This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2018-01-22
Channels
- # architecture (30)
- # beginners (56)
- # cider (16)
- # cljs-dev (12)
- # cljsrn (21)
- # clojure (169)
- # clojure-austin (1)
- # clojure-estonia (1)
- # clojure-italy (3)
- # clojure-russia (1)
- # clojure-spec (56)
- # clojure-uk (46)
- # clojurescript (53)
- # consulting (3)
- # core-async (3)
- # cursive (14)
- # data-science (16)
- # datascript (1)
- # datomic (26)
- # defnpodcast (11)
- # docs (3)
- # emacs (6)
- # fulcro (4)
- # graphql (24)
- # hoplon (8)
- # instaparse (4)
- # java (2)
- # jobs (1)
- # jobs-rus (1)
- # jobs_rus (1)
- # keechma (1)
- # luminus (2)
- # lumo (1)
- # mount (36)
- # off-topic (30)
- # om-next (5)
- # onyx (29)
- # precept (23)
- # re-frame (20)
- # reagent (2)
- # remote-jobs (9)
- # ring (2)
- # ring-swagger (3)
- # rum (3)
- # shadow-cljs (100)
- # spacemacs (17)
- # sql (10)
- # timbre (2)
- # unrepl (29)
- # yada (2)
got to do some embedded work on a arm cortex-a8, 1gz, 512MB ram. user space, linux is the os. jvm clojure seems too heavyweight for me, and cljs on node seems too janky. thinking maybe rust? not sure
@kenrestivo I had the same problem around a year ago, clojurescript works really well!
anyone else have this problem where every few months, they go "f--- clojure, I want to use forth / stack based language, read a few chapters of thinking forth, then figure out new ways to restructure their clojure code" ?
As someone who has interacted with Forth on and off for decades (I first encountered it at university in the very early 80's), no, I've never been so enamored of it that I think it's better than any of the Lisps I've used 🙂
@qqq When did you first encounter Forth? And when did you first encounter a Lisp?
@seancorfield first encounter of Lisp was in 12th grade? someone handed me a copy of SICP; but itw as very set! set-car! set-cdr! ish lisp, so not any of this "functional programming / purity"
as for Forth, I've never used it seriously -- I've written a few stack based interpreters here and there, but every time I'j mabout to go full on forth, I see a way to do similar things i Clojure
I feel like one thing stack based langauges get right is -- I want to name FUNCTIONS, not name variables
Did you ever look at Gershwin?
akiroz: i have not been able to get something as simple as a nrepl connection to an npm clojurescript repl on another machine working
i'm obviously missing something in the lumo + npm world. lumo also seems to not have the best docs
@seancorfield: I think you mentikend this to me once before; https://github.com/gershwin/gershwin is there sample code anywhere? the website seems down
looks like a DSL
does anyone know what the fastest matrix-matrix multiply on webassembly is? I know that for desktops, there's all the crazy SSE / caching tricks, but it's not clear if any/all of them applies to webassembly impls
@qqq I’ll try some people offline. My best contacts for webassembly aren’t clojurists.
@qqq how large are these matrices?
then I'm doing some image convolutoins + neural networks, which largely reduces down to gemms
@qqq does wasm even support SIMD? I didn’t think so.
Also, I'm not 100% sure WASM will be much faster than pure JS. Pure JS numeric performance is pretty fast, and in general WASM is still new and unoptimized.
Hello my friends 🙂 This is Benny, Software Engineer from Singapore. I have just created a reading list for people who are interested in Scalability or need to prepare for a System Design interview at:
Please check it out and let me know your idea! Any star ⭐ will be greatly appreciated ❤️ Thank you very much 👍
I'm Oblivious. Now I'm looking for a consulting/contracting gig! Just got the word that Friday is my last day. Wow, I'm suprised.... but in hindsight... when they are letting go of fulltime people.... consultant/contractors are not likely to be around for long.
@gonewest818: I also believe "wasm does not support SIMD" -- which is making me wonder if any of the standard matrix-multiply optimizations apply to WASM -- and if not, what the right way to write matrix multiply on WASM Is. @tbaldridge: One argument I have heard for WASM being faster -- is that there's no unboxingboxing, and you know upfront what's an int / float, so there's no checks there. I suspect (but do not know for a fact), that it's probably easy t otranslate wasm to assembly, something that one can't readily do with JS.
@qqq most of that is optimized away in JS