![]() |
|
|
#694 |
|
Feb 2019
2·32·5 Posts |
|
|
|
|
|
|
#695 |
|
If I May
"Chris Halsall"
Sep 2002
Barbados
2×67×73 Posts |
OK, just to document the empirical weirdness ("i" before "e", except after "c") of Colab.
I just experienced the "Ten hours" cut-off thing on two of my local instances. Did a restart on both, and was given two new P100s, with each VM reporting an uptime of 1 minute. Going up for dinner and some quality time with my SHMBO. Will be interesting to see how long these survive. |
|
|
|
|
|
#696 |
|
If I May
"Chris Halsall"
Sep 2002
Barbados
2×67×73 Posts |
OK, they lasted about 3:50 each. And new behavior again this morning...
I spun up three instances, and two got P100s. Interestingly, the one which got a K80 wouldn't let me do the "Runtime" - > "Manage sessions" trick to "upgrade". The session was listed, but the "Terminate" link was greyed out. I /was/ given the option on the P100 instances. |
|
|
|
|
|
#697 |
|
"Eric"
Jan 2018
USA
22×53 Posts |
I figured out a way to fix the Runtime Disconnected problem when Colab requires a reconnect within 20 minutes (!! used to be an hour). You basically click the RAM+CPU bar dropdown, then click Connect to hosted runtime. You can tell that it is gonna last when you see three green dots and busy if your instance is crunching numbers.
|
|
|
|
|
|
#698 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
2·19·101 Posts |
For the stuff I'm playing with, I'd prefer the least favorite GPU available. I often get a P100. Is there a way to change that to a K80 "on purpose?"
|
|
|
|
|
|
#699 | |
|
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
5,437 Posts |
Quote:
|
|
|
|
|
|
|
#700 | |
|
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
124758 Posts |
Quote:
Last fiddled with by kriesel on 2019-12-15 at 22:10 |
|
|
|
|
|
|
#701 | |
|
"Eric"
Jan 2018
USA
22×53 Posts |
Quote:
|
|
|
|
|
|
|
#702 | |
|
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
5,437 Posts |
Quote:
!lscpu !nvidia-smi at the beginning of the script, so I see which gpu model was allocated, before using it or logging in to google drive. So now sometimes I get none: NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running. Perhaps it's a regional thing, but from here it appears as if there are no longer any K80s to be had. They've been getting progressively harder to get a session on over weeks. Last fiddled with by kriesel on 2019-12-15 at 23:28 |
|
|
|
|
|
|
#703 |
|
"Tony Gott"
Aug 2002
Yell, Shetland, UK
1010011002 Posts |
Hmm, just got 2 K80s this morning ...
|
|
|
|
|
|
#704 |
|
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
5,437 Posts |
Now I'm repeatedly getting no GPU at all. On the account from which I made many unsuccessful tries to get a K80:
Code:
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running. Last fiddled with by kriesel on 2019-12-16 at 08:40 |
|
|
|
![]() |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Alternatives to Google Colab | kriesel | Cloud Computing | 11 | 2020-01-14 18:45 |
| Notebook | enzocreti | enzocreti | 0 | 2019-02-15 08:20 |
| Computer Diet causes Machine Check Exception -- need heuristics help | Christenson | Hardware | 32 | 2011-12-25 08:17 |
| Computer diet - Need help | garo | Hardware | 41 | 2011-10-06 04:06 |
| Workunit diet ? | dsouza123 | NFSNET Discussion | 5 | 2004-02-27 00:42 |