Re: Cost of XLogInsert CRC calculations

On E, 2005-05-16 at 12:12 +0100, Mark Cave-Ayland wrote:
> > -----Original Message-----
> > From: Mark Cave-Ayland [mailto:m(dot)cave-ayland(at)webbased(dot)co(dot)uk]
> > Sent: 16 May 2005 09:04
> > To: 'Simon Riggs'
> > Cc: 'Christopher Kings-Lynne'; 'Tom Lane'; 'Bruce Momjian';
> > 'pgsql-hackers(at)postgresql(dot)org'
> > Subject: RE: [HACKERS] Cost of XLogInsert CRC calculations
>

>
> Win32 UINT64: 0x782104059a01660 (crc0) ~158us
> Win32 UINT32: 0x78210405 (crc1), 0x59a01660 (crc0) ~58us
> FC1 UINT64: 0x782104059a01660 (crc0) ~76us
> FC1 UINT32: 0x78210405 (crc1), 0x59a01660 (crc0) ~29us

It would interesting to see how much of the time taken is actual algorithm
and how much is getting at the data (i.e the fact that the page has to go
through CPU at all, instead DMA'ing it to disk).

In your test setup you probably have the whole thing in CPU cache already,
but still it would be interesting to time the same thing with some simple
algorithm, like XOR over the whole 8k page, for comparison.

--
Hannu Krosing <hannu(at)skype(dot)net>