Elixir and Machine Learning: Nx v0.1 released

Re:Elixir and Machine Learning: Nx v0.1 released by alberth on Thursday January 06, 2022 @10:03AM (#29824347)

Off-topic: hi Jose, what's your expectation on how much more raw performant Erlang can become in the next 1-2 years?

PHP ~3x in performance from 5.x to 7.x. LuaJIT is even faster over vanilla Lua.

There were multiple attempts at creating an Erlang JIT for well over a decade now and last year BeamAsm was released but in everyday use it seems to net about a 10-15% boast.

I ask because it'd be great if we didn't have to create so many Rust NIFs in order to address Erlang raw slowness.

Curious what your take is on speeding up raw Erlang (please note, I'm not talking about concurrency & I do truly love Exilir/Erlang/OTP)

[ Parent | on HN ]

Re:Elixir and Machine Learning: Nx v0.1 released by josevalim on Thursday January 06, 2022 @10:14AM (#29824489)

I am not informed enough to give actual estimates. :)

Although your estimates for JIT improvements seem low on my experience. Compilation times are consistently half of what they were before and test suites range between 33% to 50% faster. Even Whatsapp reported server efficiency increased by 25% and many other factors are likely at play there (source: https://twitter.com/wcathcart/status/1385253969522413568).

I _assume_ we will continue seeing improvements on this front, especially if at some point we start doing cross module optimizations!

[ Parent | on HN ]

Re:Elixir and Machine Learning: Nx v0.1 released by alberth on Thursday January 06, 2022 @10:25AM (#29824681)

Thanks! Hope so too. I feel like raw performance is the #1 thing from greater adoption of the echo system.

[ Parent | on HN ]