Overview
Examples
Screenshots
Comparisons
Applications
Download
Documentation
Tutorials
Bazaar
Status & Roadmap
FAQ
Authors & License
Forums
Funding Ultimate++
Search on this site
Search in forums












SourceForge.net Logo
Home » Community » Coffee corner » Map implementation
Re: Map implementation [message #51549 is a reply to message #51547] Sun, 14 April 2019 08:03 Go to previous messageGo to previous message
mirek is currently offline  mirek
Messages: 12293
Registered: November 2005
Ultimate Member
Novo wrote on Sat, 13 April 2019 21:07
mirek wrote on Wed, 10 April 2019 11:37

Quote:

Let's take a look at two scenarios.

1. One million hash tables containing one hundred records.
2. One hash table containing one hundred million records.

You will need two completely different implementations in these cases.


That might be true, however I do not see a way how to improve Index for either (I see some accumulated knowledge how to improve it for both, but that is another story).

IMHO, this is impossible because in the first case your main concern is the memory usage. Tiny overhead multiplied by million is a huge problem. And in the second case performance is the main issue.


If the performance is the issue, then the memory is the issue too. The game starts at L1 cache size, which can correspond to hunderds of records.
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: [Skylark][witz] How would you do this
Next Topic: Strange Architecture
Goto Forum:
  


Current Time: Thu Apr 02 14:32:37 CEST 2020

Total time taken to generate the page: 0.01577 seconds