
With liblzma when built for x86 or x86_64 architectures, SSE2 intrinsics are now attempted instead of memcpy() for the LZMA/LZMA2 decoder. This SSE2 intrinsic use can yield up to 5% reduced decompression times. Or if building against the musl C library, these SSE2 intrinsics can lead to an over 15% time reduction for highly compressed files.
In addition to the XZ 5.8 SSE2 intrinsics use for liblzma, there are also CMake build system updates, language translation updates, improved LZMA/LZMA2 encoder performance for 64-bit POWER and RISC-V, rewriting of the x86_64 E2K CLMUL CRC code, and other optimizations.

The XZ 5.8 update was managed by Lasse Collin and can be downloaded from the official XZ project site .