Now
MAIN commitmail json YAML
src/sys/crypto/chacha/arch/arm/chacha_neon.c@1.7
/
diff
/
nxr@1.7
src/sys/crypto/chacha/arch/arm/chacha_neon.h@1.3 / diff / nxr@1.3
src/sys/crypto/chacha/arch/arm/chacha_neon_32.S@1.1 / diff / nxr@1.1
src/sys/crypto/chacha/arch/arm/files.chacha_arm@1.3 / diff / nxr@1.3
src/tests/sys/crypto/chacha/Makefile@1.3 / diff / nxr@1.3
src/sys/crypto/chacha/arch/arm/chacha_neon.h@1.3 / diff / nxr@1.3
src/sys/crypto/chacha/arch/arm/chacha_neon_32.S@1.1 / diff / nxr@1.1
src/sys/crypto/chacha/arch/arm/files.chacha_arm@1.3 / diff / nxr@1.3
src/tests/sys/crypto/chacha/Makefile@1.3 / diff / nxr@1.3
Implement 4-way vectorization of ChaCha for armv7 NEON.
cgd performance is not as good as I was hoping (~4% improvement over
chacha_ref.c) but it should improve substantially more if we let the
cgd worker thread keep fpu state so we don't have to pay the cost of
isb and zero-the-fpu on every 512-byte cgd block.
cgd performance is not as good as I was hoping (~4% improvement over
chacha_ref.c) but it should improve substantially more if we let the
cgd worker thread keep fpu state so we don't have to pay the cost of
isb and zero-the-fpu on every 512-byte cgd block.