| /external/eigen/unsupported/Eigen/CXX11/src/Tensor/ |
| D | TensorEvaluator.h | 106 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const { in costPerCoeff()
107 return TensorOpCost(sizeof(CoeffReturnType), 0, 0, vectorized, in costPerCoeff()
205 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
206 return TensorOpCost(sizeof(CoeffReturnType), 0, 0, vectorized,
269 costPerCoeff(bool vectorized) const {
270 return TensorOpCost(sizeof(CoeffReturnType), 0, 0, vectorized,
338 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
340 return m_argImpl.costPerCoeff(vectorized) +
341 TensorOpCost(0, 0, functor_cost, vectorized, PacketSize);
417 costPerCoeff(bool vectorized) const {
[all …]
|
| D | TensorAssign.h | 154 costPerCoeff(bool vectorized) const {
158 TensorOpCost left = m_leftImpl.costPerCoeff(vectorized);
159 return m_rightImpl.costPerCoeff(vectorized) +
163 TensorOpCost(0, sizeof(CoeffReturnType), 0, vectorized, PacketSize);
|
| D | TensorArgMax.h | 116 costPerCoeff(bool vectorized) const {
117 return m_impl.costPerCoeff(vectorized) + TensorOpCost(0, 0, 1);
258 costPerCoeff(bool vectorized) const {
261 return m_orig_impl.costPerCoeff(vectorized) +
262 m_impl.costPerCoeff(vectorized) + TensorOpCost(0, 0, compute_cost);
|
| D | TensorEvalTo.h | 155 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
158 return m_impl.costPerCoeff(vectorized) +
159 TensorOpCost(0, sizeof(CoeffReturnType), 0, vectorized, PacketSize);
|
| D | TensorConversion.h | 233 costPerCoeff(bool vectorized) const {
235 if (vectorized) {
240 return m_impl.costPerCoeff(vectorized) * (SrcCoeffRatio / PacketSize) +
243 return m_impl.costPerCoeff(vectorized) + TensorOpCost(0, 0, cast_cost);
|
| D | TensorCustomOp.h | 136 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
138 return TensorOpCost(sizeof(CoeffReturnType), 0, 0, vectorized, PacketSize);
291 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
293 return TensorOpCost(sizeof(CoeffReturnType), 0, 0, vectorized, PacketSize);
|
| D | TensorInflation.h | 203 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
211 return m_impl.costPerCoeff(vectorized) +
213 compute_cost, vectorized, PacketSize);
|
| D | TensorStriding.h | 211 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
216 if (vectorized) {
220 return m_impl.costPerCoeff(vectorized && m_inputStrides[innerDim] == 1) +
222 TensorOpCost(0, 0, compute_cost, vectorized, PacketSize);
|
| D | TensorCostModel.h | 64 bool vectorized, double packet_size) in TensorOpCost() argument
67 compute_cycles_(vectorized ? compute_cycles / packet_size in TensorOpCost()
|
| D | TensorPatch.h | 248 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
252 return m_impl.costPerCoeff(vectorized) +
253 TensorOpCost(0, 0, compute_cost, vectorized, PacketSize);
|
| D | TensorForcedEval.h | 147 EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE TensorOpCost costPerCoeff(bool vectorized) const {
148 return TensorOpCost(sizeof(CoeffReturnType), 0, 0, vectorized, PacketSize);
|
| D | TensorConvolution.h | 452 costPerCoeff(bool vectorized) const {
461 return TensorOpCost(0, 0, firstIndex_compute_cost, vectorized, PacketSize) +
462 kernel_size * (m_inputImpl.costPerCoeff(vectorized) +
463 m_kernelImpl.costPerCoeff(vectorized) +
464 TensorOpCost(0, 0, convolve_compute_cost, vectorized,
1066 costPerCoeff(bool vectorized) const {
1077 return TensorOpCost(0, 0, firstIndex_compute_cost, vectorized, PacketSize) +
1078 kernel_size * (m_inputImpl.costPerCoeff(vectorized) +
1079 m_kernelImpl.costPerCoeff(vectorized) +
1080 TensorOpCost(0, 0, convolve_compute_cost, vectorized,
|
| /external/llvm/test/Transforms/LoopVectorize/ |
| D | no_switch.ll | 5 ; CHECK: remark: source.cpp:4:5: loop not vectorized: loop contains a switch statement
6 ; CHECK: warning: source.cpp:4:5: loop not vectorized: failed explicitly specified loop vectorizati…
11 ; MOREINFO: remark: source.cpp:4:5: loop not vectorized: loop contains a switch statement
12 ; MOREINFO: remark: source.cpp:4:5: loop not vectorized: use -Rpass-analysis=loop-vectorize for mor…
13 ; MOREINFO: warning: source.cpp:4:5: loop not vectorized: failed explicitly specified loop vectoriz…
|
| D | runtime-limit.ll | 7 ;CHECK: remark: {{.*}}:0:0: vectorized loop (vectorization width: 4, interleaved count: 1)
9 ;CHECK: remark: {{.*}}:0:0: loop not vectorized: cannot prove it is safe to reorder memory operatio…
12 ;OVERRIDE: remark: {{.*}}:0:0: vectorized loop (vectorization width: 4, interleaved count: 1)
14 ;OVERRIDE: remark: {{.*}}:0:0: vectorized loop (vectorization width: 4, interleaved count: 1)
|
| D | interleaved-accesses-pred-stores.ll | 8 ; vectorized.
