ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at http://www.mlpack.org/irc/
< AlexNguyenGitter>
Hi, since this pr is up-to-date with the master branch, #2807, and it seems to be in good shape! It is great if you can have it merged! Thanks!
halaldeen-ms has joined #mlpack
halaldeen-ms has quit [Quit: Leaving]
rahul_shaw has joined #mlpack
rahul_shaw has quit [Quit: Connection closed]
ImQ009 has joined #mlpack
< rcurtin[m]>
wow, just switching from `liblapack-dev` and `libblas-dev` (the reference implementation) to `libopenblas-dev` makes the ensmallen tests run 3-4x faster, even in singlethreaded usage
< shrit[m]>
Yes,
< shrit[m]>
I have seen x16 times performance increase when I switched to OpenBLAS
< shrit[m]>
on all threads
< shrit[m]>
I wonder that using this combination it might beat all other framworks even if they are running GPU
< zoq>
If you can keep everything on the GPU I don't think so.
tharun has joined #mlpack
VENKY has joined #mlpack
tharun has quit [Quit: Connection closed]
VENKY has quit [Quit: Connection closed]
< shrit[m]>
Yeah, there will be a day where the transfere speed between all compenents will be equally the same
ImQ009 has quit [Quit: Leaving]
< rcurtin[m]>
impressive explosion from SpaceX today :)
< zoq>
yep, the start of the flip looked so goood
< rcurtin[m]>
hopefully SN10 isn't damaged by the fireball! I was surprised they had them so close
< zoq>
same, maybe they thought there is plenty of room