Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 973 974 975 976 977 978 979 980 981 982 983 .. 5022
— 502,152 events total
2020-10-31 13:24:40 hackage redis-io 1.1.0 - Yet another redis client. https://hackage.haskell.org/package/redis-io-1.1.0 (ToralfWittner)
2020-10-31 13:25:43 → lxsameer joins (~lxsameer@unaffiliated/lxsameer)
2020-10-31 13:28:07 → nschoe joins (~quassel@91-175-19-30.subs.proxad.net)
2020-10-31 13:29:42 → geowiesnot joins (~user@i15-les02-ix2-87-89-181-157.sfr.lns.abo.bbox.fr)
2020-10-31 13:30:10 hackage wai-predicates 1.0.0 - WAI request predicates https://hackage.haskell.org/package/wai-predicates-1.0.0 (ToralfWittner)
2020-10-31 13:32:37 × jud^ quits (~jud@cpe-70-113-106-222.austin.res.rr.com) (Quit: Leaving)
2020-10-31 13:32:52 × rotty quits (rotty@ghost.xx.vu) (Ping timeout: 260 seconds)
2020-10-31 13:32:54 → jud joins (~jud@unaffiliated/jud)
2020-10-31 13:33:30 → rotty joins (rotty@ghost.xx.vu)
2020-10-31 13:33:53 → mmfood joins (~mmfood@81-232-132-159-no64.tbcn.telia.com)
2020-10-31 13:39:05 × nschoe quits (~quassel@91-175-19-30.subs.proxad.net) (Ping timeout: 240 seconds)
2020-10-31 13:39:37 → ahmr88 joins (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net)
2020-10-31 13:40:55 → akad_ joins (~akad@109107030050.radomsko.vectranet.pl)
2020-10-31 13:42:27 → thir joins (~thir@p200300f27f0b7e004c18ab60065ea01b.dip0.t-ipconnect.de)
2020-10-31 13:43:38 × drbean quits (~drbean@TC210-63-209-85.static.apol.com.tw) (Ping timeout: 260 seconds)
2020-10-31 13:47:32 × mmfood quits (~mmfood@81-232-132-159-no64.tbcn.telia.com) (Quit: Leaving)
2020-10-31 13:51:10 × Ohg quits (~ohg@host31-54-224-58.range31-54.btcentralplus.com) ()
2020-10-31 13:51:30 → coot joins (~coot@37.30.55.3.nat.umts.dynamic.t-mobile.pl)
2020-10-31 13:52:02 × geekosaur quits (ac3a3be8@172.58.59.232) (Remote host closed the connection)
2020-10-31 13:53:20 → texasmynsted joins (~texasmyns@212.102.45.109)
2020-10-31 13:53:24 × MindlessDrone quits (~MindlessD@unaffiliated/mindlessdrone) (Ping timeout: 240 seconds)
2020-10-31 13:54:08 → texasmyn_ joins (~texasmyns@212.102.45.109)
2020-10-31 13:55:07 × texasmynsted quits (~texasmyns@212.102.45.109) (Read error: Connection reset by peer)
2020-10-31 13:56:09 → MindlessDrone joins (~MindlessD@unaffiliated/mindlessdrone)
2020-10-31 13:56:39 → jneira joins (501e64fa@gateway/web/cgi-irc/kiwiirc.com/ip.80.30.100.250)
2020-10-31 13:56:54 → nbloomf joins (~nbloomf@2600:1700:ad14:3020:c427:c5ca:d62:565b)
2020-10-31 13:57:17 × raoulb quits (~weechat@stateless.vsos.ethz.ch) (Quit: raoulb)
2020-10-31 13:57:37 → ddellacosta joins (~dd@86.106.121.168)
2020-10-31 13:58:40 hackage polysemy-methodology 0.1.4.0 - Domain modelling algebra for polysemy https://hackage.haskell.org/package/polysemy-methodology-0.1.4.0 (locallycompact)
2020-10-31 14:00:02 → FreeBirdLjj joins (~freebirdl@101.228.42.108)
2020-10-31 14:01:00 × nbloomf quits (~nbloomf@2600:1700:ad14:3020:c427:c5ca:d62:565b) (Client Quit)
2020-10-31 14:01:23 → Ariakenom joins (~Ariakenom@h-155-4-221-164.NA.cust.bahnhof.se)
2020-10-31 14:05:17 → is_null joins (~jpic@pdpc/supporter/professional/is-null)
2020-10-31 14:05:46 × idhugo quits (~idhugo@80-62-116-101-mobile.dk.customer.tdc.net) (Read error: Connection reset by peer)
2020-10-31 14:05:53 <merijn> CS writing question time: What's the right way to refer to "the property of being memory bound"? Memory boundedness? Any better suggestions?
