Search
Tuesday, 18th of February 2020, 7:27:42 UTC
7:47:35
beach
The French TV news is full of bad grammar. The other day, I heard (in French): Being 4 months pregnant, the fire fighters took her to the hospital.
7:47:54
beach
I can point to half a dozen of those pretty much every day.
8:02:10
no-defun-allowed
Good morning splittist.
8:36:58
shka_
no-defun-allowed: hi!
8:37:10
no-defun-allowed
Hello shka_.
8:37:34
shka_
do you happen to have adaline implementation laying around? I remotely remember that you were working on neural networks.
8:37:43
shka_
and i have dirty idea i want to try out
8:38:33
no-defun-allowed
Do you mean https://en.wikipedia.org/wiki/ADALINE?
8:39:44
no-defun-allowed
I didn't have anything very serious or readable; I still have nightmares of my backwards propagation loop.
8:40:04
no-defun-allowed
It doesn't look too hard to implement though. Let me have a go at that.
8:40:14
shka_
no, it is very simple
8:40:25
shka_
in fact it can be implemented in a hardware easily
8:46:11
no-defun-allowed
ACTION has to solve the plaster.tymoon.eu captcha
8:47:16
no-defun-allowed
I don't have any testing data, but do you want something like https://plaster.tymoon.eu/view/1675?
8:51:25
shka_
no-defun-allowed: thank you very much
11:51:19
heisig
beach: You worry too much about register allocation :)
11:53:06
heisig
Getting SICL to run is much more important than squeezing out the final 2% of a register allocation algorithm.
12:33:50
shka_
well, more importantly, register allocation will be useless unless SICL actually runs
13:08:45
beach
Yes, sure, I agree. It is just that these ideas happened to come to mind now.
13:08:55
beach
I won't implement it right away.
13:09:06
beach
But I also don't want to forget those ideas.
13:10:02
heisig
Phew. Although I'm looking forward to working on such things - once there is a way to run SICL on my machine (or better, my browser).
13:10:33
beach
Yes, the initial version will be simple. No sophisticated register allocation.
13:13:17
heisig
By the way, we also booked a room in the Sorell Hotel Seefeld, from April 25 to May 1.
13:14:10
beach
Ample opportunity for dinners.
13:24:40
heisig
OK, I just uploaded the first version of my ELS paper (http://0x0.st/iXjD.pdf).
13:26:12
beach
shka_: You need to remove ). at the end.
13:26:21
beach
... and ( at the beginning.
13:27:16
heisig
shka_: Or did you wonder why I want to run SICL in the browser?
13:27:42
beach
Ah, I might have misunderstood, yes.
13:27:42
shka_
heisig: i am more wondering about what do you mean by sicl in the browser
13:29:22
heisig
shka_: SICL is very portable. So its sounds like a natural thing to bootstrap it into WASM at some point.
13:29:53
shka_
this would be pretty cool thing to have
13:30:46
beach
heisig: I will read your paper later in the week. I may have to be reminded.
13:31:44
heisig
I should also mention that I didn't do a lot of proofreading yet. I will upload a better version this weekend.
13:31:58
heisig
But I wanted to share my progress.
Tuesday, 18th of February 2020, 19:27:42 UTC