Author Topic: A "new" compression format [subject to changes]  (Read 10339 times)

0 Members and 1 Guest are viewing this topic.

Offline timwessman

  • LV3 Member (Next: 100)
  • ***
  • Posts: 94
  • Rating: +32/-0
    • View Profile
Re: A "new" compression format [subject to changes]
« Reply #15 on: November 25, 2013, 02:35:10 pm »
I wonder if such compression would be fast enough for calcs?

Dunno about this particular algo, but as an interesting side note we gained about 2.5MB of size with the internal calc on Prime by using a huffman encoding for that text.
TW

Although I work for the HP calculator group, the comments and opinions I post here are my own.

Offline harold

  • LV5 Advanced (Next: 300)
  • *****
  • Posts: 226
  • Rating: +41/-3
    • View Profile
Re: A "new" compression format [subject to changes]
« Reply #16 on: May 25, 2014, 08:30:34 am »
Update time! Added a very simple trick. It now takes about 0.4 miliseconds on the same test data (it varies between 350 and 400, most commonly 390).
That's varying between 360 to 410 MB/s, usually 370MB/s.

There's also a very easy way to make it faster: use a simpler way to encode length/distance pairs. But that would change the format a lot, and I intended to stay close to Deflate.
Blog about bitmath: bitmath.blogspot.nl
Check the haroldbot thread for the supported commands and syntax.
You can use haroldbot from this website.