Re: Cost of XLogInsert CRC calculations

Tom Lane wrote:
> "Mark Cave-Ayland" <m(dot)cave-ayland(at)webbased(dot)co(dot)uk> writes:
> > I was just researching some articles on compression (zlib) and I saw mention
> > of the Adler-32 algorithm which is supposed to be slightly less accurate
> > than an equivalent CRC calculation but significantly faster to compute. I
> > haven't located a good paper comparing the error rates of the two different
> > checksums,
>
> ... probably because there isn't one. With all due respect to the Zip
> guys, I doubt anyone has done anywhere near the analysis on Adler-32
> that has been done on CRCs. I'd much prefer to stick with true CRC
> and drop it to 32 bits than go with a less-tested algorithm. Throwing
> more bits at the problem doesn't necessarily create a safer checksum.

Agreed. 64-bit was overkill when we added it, and it is now shown to be
a performance problem.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073