Re: memory mapped store entries

On Wed, 26 Aug 1998, Stephen R. van den Berg wrote:

> Yes, well, I think the rationale behind the pros and cons of mmapped store
> entries have been discussed at length now.

Indeed :)

> I think everyone understands the potential benefits/risks now.

To be sure that everyone's understanding is the same, could you please give a
concise enumeration (one line per item) of benefits and risks derived from
the discussion? Preferably with hard numbers given some cache configuration
(say a 20GB cache, etc.).

This way, when the patch gets out there, every admin would be able to
        - see if the patch should help her (thus, worth testing)
        - see if the patch meets the expectations (after testing)

Otherwise, there will be just one more optimization that is added just
because it "looks good".

Thank you,

Alex.
Received on Tue Jul 29 2003 - 13:15:53 MDT