d; in CeedBasisApplyTensor_Sycl_shared() local 94 cgh.set_args(num_elem, impl->d_interp_1d, d_grad_1d, d_u, d_v); in CeedBasisApplyTensor_Sycl_shared() 149 CeedCallSycl(ceed, sycl::free(impl->d_grad_1d, data->sycl_context)); in CeedBasisDestroy_Sycl_shared() 215 …CeedCallSycl(ceed, impl->d_grad_1d = sycl::malloc_device<CeedScalar>(interp_length, data->sycl_dev… in CeedBasisCreateTensorH1_Sycl_shared() 216 …sycl::event copy_grad = data->sycl_queue.copy<CeedScalar>(grad_1d, impl->d_grad_1d, interp_length,… in CeedBasisCreateTensorH1_Sycl_shared()
|
| H A D | ceed-sycl-shared.hpp | 28 CeedScalar *d_grad_1d; member
|
| /libCEED/backends/hip-shared/ |
| H A D | ceed-hip-shared-basis.c | 168 …CeedCheck(data->d_grad_1d, ceed, CEED_ERROR_BACKEND, "%s not supported; grad_1d not set", CeedEval… in CeedBasisApplyTensorCore_Hip_shared() 172 CeedScalar *d_grad_1d = data->d_grad_1d; in CeedBasisApplyTensorCore_Hip_shared() local 175 d_grad_1d = data->d_collo_grad_1d; in CeedBasisApplyTensorCore_Hip_shared() 177 void *grad_args[] = {(void *)&num_elem, &data->d_interp_1d, &d_grad_1d, &d_u, &d_v}; in CeedBasisApplyTensorCore_Hip_shared() 563 …CeedCheck(data->d_grad_1d, ceed, CEED_ERROR_BACKEND, "%s not supported; grad not set", CeedEvalMod… in CeedBasisApplyNonTensorCore_Hip_shared() 567 void *grad_args[] = {(void *)&num_elem, &data->d_grad_1d, &d_u, &d_v}; in CeedBasisApplyNonTensorCore_Hip_shared() 644 CeedCallHip(ceed, hipFree(data->d_grad_1d)); in CeedBasisDestroy_Hip_shared() 673 CeedCallHip(ceed, hipMalloc((void **)&data->d_grad_1d, interp_bytes)); in CeedBasisCreateTensorH1_Hip_shared() 674 CeedCallHip(ceed, hipMemcpy(data->d_grad_1d, grad_1d, interp_bytes, hipMemcpyHostToDevice)); in CeedBasisCreateTensorH1_Hip_shared() 768 CeedCallHip(ceed, hipMalloc((void **)&data->d_grad_1d, grad_bytes)); in CeedBasisCreateH1_Hip_shared() [all …]
|
| H A D | ceed-hip-shared.h | 32 CeedScalar *d_grad_1d; member
|
| /libCEED/backends/cuda-shared/ |
| H A D | ceed-cuda-shared-basis.c | 103 …CeedCheck(data->d_grad_1d, ceed, CEED_ERROR_BACKEND, "%s not supported; grad_1d not set", CeedEval… in CeedBasisApplyTensorCore_Cuda_shared() 107 CeedScalar *d_grad_1d = data->d_grad_1d; in CeedBasisApplyTensorCore_Cuda_shared() local 110 d_grad_1d = data->d_collo_grad_1d; in CeedBasisApplyTensorCore_Cuda_shared() 112 void *grad_args[] = {(void *)&num_elem, &data->d_interp_1d, &d_grad_1d, &d_u, &d_v}; in CeedBasisApplyTensorCore_Cuda_shared() 500 …CeedCheck(data->d_grad_1d, ceed, CEED_ERROR_BACKEND, "%s not supported; grad not set", CeedEvalMod… in CeedBasisApplyNonTensorCore_Cuda_shared() 505 void *grad_args[] = {(void *)&num_elem, &data->d_grad_1d, &d_u, &d_v}; in CeedBasisApplyNonTensorCore_Cuda_shared() 583 CeedCallCuda(ceed, cudaFree(data->d_grad_1d)); in CeedBasisDestroy_Cuda_shared() 612 CeedCallCuda(ceed, cudaMalloc((void **)&data->d_grad_1d, interp_bytes)); in CeedBasisCreateTensorH1_Cuda_shared() 613 CeedCallCuda(ceed, cudaMemcpy(data->d_grad_1d, grad_1d, interp_bytes, cudaMemcpyHostToDevice)); in CeedBasisCreateTensorH1_Cuda_shared() 702 CeedCallCuda(ceed, cudaMalloc((void **)&data->d_grad_1d, grad_bytes)); in CeedBasisCreateH1_Cuda_shared() [all …]
|
| H A D | ceed-cuda-shared.h | 31 CeedScalar *d_grad_1d; member
|
| /libCEED/include/ceed/jit-source/sycl/ |
| H A D | sycl-shared-basis-tensor.h | 90 …m_elem, global const CeedScalar *restrict d_interp_1d, global const CeedScalar *restrict d_grad_1d, in Grad() argument 104 loadMatrix(BASIS_Q_1D * (BASIS_HAS_COLLOCATED_GRAD ? BASIS_Q_1D : BASIS_P_1D), d_grad_1d, s_G); in Grad() 125 …m_elem, global const CeedScalar *restrict d_interp_1d, global const CeedScalar *restrict d_grad_1d, in GradTranspose() argument 139 loadMatrix(BASIS_Q_1D * (BASIS_HAS_COLLOCATED_GRAD ? BASIS_Q_1D : BASIS_P_1D), d_grad_1d, s_G); in GradTranspose()
