On 05/12/2024 14:31, Sam Russell wrote:
Hi,
I've been looking into other CRC32 implementations for architectures that don't have accelerated
polynomial multiplication (PCLMUL/VMULL for x86-SSE/AVX or ARM-NEON), and there are improvements we
can make, but they make tradeoffs for memory (both fo
Hi,
I've been looking into other CRC32 implementations for architectures that
don't have accelerated polynomial multiplication (PCLMUL/VMULL for
x86-SSE/AVX or ARM-NEON), and there are improvements we can make, but they
make tradeoffs for memory (both for lookup tables and for working set). The
"b
