/external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
D | nvptx_compiler.h | 101 const string& ptx, int cc_major, int cc_minor, 117 CompilationCacheKey(std::string ptx, int cc_major, int cc_minor) in CompilationCacheKey() 118 : ptx(std::move(ptx)), cc_major(cc_major), cc_minor(cc_minor) {} in CompilationCacheKey() 121 int cc_minor; member 127 key.cc_minor); in operator() 133 return a.cc_major == b.cc_major && a.cc_minor == b.cc_minor && in operator()
|
D | nvptx_compiler.cc | 540 const string& ptx, int cc_major, int cc_minor, in CompilePtx() argument 582 absl::StrCat("-arch=sm_", cc_major, cc_minor)}; in CompilePtx() 737 int cc_major, cc_minor; in RunBackend() local 739 &cc_minor)) { in RunBackend() 743 cc_minor = 0; in RunBackend() 749 TF_ASSIGN_OR_RETURN(ptx, CompileToPtx(&llvm_module, {cc_major, cc_minor}, in RunBackend() 764 CompilePtxOrGetCachedResult(ptx, cc_major, cc_minor, module->config()); in RunBackend() 793 ptx, cubin, {cc_major, cc_minor}, std::move(thunk_schedule), in RunBackend() 804 const string& ptx, int cc_major, int cc_minor, in CompilePtxOrGetCachedResult() argument 819 std::forward_as_tuple(ptx, cc_major, cc_minor), in CompilePtxOrGetCachedResult() [all …]
|
D | cudnn_conv_algorithm_picker.cc | 117 int cc_major, cc_minor; in GetComputeCapability() local 119 &cc_minor); in GetComputeCapability() 121 cc.set_minor(cc_minor); in GetComputeCapability()
|
/external/tensorflow/tensorflow/core/common_runtime/gpu/ |
D | gpu_device_test.cc | 45 int* cc_minor) { in GetComputeCapability() argument 49 if (!se->GetDeviceDescription().cuda_compute_capability(cc_major, cc_minor)) { in GetComputeCapability() 51 *cc_minor = 0; in GetComputeCapability() 225 int cc_major, cc_minor; in TEST_F() local 226 TF_ASSERT_OK(GetComputeCapability(PlatformGpuId(0), &cc_major, &cc_minor)); in TEST_F() 249 int cc_major, cc_minor; in TEST_F() local 250 TF_ASSERT_OK(GetComputeCapability(kPlatformGpuId, &cc_major, &cc_minor)); in TEST_F()
|
D | gpu_device.cc | 904 int cc_major = 0, cc_minor = 0; in SingleVirtualDeviceMemoryLimit() local 906 &cc_minor)) { in SingleVirtualDeviceMemoryLimit() 1165 int cc_minor; in GetShortDeviceDescription() local 1166 if (!desc.cuda_compute_capability(&cc_major, &cc_minor)) { in GetShortDeviceDescription() 1168 cc_minor = 0; in GetShortDeviceDescription() 1173 ", compute capability: ", cc_major, ".", cc_minor); in GetShortDeviceDescription() 1545 int cc_minor; in GetValidDeviceIds() local 1546 if (!description.cuda_compute_capability(&cc_major, &cc_minor)) { in GetValidDeviceIds() 1549 cc_minor = 0; in GetValidDeviceIds() 1553 << " minor: " << cc_minor in GetValidDeviceIds()
|
/external/tensorflow/tensorflow/stream_executor/ |
D | stream_executor_pimpl.cc | 330 int cc_major, cc_minor; in GetConvolveAlgorithms() local 331 GetDeviceDescription().cuda_compute_capability(&cc_major, &cc_minor); in GetConvolveAlgorithms() 333 cc_minor, out_algorithms); in GetConvolveAlgorithms() 352 int cc_major, cc_minor; in GetConvolveBackwardDataAlgorithms() local 353 GetDeviceDescription().cuda_compute_capability(&cc_major, &cc_minor); in GetConvolveBackwardDataAlgorithms() 355 with_winograd_nonfused, cc_major, cc_minor, out_algorithms); in GetConvolveBackwardDataAlgorithms() 365 int cc_major, cc_minor; in GetConvolveBackwardFilterAlgorithms() local 366 GetDeviceDescription().cuda_compute_capability(&cc_major, &cc_minor); in GetConvolveBackwardFilterAlgorithms() 368 with_winograd_nonfused, cc_major, cc_minor, out_algorithms); in GetConvolveBackwardFilterAlgorithms()
|
D | dnn.cc | 30 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveAlgorithms() argument 40 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveBackwardDataAlgorithms() argument 46 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveBackwardFilterAlgorithms() argument
|
D | dnn.h | 1272 bool with_winograd_nonfused, int cc_major, int cc_minor, 1365 bool with_winograd_nonfused, int cc_major, int cc_minor, 1412 bool with_winograd_nonfused, int cc_major, int cc_minor,
|
/external/tensorflow/tensorflow/core/kernels/ |
D | gpu_utils.cc | 48 int cc_major, cc_minor; in GetComputeCapability() local 50 &cc_minor); in GetComputeCapability() 52 cc.set_minor(cc_minor); in GetComputeCapability()
|
/external/tensorflow/tensorflow/contrib/fused_conv/kernels/ |
D | fused_conv2d_bias_activation_op.cc | 304 int cc_major, cc_minor; in GetComputeCapability() local 306 &cc_minor); in GetComputeCapability() 308 cc.set_minor(cc_minor); in GetComputeCapability() 460 int cc_major, cc_minor; in launch() local 462 &cc_minor); in launch() 464 ctx, ((cc_major == 6 && cc_minor >= 1) || cc_major > 6), in launch()
|
/external/tensorflow/tensorflow/stream_executor/cuda/ |
D | cuda_dnn.h | 203 bool with_winograd_nonfused, int cc_major, int cc_minor, 210 bool with_winograd_nonfused, int cc_major, int cc_minor, 214 bool with_winograd_nonfused, int cc_major, int cc_minor,
|
D | cuda_blas.cc | 1598 int cc_major, cc_minor; in DoBlasGemm() local 1600 &cc_minor); in DoBlasGemm() 1895 int cc_major, cc_minor; in DoBlasGemmWithAlgorithmImpl() local 1897 &cc_major, &cc_minor) && in DoBlasGemmWithAlgorithmImpl() 1900 << cc_minor << " devices don't support explicit gemm algorithms."; in DoBlasGemmWithAlgorithmImpl() 2213 int cc_major, cc_minor; in DoBlasGemmBatchedInternal() local 2215 &cc_major, &cc_minor) && in DoBlasGemmBatchedInternal() 2366 int cc_major, cc_minor; in DoBlasGemmStridedBatched() local 2368 &cc_major, &cc_minor)) { in DoBlasGemmStridedBatched()
|
D | cuda_driver.cc | 1351 int* cc_minor, in GetComputeCapability() argument 1354 *cc_minor = 0; in GetComputeCapability() 1367 cc_minor, CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, device); in GetComputeCapability()
|
D | cuda_dnn.cc | 3043 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveAlgorithms() argument 3104 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveBackwardDataAlgorithms() argument 3139 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveBackwardFilterAlgorithms() argument 3429 int cc_major, cc_minor; in DoFusedConvolve() local 3431 &cc_minor); in DoFusedConvolve() 3432 if (cc_major < 6 || (cc_major == 6 && cc_minor < 1)) { in DoFusedConvolve()
|
/external/tensorflow/tensorflow/stream_executor/rocm/ |
D | rocm_dnn.h | 194 bool with_winograd_nonfused, int cc_major, int cc_minor, 201 bool with_winograd_nonfused, int cc_major, int cc_minor, 205 bool with_winograd_nonfused, int cc_major, int cc_minor,
|
D | rocm_driver.cc | 1056 int* cc_minor, in GetComputeCapability() argument
|
D | rocm_dnn.cc | 2949 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveAlgorithms() argument 2970 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveBackwardDataAlgorithms() argument 2985 bool with_winograd_nonfused, int cc_major, int cc_minor, in GetConvolveBackwardFilterAlgorithms() argument
|
/external/tensorflow/tensorflow/stream_executor/gpu/ |
D | gpu_driver.h | 398 static port::Status GetComputeCapability(int* cc_major, int* cc_minor,
|