56 ; groups are separately vectorized. The store group contains gaps and is not
57 ; vectorized.
115 ; predicated block. The load group is vectorized, and the store groups contain
116 ; gaps and are not vectorized.
|
| D | vect.stats.ll | 9 ; CHECK: 1 loop-vectorize - Number of loops vectorized
14 define void @vectorized(float* nocapture %a, i64 %size) {
|
| D | memdep.ll | 8 ; No plausible dependence - can be vectorized.
34 ; Plausible dependence of distance 1 - can't be vectorized.
60 ; Plausible dependence of distance 2 - can be vectorized with a width of 2.
92 ; Plausible dependence of distance 1 - cannot be vectorized (without reordering
|
| D | control-flow.ll | 13 ; CHECK: remark: source.cpp:5:9: loop not vectorized: loop control flow is not understood by vector…
14 ; CHECK: remark: source.cpp:5:9: loop not vectorized: use -Rpass-analysis=loop-vectorize for more i…
|
| /external/llvm/test/Transforms/LoopVectorize/X86/ |
| D | vect.omp.force.small-tc.ll | 9 ; CHECK: 1 loop-vectorize - Number of loops vectorized
24 ; This loop will be vectorized, although the trip count is below the threshold, but vectorization i…
26 define void @vectorized(float* noalias nocapture %A, float* noalias nocapture readonly %B) {
50 ; This loop will not be vectorized as the trip count is below the threshold.
|
| D | vect.omp.force.ll | 9 ; CHECK: 1 loop-vectorize - Number of loops vectorized
25 ; This loop will be vectorized, although the scalar cost is lower than any of vector costs, but vec…
28 define void @vectorized(float* noalias nocapture %A, float* noalias nocapture %B) {
55 ; This method will not be vectorized, as scalar cost is lower than any of vector costs.
|
| D | already-vectorized.ll | 4 ; The vectorizer used to mark the un-vectorized loop only as already vectorized
5 ; thus, trying to vectorize the vectorized loop again
|
| D | vectorization-remarks-missed.ll | 26 ; CHECK: remark: source.cpp:4:5: loop not vectorized: could not determine number of loop iterations
27 ; CHECK: remark: source.cpp:4:5: loop not vectorized: use -Rpass-analysis=loop-vectorize for more i…
28 ; CHECK: remark: source.cpp:13:5: loop not vectorized: vectorization and interleaving are explicitl…
29 ; CHECK: remark: source.cpp:19:5: loop not vectorized: cannot identify array bounds
30 ; CHECK: remark: source.cpp:19:5: loop not vectorized: use -Rpass-analysis=loop-vectorize for more …
31 ; CHECK: warning: source.cpp:19:5: loop not vectorized: failed explicitly specified loop vectorizat…
|
| D | no_fpmath.ll | 3 ; CHECK: remark: no_fpmath.c:6:11: loop not vectorized: cannot prove it is safe to reorder floating…
4 ; CHECK: remark: no_fpmath.c:6:14: loop not vectorized:
5 ; CHECK: remark: no_fpmath.c:17:14: vectorized loop (vectorization width: 2, interleaved count: 2)
|
| /external/llvm/docs/ |
| D | Vectorizers.rst | 89 Many loops cannot be vectorized including loops with complicated control flow,
96 ``-Rpass=loop-vectorize`` identifies loops that were successfully vectorized.
117 The command line ``-Rpass-missed=loop-vectorized`` prints the remark:
121 …no_switch.cpp:4:5: remark: loop not vectorized: vectorization is explicitly enabled [-Rpass-missed…
124 switch statement cannot be vectorized.
128 …no_switch.cpp:4:5: remark: loop not vectorized: loop contains a switch statement [-Rpass-analysis=…
293 Access to global structures can also be vectorized, with alias analysis being
298 ignored (as other compilers do) are still being left un-vectorized.
331 example, the loop below will be vectorized on Intel x86 if the SSE4.1 roundps
394 operations, PHI-nodes, can all be vectorized using this technique.
|
| /external/eigen/doc/ |
| D | UnalignedArrayAssert.dox | 116 …vectorized and non-vectorized code are not mutually ABI-compatible. This ABI compatibility is very…
|