Not that I actually need it but the websocket server currently runs on a machine with 128GB RAM.
Works just fine, the websocket "router" currently juggles hundreds of concurrent and persistent connections without using noticeable amounts of RAM and CPU. Due to a mistake I made processes would sometimes start to use 100% of a core (there are 16 of them), I just noticed it by coincidence because it didn't affect the performance of the websocket server (since it needs so little CPU anyway). Hopefully I just uploaded a fix but more on that in a future post if it indeed did fix the problem. On Thu, Mar 12, 2015 at 2:33 PM, Jakob Eriksson <ja...@aurorasystems.eu> wrote: > > http://www.infoq.com/news/2015/03/petabyte-jvms?utm_source=hacker%20news&utm_medium=link&utm_campaign=petabyte%20news > > They describe a machine with NUMA, and basically run a handful of JVMs, > each > confined to a NUMA region. Then synchronize with a huge disk cluster. > > I wonder how PicoLisp would fare on large machines with large sets of data? > > I have a feeling "really great". :) > > Has someone on the list tried PicoLisp on a monster PC? > > best regards, > Jakob > -- > UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe >
