Re: Fastcode MM Rule

Hello Pierre,
Thanks for your answers.

> However, it's not only the memory usage that is a concern: If the blocks
> are smaller they can be more tightly packed and that speeds up access
> speeds, something which our B&V tool does not really measure. (We write
> the beginning and end of blocks, but never again read them.)

Interestingly Eric's new LinkedList benchmark shows some of this behaviour.
BucketMM is about 10% slower than Recycler or FastMM3 because it requires a
32-byte allocation for the 16-byte TLinkedListItem record
By removing the "List" variable in the record (which isn't used at all, no
idea why it's there) to make it 12-byte record, BucketMM has the same speed
as the others.

Kind regards,
Robert


.

Relevant Pages

  • Re: The Speed Trap - BBC1 Scotland
    ... >> Yet you're not prepared to consider removing any just now. ... dramatically reduced traffic speeds and actually made streets safer." ... >>> against me because I'm rude to you, ... if they were in a more intimate relationship with other road ...
    (uk.rec.cycling)
  • Re: The Speed Trap - BBC1 Scotland
    ... -"Hans Monderman, a traffic engineer working in Holland, discovered that removing all traffic signs, speed humps, line markings and traffic lights dramatically reduced traffic speeds and actually made streets safer." ...
    (uk.rec.cycling)