Messages in this thread | | | From | David Laight <> | Subject | RE: [PATCH 04/18] csum_and_copy_..._user(): pass 0xffffffff instead of 0 as initial sum | Date | Thu, 23 Jul 2020 15:36:52 +0000 |
| |
From: Al Viro > Sent: 23 July 2020 16:21 ... > The point is, your "~4.5 cycles per vector" is pretty much noise and the > difference between the 3-argument and 4-argument variants could easily be > in the same range. It might be a valid microoptimization, it might be not. > 3-argument variant is simpler and IMO in absence of strong data we ought > to go with that.
There is definitely more to be gained by rewriting the x86-86 asm.
David
- Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
| |