dumblob

Crossbeam and such (incl. Actix) are nice, very powerful "frameworks" but are very heavyweight and their performance is also not the top for "light" computations like HVM does. So I'm...

@VictorTaelin that's 99% true with the 1% caveat :wink:. Namely a shared queue doesn't scale linearly with number of processors (cores, GPUs, FPGAs, etc.). You can see it in [Weave](https://github.com/mratsim/weave/...

Any ideas? Does blosc2 do **much** better on this front (I'm assuming Python bindings which I didn't try yet)?

Ah, thanks. That clarifies it a lot. Any concrete suggestions how to approach compression of short inputs?

These projects use IMHO state of the art "multithreaded" queues: https://github.com/pramalhe/ConcurrencyFreaks https://github.com/mratsim/weave (e.g. https://github.com/mratsim/weave/pull/21 ) http://daugaard.org/blog/writing-a-fast-and-versatile-spsc-ring-buffer http://www.vitorian.com/x1/archives/370 Also the queue size might be between the size of L1 and 2*L2...

Did you solve this issue somehow?

>I've briefly played with integrating some concatenative capabilities into Tao, although I believe that currying and function composition net a considerable number of the benefits that concatenative features bring. If...

#### MOTIVATIONAL INTERMEZZO What I mean with **practical** is only one thing: a programmer just wants to **get things done** - i.e. as fast as possible and with as low...

Very interesting observations! I'm glad you are raising them as I can then better convey my thoughts. >>without taking any future into consideration because that's the job of the programming...

@zesterer speaking of advanced types you might be interested in the language [Formality](https://github.com/Soonad/FormalityFM ) (see the broader scope in https://github.com/Soonad/Whitepaper ).