Zandronum Chat on our Discord Server Get the latest version: 3.2
Source Code
My View | View Issues | Change Log | Roadmap | All Projects Issue Support Ranking | Rules | My Account
  

View Revisions: Issue #1268 [ All Revisions ] [ Back to Issue ]
Summary 0001268: Huffman optimization
Revision 2013-01-29 17:23 by Dusk
Description I've heard that the current Huffman table is not as optimized as it should be (note: someone should upload the huffman.zip somewhere on Zandronum's main site because it's linked to skulltag.com and no one knows how stable that site will be).
If Huffman is used in compressing and sending packets to the clients, optimization of this table may help in coop/major monster games (like ZDW) where people tend to time out due to bandwidth issues.

If the idea is considered, to allow for backwards compatibility, maybe there could be an sv_optimizedhuffman cvar that will tell the server to transmit data based on a new huffman table (if one should be made). Making a new table just as a basic brainstorm could be done with monitoring the raw bytes pre-huffman encoding and storing the amount in the table, and then generating a new frequency table from there. It would be interesting to see if there is a significant difference or not.

Another even further idea would be to allow for dynamic huffman tables, not that I expect that to actually be implemented.
Revision 2013-01-29 17:16 by Dusk
Description I've heard that the current Huffman table is not as optimized as it should be (note: someone should upload the huffman.zip somewhere on Zandronum's main site because it's linked to skulltag.com and no one knows how stable that site will be).
If Huffman is used in compressing and sending packets to the clients, optimization of this table may help in coop/major monster games (like ZDW) where people tend to time out due to bandwidth issues.

If the idea is considered, to allow for backwards compatibility, maybe there could be an sv_optimizedhuffman cvar that will tell the server to transmit the new huffman table (if one should be made). Making a new table just as a basic brainstorm could be done with monitoring the raw bytes pre-huffman encoding and storing the amount in the table, and then generating a new frequency table from there. It would be interesting to see if there is a significant difference or not.

Another even further idea would be to allow for dynamic huffman tables, not that I expect that to actually be implemented.




Questions or other issues? Contact Us.

Links

Copyright © 2000 - 2025 MantisBT Team
Powered by Mantis Bugtracker