Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2013-03-09

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

All times shown according to UTC.

Time Nick Message
00:21 MeierM joined #pdl
12:03 dimuls joined #pdl
12:04 dimuls hi guys
12:04 dimuls does pdl support mmx?
12:04 dimuls or other similar technologies
12:04 dimuls &
12:04 dimuls ?
12:04 dimuls cuda?
12:31 dimuls joined #pdl
14:02 jberger run4flat, ping for dimuls
14:07 dimuls here
14:07 dimuls jberger: pong
14:07 jberger run4flat is really the one with the knowledge on these things
14:08 jberger :-)
14:08 dimuls thanx:3
14:08 jberger I can maybe help you find some links/modules if you tell me what you are trying to do
14:08 jberger but I can't say I'm really that knowledgable in this area
14:09 dimuls i'm working on tf-idf documents representation model
14:09 jberger ... and you lost me :-)
14:09 dimuls i need to compute some matrixes faster
14:09 jberger ok, that I get
14:09 dimuls don't worry
14:10 jberger there has been recent work on PDL threading etc
14:10 dimuls yes, i know
14:10 dimuls i know how to use it
14:10 jberger oh ok
14:10 jberger in that case, you really need run4flat
14:10 jberger :-)
14:11 dimuls but i thing, if pdl will use some cpu vector computing tools it will work much faster
14:12 dimuls *features
14:12 dimuls run4flat: ping
14:35 run4flat_ joined #pdl
14:36 run4flat_ hello, everyone
14:36 * run4flat_ reads the log
14:36 run4flat_ ah, mmx and cuda
14:36 run4flat_ dimuls, PDL does not directly support mmx
14:37 run4flat_ it does not, and can not, directly support CUDA
14:37 run4flat_ it is possible to shuttle PDL data to the video card and invoke custom CUDA kernels, however
14:37 run4flat_ which is something I have done in the past
14:38 run4flat_ as for mmx, it is a technology that I have only recently heard about and know nothing about
14:38 run4flat_ but I suspect it is something that we (PDL developers) should look into using
14:39 run4flat_ dimuls, at the moment, PDL's matrix multiplication only runs on one thread, which kinda kills me
14:40 run4flat_ wow, looks like mmx technology has been around for over 10 years
15:16 Molaf__ joined #pdl
17:32 dimuls joined #pdl
19:16 jberger joined #pdl
20:22 chm joined #pdl
20:24 chm ping run4flat
22:57 Meiermann joined #pdl

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