mirror of
https://gitlab.com/libeigen/eigen.git
synced 2024-12-21 07:19:46 +08:00
f85038b7f3
Clang does a poor job of optimizing the GEBP microkernel on 32-bit ARM, leading to excessive 16-byte register spills, slowing down basic f32 matrix multiplication by approx 50%. By specializing `gebp_traits`, we can eliminate the register spills. Volatile inline ASM both acts as a barrier to prevent reordering and enforces strict register use. In a simple f32 matrix multiply example, this modification reduces 16-byte spills from 109 instances to zero, leading to a 1.5x speed increase (search for `16-byte Spill` in the assembly in https://godbolt.org/z/chsPbE). This is a replacement of !379. See there for further discussion. Also moved `gebp_traits` specializations for NEON to `Eigen/src/Core/arch/NEON/GeneralBlockPanelKernel.h` to be alongside other NEON-specific code. Fixes #2138. |
||
|---|---|---|
| .. | ||
| src | ||
| Cholesky | ||
| CholmodSupport | ||
| Core | ||
| Dense | ||
| Eigen | ||
| Eigenvalues | ||
| Geometry | ||
| Householder | ||
| IterativeLinearSolvers | ||
| Jacobi | ||
| KLUSupport | ||
| LU | ||
| MetisSupport | ||
| OrderingMethods | ||
| PardisoSupport | ||
| PaStiXSupport | ||
| QR | ||
| QtAlignedMalloc | ||
| Sparse | ||
| SparseCholesky | ||
| SparseCore | ||
| SparseLU | ||
| SparseQR | ||
| SPQRSupport | ||
| StdDeque | ||
| StdList | ||
| StdVector | ||
| SuperLUSupport | ||
| SVD | ||
| UmfPackSupport | ||