Spent several hours last week in Portland with @TimSweeneyEpic and Jim. Great insights from Tim on the server side computing challenges and opportunities, based on learnings from fortnite user growth.pic.twitter.com/gbqlTr8OzD
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
I assume you’re joking but massively parallel simulations do not require coherence across nodes or a unified address space. 99% of the codes use MPI, which provides messaging semantics. A few folks use PGAS, which is not coherent but a kind of unified addressing.
Message passing is sort of the assembly language of distributed systems. The hard higher-level part is coordinating consistent operations involving multiple objects while lots of things are going on in parallel.
Or run gameplay on the cloud and render on the client. Nodes much closer to the edge, costs reduced with no GPU on the cloud, massive scalability with any cloud provider, ability to predict ahead on the client, etc... That's what we do @Polystream :)
We (InsightfulVR) have spent 3 years building a platform that UE4 plugs into which enables latency tolerant distributed rendering... to power AAA games on mobile SoCs - let us know if you'd like to see it at some point
Wow I had to do a double take in this. But yeah it could make sense. I think unifying the entire memory will be too much but an region could be cool. You could send binary objects without caring too much about remapping things
I've tried NVidia's streaming service with this shitty optimised MHW and it was working fine, but clearly it's not a solution for FPS. 60 IPS is too low for many people, and the input delay is too important. (I've an excellent connexion...)
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.