/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/NVPTX/ |
D | sqrt-approx.ll | 13 ; CHECK: rsqrt.approx.f32 21 ; CHECK: rsqrt.approx.ftz.f32 29 ; CHECK: rsqrt.approx.f64 37 ; There's no rsqrt.approx.ftz.f64 instruction; we just use the non-ftz version. 38 ; CHECK: rsqrt.approx.f64 63 ; reciprocal(rsqrt.approx.f64(x)). There's no non-ftz approximate reciprocal, 65 ; CHECK: rsqrt.approx.f64 74 ; CHECK: rsqrt.approx.f64 80 ; -- refined sqrt and rsqrt -- 82 ; The sqrt and rsqrt refinement algorithms both emit an rsqrt.approx, followed [all …]
|
D | fast-math.ll | 43 ; reciprocal(rsqrt(x)) for sqrt(x), and emit a vanilla divide. 46 ; CHECK: rsqrt.approx.f64 55 ; CHECK-LABEL: rsqrt( 56 ; CHECK-NOT: rsqrt.approx 58 ; CHECK-NOT: rsqrt.approx 59 define float @rsqrt(float %a) { 68 ; CHECK: rsqrt.approx.f32 80 ; CHECK: rsqrt.approx.ftz.f32
|
/external/tensorflow/tensorflow/core/kernels/ |
D | cwise_op_rsqrt.cc | 19 REGISTER5(UnaryOp, CPU, "Rsqrt", functor::rsqrt, float, Eigen::half, double, 23 REGISTER3(UnaryOp, GPU, "Rsqrt", functor::rsqrt, float, Eigen::half, double); 26 REGISTER2(UnaryOp, SYCL, "Rsqrt", functor::rsqrt, float, double);
|
D | batch_norm_op.h | 55 ((var + var.constant(variance_epsilon)).rsqrt() * gamma) in operator() 64 ((var + var.constant(variance_epsilon)).rsqrt()) in operator() 117 scratch1.device(d) = (var + var.constant(variance_epsilon)).rsqrt(); in operator()
|
D | cwise_op_gpu_rsqrt.cu.cc | 23 DEFINE_UNARY3(rsqrt, Eigen::half, float, double);
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/X86/ |
D | extractps.ll | 12 … call <4 x float> @llvm.x86.sse.rsqrt.ss( <4 x float> %2 ) ; <<4 x float>>:3 [#uses=1] 20 … call <4 x float> @llvm.x86.sse.rsqrt.ss( <4 x float> %2 ) ; <<4 x float>>:3 [#uses=1] 26 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
|
D | fold-load-unops.ll | 41 %res = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %ins) 128 %res = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %ins) 144 %res = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %ld) 217 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
|
D | sse-scalar-fp-arith-unary.ll | 40 %y = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %x) 76 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>)
|
/external/llvm/test/CodeGen/X86/ |
D | extractps.ll | 12 … call <4 x float> @llvm.x86.sse.rsqrt.ss( <4 x float> %2 ) ; <<4 x float>>:3 [#uses=1] 20 … call <4 x float> @llvm.x86.sse.rsqrt.ss( <4 x float> %2 ) ; <<4 x float>>:3 [#uses=1] 26 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
|
D | sse-scalar-fp-arith-unary.ll | 35 %y = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %x) 71 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>)
|
D | fold-load-unops.ll | 40 %res = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %ins) 112 %res = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %ins) 152 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
|
/external/swiftshader/third_party/LLVM/test/CodeGen/X86/ |
D | extractps.ll | 12 … call <4 x float> @llvm.x86.sse.rsqrt.ss( <4 x float> %2 ) ; <<4 x float>>:3 [#uses=1] 20 … call <4 x float> @llvm.x86.sse.rsqrt.ss( <4 x float> %2 ) ; <<4 x float>>:3 [#uses=1] 26 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
|
/external/tensorflow/tensorflow/core/api_def/python_api/ |
D | api_def_Rsqrt.pbtxt | 4 name: "math.rsqrt" 7 name: "rsqrt"
|
/external/swiftshader/third_party/llvm-7.0/llvm/test/Transforms/Reassociate/ |
D | factorize-again.ll | 9 ; CHECK-NEXT: [[TMP4:%.*]] = call float @llvm.rsqrt.f32(float undef) 22 %4 = call float @llvm.rsqrt.f32(float undef) 41 declare float @llvm.rsqrt.f32(float) #1
|
/external/tensorflow/tensorflow/core/api_def/base_api/ |
D | api_def_RsqrtGrad.pbtxt | 4 summary: "Computes the gradient for the rsqrt of `x` wrt its input." 6 Specifically, `grad = dy * -0.5 * y^3`, where `y = rsqrt(x)`, and `dy`
|
/external/tensorflow/tensorflow/python/compiler/tensorrt/test/ |
D | unary_test.py | 55 q = gen_math_ops.rsqrt(q) 70 q = gen_math_ops.rsqrt(q) 91 q = gen_math_ops.rsqrt(q)
|
/external/llvm/test/Transforms/Reassociate/ |
D | factorize-again.ll | 11 %4 = call float @llvm.rsqrt.f32(float undef) 30 declare float @llvm.rsqrt.f32(float) #1
|
/external/skia/tests/ |
D | SkVxTest.cpp | 36 float4 RSqrt(float4 x) { return rsqrt(x); } in RSqrt() 109 REPORTER_ASSERT(r, all(rsqrt(float4{2,3,4,5}) < float4{1.0f,1.0f,1.0f,0.5f})); in DEF_TEST() 113 REPORTER_ASSERT(r, all(rsqrt(float2{2,3}) < float2{1.0f,1.0f})); in DEF_TEST()
|
/external/eigen/Eigen/src/Core/arch/CUDA/ |
D | MathFunctions.h | 81 return make_double2(rsqrt(a.x), rsqrt(a.y));
|
/external/deqp-deps/glslang/Test/ |
D | hlsl.intrinsics.vert | 45 rsqrt(inF0); 119 rsqrt(inF0); 189 rsqrt(inF0); 259 rsqrt(inF0); 314 rsqrt(inF0); \
|
/external/tensorflow/tensorflow/contrib/labeled_tensor/ |
D | __init__.py | 59 rsqrt = _core.rsqrt variable
|
/external/llvm/test/CodeGen/NVPTX/ |
D | rsqrt.ll | 8 ; CHECK: rsqrt.approx.f32
|
/external/tensorflow/tensorflow/python/ops/signal/ |
D | mfcc_ops.py | 110 return dct2 * math_ops.rsqrt(
|
/external/llvm/test/MC/Disassembler/Mips/mips4/ |
D | valid-xfail-mips4.txt | 41 0x46 0x20 0xe0 0xd6 # CHECK: rsqrt.d $f3, $f28 42 0x46 0x00 0x41 0x16 # CHECK: rsqrt.s $f4, $f8
|
/external/skqp/tests/ |
D | MathTest.cpp | 257 static void test_rsqrt(skiatest::Reporter* reporter, RSqrtFn rsqrt) { in test_rsqrt() argument 264 float estimate = rsqrt(input); in test_rsqrt() 274 float estimate = rsqrt(input); in test_rsqrt() 284 float estimate = rsqrt(input); in test_rsqrt()
|