I am in the process of testing starlit for the Hutter prize. I expect it to take about 2 days to compress 1 GB to 115 MB and 2 days to decompress. That won't beat NNCP at 108 MB but it is the best within the hardware constraints (10 MB memory, 100 CPU hours in 1 thread). NNCP uses a GPU. Starlit is running in an Ubuntu window under Windows 10 on a new Lenovo laptop with 16 GB. I have wifi off to keep Windows from installing updates and killing the program. When I'm done testing I will update the benchmark.
Starlit reorders the articles in enwik9 by mutual information, then uses dictionary preprocessing from phda9 (previous Hutter prize winner) and compresses with a reduced version of cmix. But yeah, it helps to quote some context when posting to the list. On Wed, Jun 9, 2021, 11:11 PM <[email protected]> wrote: > Matt are you going to update this new score below or why not ? > https://bellard.org/nncp/ > > And are you trying starlit on 1GB ? Or why not? > > And do I use 7zip to zip my software? > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + > delivery options <https://agi.topicbox.com/groups/agi/subscription> > Permalink > <https://agi.topicbox.com/groups/agi/Ta8a0e64036790330-M39533fa13d100822a89fde34> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Ta8a0e64036790330-M9bfc816718b726d16164c120 Delivery options: https://agi.topicbox.com/groups/agi/subscription
