GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-08-15 21:26:00 +08:00
Code Issues Packages Projects Releases Wiki Activity

Revert b55b5c7280

Browse Source 
.
This commit is contained in:
Rasmus Munk Larsen 2019-02-14 13:46:13 -08:00
parent efeabee445
commit 65e23ca7e9
1 changed files with 5 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
Eigen/src/Core/products/GeneralMatrixVector.h
View File

@ -255,20 +255,11 @@ EIGEN_DEVICE_FUNC EIGEN_DONT_INLINE void general_matrix_vector_product<Index,Lhs
conj_helper<LhsScalar,RhsScalar,ConjugateLhs,ConjugateRhs> cj; conj_helper<LhsScalar,RhsScalar,ConjugateLhs,ConjugateRhs> cj;
conj_helper<LhsPacket,RhsPacket,ConjugateLhs,ConjugateRhs> pcj; conj_helper<LhsPacket,RhsPacket,ConjugateLhs,ConjugateRhs> pcj;
// TODO: fine tune the following heuristic. The rationale is that if the // TODO: fine tune the following heuristic. The rationale is that if the matrix is very large,
// matrix is very large, processing multiple rows at once might be counter // processing 8 rows at once might be counter productive wrt cache.
// productive wrt cache. const Index n8 = lhs.stride()*sizeof(LhsScalar)>32000 ? 0 : rows-7;
#if EIGEN_ARCH_ARM_OR_ARM64 const Index n4 = rows-3;
// This threshold was empirically determined using a Pixel2. const Index n2 = rows-1;
// The little cores are a lot more sensitive to this number
// than the big cores.
const Index cache_threshold = 1024;
#else
const Index cache_threshold = 1024 * 256;
#endif
const Index row_bytes = lhs.stride() * sizeof(LhsScalar);
const Index n8 = (8 * row_bytes > cache_threshold) ? 0 : (rows - 7);
// TODO: for padded aligned inputs, we could enable aligned reads // TODO: for padded aligned inputs, we could enable aligned reads
enum { LhsAlignment = Unaligned }; enum { LhsAlignment = Unaligned };
@ -329,9 +320,6 @@ EIGEN_DEVICE_FUNC EIGEN_DONT_INLINE void general_matrix_vector_product<Index,Lhs
res[(i+6)*resIncr] += alpha*cc6; res[(i+6)*resIncr] += alpha*cc6;
res[(i+7)*resIncr] += alpha*cc7; res[(i+7)*resIncr] += alpha*cc7;
} }
if (i == rows) return;
const Index n4 = (4 * row_bytes > cache_threshold) ? 0 : (rows - 3);
for(; i<n4; i+=4) for(; i<n4; i+=4)
{ {
ResPacket c0 = pset1<ResPacket>(ResScalar(0)), ResPacket c0 = pset1<ResPacket>(ResScalar(0)),
@ -367,9 +355,6 @@ EIGEN_DEVICE_FUNC EIGEN_DONT_INLINE void general_matrix_vector_product<Index,Lhs
res[(i+2)*resIncr] += alpha*cc2; res[(i+2)*resIncr] += alpha*cc2;
res[(i+3)*resIncr] += alpha*cc3; res[(i+3)*resIncr] += alpha*cc3;
} }
if (i == rows) return;
const Index n2 = (2 * row_bytes > cache_threshold) ? 0 : (rows - 1);
for(; i<n2; i+=2) for(; i<n2; i+=2)
{ {
ResPacket c0 = pset1<ResPacket>(ResScalar(0)), ResPacket c0 = pset1<ResPacket>(ResScalar(0)),