New MEMORY USAGE measurement proposal for MM B&V
From: Ivo Tops (ivotops_at_topsoftwaresite.nl)
Date: 02/25/05
- Next message: Ivo Tops: "Two new tests for MM B&V"
- Previous message: Arash Partow: "Re: Suggestion for Fastcode challenge "Hash""
- Next in thread: Pierre le Riche: "Re: New MEMORY USAGE measurement proposal for MM B&V"
- Reply: Pierre le Riche: "Re: New MEMORY USAGE measurement proposal for MM B&V"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Fri, 25 Feb 2005 08:03:47 +0100 To: pleriche@hotmail.com, marianndkc@home3.gvdnet.dk
I think the Relative performance is not really straightforward and
usefull. When testing a Memory manager I want to know speed and memory
efficiency. The combined relative performance does not really tell me
what I want to know.
These two would;
MEMORY USAGE
========================
Add to each test in code the peak and average useage in the regular
memory manager (test once). And give the number as a relative to this.
So a memory usage of 140% would mean this memory manager used 40% more
memory as the delphi standard.
PERFORMANCE
========================
Switch from displaying Ticks to displaying seconds for easier reading.
Current measurement is fine.
Kind regards,
Ivo Tops
- Next message: Ivo Tops: "Two new tests for MM B&V"
- Previous message: Arash Partow: "Re: Suggestion for Fastcode challenge "Hash""
- Next in thread: Pierre le Riche: "Re: New MEMORY USAGE measurement proposal for MM B&V"
- Reply: Pierre le Riche: "Re: New MEMORY USAGE measurement proposal for MM B&V"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
