Optimized risc-v memcpy

teknoraver · January 28, 2026, 1:12am

Hi,

I’ve implented a (hopefully) faster memcpy() function for the RISC-V kernel.

Performances when copying unaligned buffers are up to +130% faster.

Please give it a try and let me know. More details in the commit message.

Patch is at:

github.com/teknoraver/linux

riscv: memcpy: fast copy for unaligned buffers

committed 01:00AM - 28 Jan 26 UTC

teknoraver

+73 -4

The RISC-V memcpy() does an 8 byte wide copy when the two buffers have the same …alignment, and fallbacks to a single byte copy otherwise. Implement a unaligned aware 8 byte copy when buffers are unaligned which copies 8 bytes at time by doing proper shifting. Benchmarks shows that the aligned code path is unaffected, while the unaligned one gets a ~2.3x boost. Benchmark with the current implementation: memcpy: aligned copy of 400 MBytes in 429 msecs (931 MB/s) memcpy: unaligned copy of 400 MBytes in 1202 msecs (332 MB/s) Benchmark with the new unaligned copy: memcpy: aligned copy of 400 MBytes in 428 msecs (933 MB/s) memcpy: unaligned copy of 400 MBytes in 519 msecs (770 MB/s) Signed-off-by: Matteo Croce <[email protected]>