Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2016-07-25

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:25 colomon_ joined #moarvm
02:05 colomon joined #moarvm
02:52 colomon joined #moarvm
06:32 domidumont joined #moarvm
06:53 domidumont joined #moarvm
06:58 domidumont joined #moarvm
07:10 zakharyas joined #moarvm
08:06 domidumont joined #moarvm
09:32 Ven_ joined #moarvm
10:11 domidumont joined #moarvm
11:29 brrt joined #moarvm
11:41 brrt ehm, you may with reason not believe me
11:41 jnthn yeah, right
11:41 brrt but i think i finally, really, truly know how to implement the register allocator 8-)
11:41 jnthn \o/
11:42 brrt so now the issue is to make time for it
11:43 brrt one of the interesting bits is that i think i've discovered that linear scan (known as fast) is a subset of graph coloring
11:43 brrt that is going to be a blogpost soonish
11:44 brrt graph coloring is known as slow, because a): graph coloring is .... whatsitcalled? traveling-salesman-equivalent complexity
11:44 brrt and linear scan is known as fast because linear
11:45 brrt but in reality linear scan is just a special version of graph coloring
11:46 timotimo NP-complete is the word you're looking for?
11:46 brrt i.. think so
11:46 brrt yeah, thanks :-)
12:02 arnsholt That is indeed NP-complete
12:03 arnsholt When I took a databases class some years ago, we spent quite a bit of time discussing serialization algorithms/strategies (that being rather important for databases)
12:03 brrt uhuh
12:03 arnsholt First we discussed locking, which has some obvious drawbacks, then we covered something called view-serializability (IIRC)
12:03 brrt (fwiw, i'm not sure a live range in this model can still meanigfully live in two registers.... that just has to go then)
12:03 arnsholt Think we spent like two lectures on it
12:04 brrt do continue
12:04 arnsholt And then at the end they just casually drop, "but this turns out to be NP-complete"
12:04 arnsholt >>>>>>.<<<<<<<
12:04 brrt lol
12:04 brrt yeah, sucks when that happens
12:05 brrt that basically happened to my idea of 'let's find the ideal combination of renewable generators for a given area'
12:05 brrt 'this is going to be easy, just fill them up with the most ideal spots first... oh wait'
12:05 arnsholt Yeah, turns out a rather large fraction of interesting problems are NP-complete
12:09 brrt or solved by union-find
12:10 timotimo when you get a job, you better find a union to join ...
12:45 brrt :-)
12:45 brrt okay, i'm off, i have to write thesis reports and stuff
12:45 brrt see you later
16:27 FROGGS joined #moarvm
18:56 domidumont joined #moarvm

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary