devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add a generic Eigen backend based on clang vector extensions

Browse Source 
The goal of this MR is to implement a generic SIMD backend (packet ops) for Eigen that uses clang vector extensions instead of platform-dependent intrinsics. Ideally, this should make it possible to build Eigen and achieve reasonable speed on any platform that has a recent clang compiler, without having to write any inline assembly or intrinsics.

Caveats:

* The current implementation is a proof of concept and supports vectorization for float, double, int32_t, and int64_t using fixed-size 512-bit vectors (a somewhat arbitrary choice). I have not done much to tune this for speed yet.
* For now, there is no way to enable this other than setting -DEIGEN_VECTORIZE_GENERIC on the command line.
* This only compiles with newer versions of clang. I have tested that it compiles and all tests pass with clang 19.1.7.

https://clang.llvm.org/docs/LanguageExtensions.html#vectors-and-extended-vectors

Closes #2998 and #2997

See merge request libeigen/eigen!2051

Co-authored-by: Rasmus Munk Larsen <rmlarsen@google.com>
Co-authored-by: Antonio Sánchez <cantonios@google.com>
This commit is contained in:
Rasmus Munk Larsen
2025-11-06 21:52:19 +00:00
parent 7c7d84735e
commit ec93a6d098
17 changed files with 1161 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
Eigen/src/Core/GenericPacketMath.h
View File

@@ -1224,7 +1224,7 @@ EIGEN_DEVICE_FUNC inline typename unpacket_traits<Packet>::type pfirst(const Pac
template <typename Packet>
EIGEN_DEVICE_FUNC inline std::conditional_t<(unpacket_traits<Packet>::size % 8) == 0,
typename unpacket_traits<Packet>::half, Packet>
predux_half_dowto4(const Packet& a) {
predux_half(const Packet& a) {
return a;
}
@@ -1342,7 +1342,7 @@ struct pmadd_impl {
return psub(c, pmul(a, b));
}
static EIGEN_DEVICE_FUNC EIGEN_ALWAYS_INLINE Packet pnmsub(const Packet& a, const Packet& b, const Packet& c) {
return pnegate(pmadd(a, b, c));
return pnegate(padd(pmul(a, b), c));
}
};
@@ -1476,8 +1476,8 @@ struct PacketBlock {
Packet packet[N];
};
template <typename Packet>
EIGEN_DEVICE_FUNC inline void ptranspose(PacketBlock<Packet, 1>& /*kernel*/) {
template <typename Packet, int size = 1>
EIGEN_DEVICE_FUNC inline void ptranspose(PacketBlock<Packet, size>& /*kernel*/) {
// Nothing to do in the scalar case, i.e. a 1x1 matrix.
}