llama.cpp
9c5b594c - iq3_s: another small ARM_NEON improvement

Commit

1 year ago

iq3_s: another small ARM_NEON improvement 10.7 -> 11.0 t/s. Using vmulq_s8 is faster than the xor - sub trick that works best on AVX2.

References

Author

Iwan Kawrakow

Parents