Age | Commit message (Collapse) | Author | |
---|---|---|---|
2024-02-16 | Remove redundant shift | Ian Moffett | |
Signed-off-by: Ian Moffett <ian@osmora.org> | |||
2024-02-16 | Update copyright | Ian Moffett | |
Signed-off-by: Ian Moffett <ian@osmora.org> | |||
2023-10-20 | Ensure cpuid macro is only defined on AMD64 | sigsegv7 | |
Signed-off-by: sigsegv7 <ian@vegaa.systems> | |||
2023-09-18 | Fix overflow issue + odd behaviour | sigsegv7 | |
It is best to check for an overflow risk and lower the step BEFORE we actually perform the inversion Signed-off-by: sigsegv7 <ian@vegaa.systems> | |||
2023-09-18 | Shift step left by 1 if SSE is supported | sigsegv7 | |
This is a faster way of increasing the step as they are in powers of two. Signed-off-by: sigsegv7 <ian@vegaa.systems> | |||
2023-09-18 | Optimize with SSE2/SSE3 and 128 bit inversions | sigsegv7 | |
Signed-off-by: sigsegv7 <ian@vegaa.systems> |