kale[m] has quit [Read error: Connection reset by peer]
kale[m] has joined #ste||ar
kale[m] has quit [Ping timeout: 260 seconds]
kale[m] has joined #ste||ar
bita_ has quit [Ping timeout: 260 seconds]
<rori>
freenode_gonidelis[m]: I don't have the link in my bookmarks anymore, but let me know if you have trouble finding it, I might have a look
nikunj97 has joined #ste||ar
Nikunj__ has joined #ste||ar
nikunj97 has quit [Ping timeout: 240 seconds]
kale[m] has quit [Ping timeout: 260 seconds]
kale[m] has joined #ste||ar
kale[m] has quit [Ping timeout: 260 seconds]
kale[m] has joined #ste||ar
kale[m] has quit [Read error: Connection reset by peer]
kale[m] has joined #ste||ar
Nikunj__ is now known as nikunj97
hkaiser has joined #ste||ar
kale[m] has quit [Ping timeout: 256 seconds]
kale[m] has joined #ste||ar
<nikunj97>
heller1, jbjnr, hkaiser, ms[m]: The Arm paper is almost ready. Let me know if you want to have a look at it. We're thinking to submit it to EAHPC 2020 at IEEE Cluster. Still need permission from Fujitsu to make the paper public though :/
<hkaiser>
nikunj97: I'd love to see it
<hkaiser>
nikunj97: also: we'll get access to a machine that has the same architecture as the new Japanese machine ARM64fx
<hkaiser>
would it make sense to run your benchmarks there?
<hkaiser>
(note: vectorization is not NEON)
<nikunj97>
hkaiser, sure. We can!
<nikunj97>
yea, I expect it to be SVE
<hkaiser>
64fx
<hkaiser>
also, see pm, pls
nanmiao11 has joined #ste||ar
diehlpk__ has joined #ste||ar
weilewei has joined #ste||ar
<parsa>
kale[m]: ping
Yorlik has joined #ste||ar
diehlpk_work has joined #ste||ar
diehlpk__ has quit [Ping timeout: 260 seconds]
bita_ has joined #ste||ar
karame_ has joined #ste||ar
<kale[m]>
parsa: pong
<nanmiao11>
lol
diehlpk__ has joined #ste||ar
kale[m] has quit [Ping timeout: 260 seconds]
kale[m] has joined #ste||ar
rtohid has joined #ste||ar
diehlpk__ has quit [Ping timeout: 260 seconds]
karame_ has quit [Remote host closed the connection]
kale[m] has quit [Ping timeout: 246 seconds]
kale[m] has joined #ste||ar
rtohid has quit [Remote host closed the connection]
kale[m] has quit [Ping timeout: 244 seconds]
kale[m] has joined #ste||ar
weilewei has quit [Remote host closed the connection]
nanmiao11 has quit [Remote host closed the connection]
nanmiao11 has joined #ste||ar
diehlpk__ has joined #ste||ar
rtohid has joined #ste||ar
karame_ has joined #ste||ar
weilewei has joined #ste||ar
rajpalsanyam1997 has quit [*.net *.split]
klaus[m]1 has quit [*.net *.split]
klaus[m]1 has joined #ste||ar
rajpalsanyam1997 has joined #ste||ar
diehlpk__ has quit [Ping timeout: 260 seconds]
diehlpk__ has joined #ste||ar
diehlpk__ has quit [Ping timeout: 260 seconds]
kale[m] has quit [Ping timeout: 240 seconds]
kale[m] has joined #ste||ar
<gonidelis[m]>
hkaiser: just submited the evaluation ;)
<K-ballo>
do we know who LiliumAtratum is?
<jbjnr>
no
<zao>
Developer on some mysterious CAD software :)
Nikunj__ has joined #ste||ar
nikunj97 has quit [Ping timeout: 260 seconds]
<weilewei>
on Piz Daints, if I would like to have a mpi wrapper module to gcc, what module should I load?
<weilewei>
(day 0 on Piz Daints...)
<zao>
weilewei: `module load daint-mc` or `module load daint-gpu` and look around with `module avail`, maybe?
<zao>
(I've never been on a CSCS system)
<weilewei>
zao right I did `module load daint-gpu` but I only see gcc modules but not sure which one is mpi wrapper to it
Nikunj__ has quit [Read error: Connection reset by peer]
<zao>
There's no `mpicc` or similar in path? Look for something MPICH otherwise?
<weilewei>
yea, maybe tmr when jbjnr wakes up, he will know...
<weilewei>
I am out of clues as well...
<weilewei>
:') Daint has so many modules
<weilewei>
comparing Summit...
<weilewei>
Summit modules are more straigtfowrad to use, to my personal feeling
<zao>
You should see our production clusters, I think we have thousands of versions of packages installed.
<weilewei>
X) that's too many...
<zao>
Mine is vanilla EasyBuild tho, so very unsurprising compared to the Cray stuff.
<weilewei>
haven't used EasyBuild though
<weilewei>
Does users really need that much libraries?
<zao>
The wonders of supporting any little PhD that wants to run on the resource, there's a lot of weird stuff, especially ML/AI things tend to spawn a lot of Python packages.
<weilewei>
well, that makes sense... if everyone asks for a library
<zao>
We've got new toolchains twice a year, with new versions of libraries and software with them as needed.
<zao>
The R bundle has something like 750 extensions in it alone.
<weilewei>
hmmm, I would say too many is better than nothing
<weilewei>
yesterday, I compiled code in a new cluster with limited modules, I need to compile everything from scratch
<weilewei>
and login node (no gpu) is different environment than compute node... and compute node has no internet connection, so I need to git pull things from login node, and then go to compute node to compile
<zao>
No libraries available either, or just things like autodetect of compute capability and test suites that fail?
<weilewei>
there are some libraries, but not enough to compile my code