2021-08-06 22:55
hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar-group.org | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | This channel is logged: irclog.cct.lsu.edu
00:22
tufei_ has joined #ste||ar
00:23
tufei has quit [Remote host closed the connection]
00:30
hkaiser has quit [Read error: Connection reset by peer]
03:32
K-ballo1 has joined #ste||ar
03:34
K-ballo has quit [Ping timeout: 260 seconds]
03:34
K-ballo1 is now known as K-ballo
04:32
sivoais has quit [*.net *.split]
04:32
sivoais has joined #ste||ar
06:27
<
satacker[m] >
It does
07:56
Yorlik has joined #ste||ar
09:39
<
satacker[m] >
I ODR used the std::declval which is wrong
09:40
<
satacker[m] >
Sending a PR
10:15
bhumit[m] has joined #ste||ar
12:53
hkaiser has joined #ste||ar
13:22
<
pansysk75[m] >
on a machine with 2 numa domains (20 + 20 cores), how can i launch hpx to run on just on the cores of one numa domain?
13:23
<
pansysk75[m] >
is it --hpx::threads 20 and --hpx::affinity numa?
13:24
<
hkaiser >
just the first
13:24
<
hkaiser >
the second is not needed
13:24
<
hkaiser >
--hpx:print-bind will show you what cores are used
13:25
<
pansysk75[m] >
great, thank you!
13:25
<
pansysk75[m] >
how about if I want to launch on 20 cores that are not within the same numa?
13:25
<
pansysk75[m] >
is there any way to select, with core IDs or with a bit mask or sth?
13:28
<
pansysk75[m] >
nice!
15:36
tufei_ has quit [Quit: Leaving]
15:43
tufei has joined #ste||ar
17:59
diehlpk has joined #ste||ar
19:06
K-ballo has quit [Ping timeout: 240 seconds]
19:06
K-ballo1 has joined #ste||ar
19:08
K-ballo1 is now known as K-ballo
19:21
diehlpk has quit [Quit: Leaving.]