|
| /libCEED/backends/magma/ |
| H A D | ceed-magma.h | 56 CeedScalar *d_grad_1d; member
|
| H A D | ceed-magma-basis.c | 194 …void *args[] = {&impl->d_interp_1d, &impl->d_grad_1d, &d_u, &u_elem_stride, &u_comp_str… in CeedBasisApplyCore_Magma() 493 CeedCallBackend(magma_free(impl->d_grad_1d)); in CeedBasisDestroy_Magma() 553 CeedCallBackend(magma_malloc((void **)&impl->d_grad_1d, Q_1d * P_1d * sizeof(grad_1d[0]))); in CeedBasisCreateTensorH1_Magma() 554 magma_setvector(Q_1d * P_1d, sizeof(grad_1d[0]), grad_1d, 1, impl->d_grad_1d, 1, data->queue); in CeedBasisCreateTensorH1_Magma()
|
| /libCEED/backends/sycl-ref/ |
| H A D | ceed-sycl-ref.hpp | 57 CeedScalar *d_grad_1d; member
|
| H A D | ceed-sycl-ref-basis.sycl.cpp | 147 const CeedScalar *grad_1d = impl->d_grad_1d; in CeedBasisApplyGrad_Sycl() 528 CeedCallSycl(ceed, sycl::free(impl->d_grad_1d, data->sycl_context)); in CeedBasisDestroy_Sycl() 602 …CeedCallSycl(ceed, impl->d_grad_1d = sycl::malloc_device<CeedScalar>(interp_length, data->sycl_dev… in CeedBasisCreateTensorH1_Sycl() 603 …sycl::event copy_grad = data->sycl_queue.copy<CeedScalar>(grad_1d, impl->d_grad_1d, interp_length,… in CeedBasisCreateTensorH1_Sycl()
|
| /libCEED/backends/hip-ref/ |
| H A D | ceed-hip-ref-basis.c | 57 …s[] = {(void *)&num_elem, (void *)&is_transpose, &data->d_interp_1d, &data->d_grad_1d, &d_u, &d_v}; in CeedBasisApplyCore_Hip() 368 CeedCallHip(ceed, hipFree(data->d_grad_1d)); in CeedBasisDestroy_Hip() 416 CeedCallHip(ceed, hipMalloc((void **)&data->d_grad_1d, interp_bytes)); in CeedBasisCreateTensorH1_Hip() 417 CeedCallHip(ceed, hipMemcpy(data->d_grad_1d, grad_1d, interp_bytes, hipMemcpyHostToDevice)); in CeedBasisCreateTensorH1_Hip()
|
| H A D | ceed-hip-ref.h | 82 CeedScalar *d_grad_1d; member
|
| /libCEED/backends/cuda-ref/ |
| H A D | ceed-cuda-ref-basis.c | 57 …s[] = {(void *)&num_elem, (void *)&is_transpose, &data->d_interp_1d, &data->d_grad_1d, &d_u, &d_v}; in CeedBasisApplyCore_Cuda() 369 CeedCallCuda(ceed, cudaFree(data->d_grad_1d)); in CeedBasisDestroy_Cuda() 417 CeedCallCuda(ceed, cudaMalloc((void **)&data->d_grad_1d, interp_bytes)); in CeedBasisCreateTensorH1_Cuda() 418 CeedCallCuda(ceed, cudaMemcpy(data->d_grad_1d, grad_1d, interp_bytes, cudaMemcpyHostToDevice)); in CeedBasisCreateTensorH1_Cuda()
|
| H A D | ceed-cuda-ref.h | 77 CeedScalar *d_grad_1d; member
|
| /libCEED/backends/sycl-gen/ |
| H A D | ceed-sycl-gen-operator-build.sycl.cpp | 315 … h_G.inputs[i] = has_collo_grad ? basis_impl->d_collo_grad_1d : basis_impl->d_grad_1d; in CeedOperatorBuildKernel_Sycl_gen() 373 … h_G.outputs[i] = has_collo_grad ? basis_impl->d_collo_grad_1d : basis_impl->d_grad_1d; in CeedOperatorBuildKernel_Sycl_gen()
|
| /libCEED/backends/hip-gen/ |
| H A D | ceed-hip-gen-operator-build.cpp | 366 …is_input) data->G.inputs[i] = has_collo_grad ? basis_data->d_collo_grad_1d : basis_data->d_grad_1d; in CeedOperatorBuildKernelFieldData_Hip_gen() 367 … else data->G.outputs[i] = has_collo_grad ? basis_data->d_collo_grad_1d : basis_data->d_grad_1d; in CeedOperatorBuildKernelFieldData_Hip_gen()
|
| /libCEED/backends/cuda-gen/ |
| H A D | ceed-cuda-gen-operator-build.cpp | 339 …is_input) data->G.inputs[i] = has_collo_grad ? basis_data->d_collo_grad_1d : basis_data->d_grad_1d; in CeedOperatorBuildKernelFieldData_Cuda_gen() 340 … else data->G.outputs[i] = has_collo_grad ? basis_data->d_collo_grad_1d : basis_data->d_grad_1d; in CeedOperatorBuildKernelFieldData_Cuda_gen()
|