2020-10-31 14:06:07 → nschoe joins (~quassel@2a01:e0a:3c4:c7b0:b8f0:6b21:1cae:fc94)
2020-10-31 14:06:07 → idhugo joins (~idhugo@80-62-116-101-mobile.dk.customer.tdc.net)
2020-10-31 14:06:36 → nbloomf joins (~nbloomf@2600:1700:ad14:3020:c427:c5ca:d62:565b)
2020-10-31 14:08:27 × ahmr88 quits (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net) (Remote host closed the connection)
2020-10-31 14:09:43 <jkaye[m]> I think it's just "memory bound"
2020-10-31 14:09:46 <jkaye[m]> What's an example of how you'd want to use it?
2020-10-31 14:12:07 <merijn> jkaye[m]: "something is memory bound" and I need something more along the lines of "because of X being memory bound..." but that's very long and wordy in context. I'm not sure how I want to use it yet, I'm trying to figure out a non-awkward wording :p
2020-10-31 14:13:07 <pjb> merijn: X being memory bound is X being bloated.
2020-10-31 14:13:46 <jkaye[m]> Really? You believe there are no memory-bound tasks in the world? Stop trolling.
2020-10-31 14:14:12 <pjb> You can always trade time for space.
2020-10-31 14:14:16 <jkaye[m]> I do think that "memory bound" is probably the most succinct you're going to get, at least that I've heard that would be easily recognizable
2020-10-31 14:14:27 <pjb> Agreed.
2020-10-31 14:14:52 <jkaye[m]> Trading time for space is how you solve a memory-bound problem :)
2020-10-31 14:15:28 × rusua quits (uid124537@gateway/web/irccloud.com/x-eqeurrcaunytkiar) (Quit: Connection closed for inactivity)
2020-10-31 14:16:09 × Franciman quits (~francesco@host-79-36-167-172.retail.telecomitalia.it) (Quit: Leaving)
2020-10-31 14:16:15 → wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net)
2020-10-31 14:21:46 → p-core joins (~Thunderbi@2001:718:1e03:5128:2ab7:7f35:48a1:8515)
2020-10-31 14:22:36 × jakob_ quits (~textual@p200300f49f1622007caf13ec66be0ccb.dip0.t-ipconnect.de) (Quit: My Laptop has gone to sleep. ZZZzzz…)
2020-10-31 14:23:04 → Amras joins (~Amras@unaffiliated/amras0000)
2020-10-31 14:23:32 <merijn> Eh, hard disagree
2020-10-31 14:23:38 <merijn> memory bound is about space
2020-10-31 14:23:49 → kmwyard joins (uid471732@gateway/web/irccloud.com/x-tmjuaykfgmfaiqvc)
2020-10-31 14:23:50 <merijn> It's whether you are bottle-necked by memory transfer speeds
2020-10-31 14:23:50 <nshepperd2> PSPACE is not equal to EXPSPACE, so i don't think you can always trade time for space
2020-10-31 14:24:30 <merijn> If you are bottle-necked by memory transfer speed you can't necessarily trade that for time, definitely not without making things worse in terms of performance
2020-10-31 14:25:03 <merijn> nshepperd2: That's talking about complexity anyway, and while complexity analysis is interesting, it's also mostly intellectual masturbation in terms of practical HPC :)
2020-10-31 14:25:06 × jespada quits (~jespada@90.254.243.98) (Ping timeout: 258 seconds)
2020-10-31 14:25:40 × petersen quits (~petersen@redhat/juhp) (Ping timeout: 246 seconds)
2020-10-31 14:26:07 <nshepperd2> true, but i think it does suggest that there are problems that objectively take more space to solve than others
2020-10-31 14:26:33 <merijn> nshepperd2: I don't know PSPACE and EXPSPACE enough agree or disagree ;)
2020-10-31 14:29:37 → LKoen joins (~LKoen@lstlambert-657-1-123-43.w92-154.abo.wanadoo.fr)
2020-10-31 14:29:37 <pjb> merijn: performace is a relative term. if memory bound is a problem it means your space performace is bad. You need to take less space! And a way to do that, is to lose time to compress, or to recompute what is stored, etc.
2020-10-31 14:30:54 <merijn> eh, no
2020-10-31 14:31:10 <merijn> compute vs memory bound is about bottlenecks in roofline models
2020-10-31 14:31:27 → jespada joins (~jespada@90.254.243.98)
2020-10-31 14:31:31 <pjb> I take memory bound here as meaning using too much memory. But it may be doing many memory exchanges. CPU bound vs memory-bus bound vs. I/O bound.
2020-10-31 14:31:36 <merijn> You can't solve being memory bound by "being slower and using less space", because then performance goes even lower
2020-10-31 14:32:07 <merijn> pjb: That's fine, but then you're taking memory bound to mean something different from literally everyone in HPC/performance oriented fields
2020-10-31 14:32:18 <pjb> In that case, you can up to a certain measure do some computation in the CPU (with registers) to avoid accessing the memory, but there's not much you can do in general.
2020-10-31 14:32:39 → mirrorbird joins (~psutcliff@2a00:801:3f2:56a2:5835:27e4:34fc:e9d)
2020-10-31 14:33:12 <merijn> Anyway, this discussion is woefully unhelpful to my original question >.>
2020-10-31 14:33:31 <pjb> It's a good thing to be memory bound in a way… Look at the alternatives! If you put the data in secondary memory, then you can become I/O bound. Not necessarily better.
2020-10-31 14:34:20 → Nachtgespenst joins (~user@unaffiliated/siracusa)
2020-10-31 14:34:28 → carlomagno joins (~cararell@148.87.23.13)
2020-10-31 14:34:28 <pjb> perhaps you could shorten it as "unbalanced". As in "unbalanced architecture" where the processor is too powerfull compared to the memory bus?
2020-10-31 14:35:06 <pjb> or unbalanced as in lacking cache memory vs. main memory?
2020-10-31 14:38:22 → urodna joins (~urodna@unaffiliated/urodna)
2020-10-31 14:38:31 × taurux quits (~taurux@net-130-25-101-121.cust.vodafonedsl.it) (Ping timeout: 258 seconds)
2020-10-31 14:38:41 → bennofs1 joins (~benno@dslb-188-106-255-000.188.106.pools.vodafone-ip.de)
2020-10-31 14:39:30 nshepperd2 . o O ("memory bound" is when your ROI of performance per $ of additional memory is > performance per $ of processors)
2020-10-31 14:40:38 → taurux joins (~taurux@net-130-25-101-160.cust.vodafonedsl.it)
2020-10-31 14:41:42 <merijn> nshepperd2: In HPC roofline modelling the machine is usually a given, so you try and identify where you're bottlenecked so you can deal with that. If your implementation's memory bandwidth usage is equal to the max bandwith of your hardware, you are memory bound and any optimisation of compute is pointless, since it can't speed anything up
2020-10-31 14:41:47 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-10-31 14:42:06 <nshepperd2> makes sense
2020-10-31 14:42:21 <merijn> nshepperd2: If you're compute bound that means your implementation's FLOPS (traditionally, but that's a shitty metric for no floating point stuff) approaches theoretical peak FLOPS
2020-10-31 14:43:41 × kish quits (~oracle@unaffiliated/oracle) (Remote host closed the connection)
2020-10-31 14:43:56 <merijn> Although in parallel systems you also get stuff like occupancy and efficiency messing things up :p
2020-10-31 14:43:59 → tenniscp25 joins (~tenniscp2@134.196.209.118)
2020-10-31 14:44:12 × FreeBirdLjj quits (~freebirdl@101.228.42.108) (Remote host closed the connection)
2020-10-31 14:44:14 <pjb> square[i]=f(i); loop(g(square[i])) can be optimized and lighten the memory load with loop(g(f(i))) if f(i) can be computed in the CPU.
2020-10-31 14:44:30 <pjb> Ie. remove temp variables, and let the compiler do its job.
2020-10-31 14:44:37 × Axman6 quits (~Axman6@pdpc/supporter/student/Axman6) (Ping timeout: 264 seconds)
2020-10-31 14:44:59 → kish joins (~oracle@unaffiliated/oracle)
2020-10-31 14:46:38 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Ping timeout: 260 seconds)
2020-10-31 14:47:59 → conal joins (~conal@64.71.133.70)
←Prev  Next→
Page 1 .. 973 974 975 976 977 978 979 980 981 982 983 .. 5022

All times are in UTC.