trampgeek
commented
1 year ago Hi Björn and Sebastian
I'm assuming you're getting these failures when using Jobe from CodeRunner, right?
This does sound to me like a memory error. numpy is greedy and I recently raised the default memory allocation for Python to 1000 because of numpy. When the memory limit is reached, the call to malloc returns null and much/most software never checks this. So the code can easily get stuck in an endless loop trying to allocate enough memory.
So ... I'm wondering how you set the memory limit. Did you perhaps change the default setting within Jobe itself? If that's the case, it's possible that that default is being overridden somewhere else along the line, e.g. by the question itself or its prototype
Try customising a CodeRunner question that's failing (click "Customise" near the top of the question-authoring form), opening the Advanced customisation section, and entering 1000 into the Memory Limit field.
If that doesn't help, post back and I'll think further on it.
Richard
From: Björn Fjällström @.> Sent: Saturday, 30 September 2023 3:35 am To: trampgeek/jobe @.> Cc: Subscribed @.**> Subject: [trampgeek/jobe] Time limit exceeded * when running python3 scripts with numpy (Issue #66)
We have a debian installation of jobe, that gets time limit exceeded when importing numpy.
lsb_release -a
No LSB modules are available. Distributor ID: Debian Description: Debian GNU/Linux 12 (bookworm) Release: 12 Codename: bookworm
free -m
total used free shared buff/cache availableMem: 8902 583 8239 7 342 8319 Swap: 1955 0 1955
We have 16 processors on this VM.
We have tired to raise the limits of memory to 2000 for python, and CPU runtime with no other result that the tests take longer to respond.
According to debugging in README.md:
" If something unexpected happened with the actual run of a program, find the run in /home/jobe/runs and try executing the program manually. [The run directory contains the source file, the bash command used to run it, plus the compile output and (if it ran) the stderr and stdout outputs. "
We have tried to catch the program and run them in CLI on the server: while :;do time python3 jobe_/__test ;echo $?; sleep 1;done
Output from out program: Test [420/8991] [1, 3, 5, 7, 9, 2, 4, 6, 8, 10]
real 0m0.166s user 0m0.931s sys 0m0.883s 0
Where the program is: " import numpy as np def task12(param1, param2): print("Test") return [1, 3, 5, 7, 9, 2, 4, 6, 8, 10] "
testsubmit.py is only complaining on c, witch does not seem relevant to our case.
***** FAILED TEST ** (8 results) [3360/8998]
{'run_id': None, 'outcome': 11, 'cmpinfo': 'prog.c: In function ‘silly’:\nprog.c:4:6: error: infinite rec
ursion detected [-Werror=infinite-recursion]\n 4 | void silly(int i) {\n | ^~~~~\nprog.c:7:9
: note: recursive call\n 7 | silly(j);\n | ^~~~\nprog.c:9:9: note: recursive
call\n 9 | silly(j + 1);\n | ^~~~\ncc1: all warnings being treated as err
ors\n', 'stdout': '', 'stderr': ''}
Infinite recursion (stack error) on C
Jobe result: Compile error
Compiler output:
prog.c: In function ‘silly’:
prog.c:4:6: error: infinite recursion detected [-Werror=infinite-recursion]
4 | void silly(int i) {
| ^~~~~
prog.c:7:9: note: recursive call
7 | silly(j);
| ^~~~
prog.c:9:9: note: recursive call
9 | silly(j + 1);
| ^~~~
cc1: all warnings being treated as errors
We don't understand why it completes quickly in CLI, but times out in moodle. Any suggestions or tips?
Regards Björn and Sebastian DSV Stockholm university
— Reply to this email directly, view it on GitHubhttps://github.com/trampgeek/jobe/issues/66, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAIM2VTGYA3M5IH5B25IA2TX43MCDANCNFSM6AAAAAA5MSFWQI. You are receiving this because you are subscribed to this thread.Message ID: @.***>
This email may be confidential and subject to legal privilege, it may not reflect the views of the University of Canterbury, and it is not guaranteed to be virus free. If you are not an intended recipient, please notify the sender immediately and erase all copies of the message and any attachments.
spetzreborn
We have a debian installation of jobe, that gets time limit exceeded when importing numpy.
lsb_release -a
No LSB modules are available. Distributor ID: Debian Description: Debian GNU/Linux 12 (bookworm) Release: 12 Codename: bookworm
free -m
Mem: 8902 583 8239 7 342 8319 Swap: 1955 0 1955
We have 16 processors on this VM.
We have tired to raise the limits of memory to 2000 for python, and CPU runtime with no other result that the tests take longer to respond.
According to debugging in README.md:
" If something unexpected happened with the actual run of a program, find the run in /home/jobe/runs and try executing the program manually. [The run directory contains the source file, the bash command used to run it, plus the compile output and (if it ran) the stderr and stdout outputs. "
We have tried to catch the program and run them in CLI on the server: while :;do time python3 jobe_/__test ;echo $?; sleep 1;done
Output from out program: Test [420/8991] [1, 3, 5, 7, 9, 2, 4, 6, 8, 10]
real 0m0.166s
user 0m0.931s
sys 0m0.883s
0
Where the program is: " import numpy as np def task12(param1, param2): print("Test") return [1, 3, 5, 7, 9, 2, 4, 6, 8, 10] "
testsubmit.py is only complaining on c, witch does not seem relevant to our case.
***** FAILED TEST ** (8 results) [3360/8998]
{'run_id': None, 'outcome': 11, 'cmpinfo': 'prog.c: In function ‘silly’:\nprog.c:4:6: error: infinite rec ursion detected [-Werror=infinite-recursion]\n 4 | void silly(int i) {\n | ^~~~~\nprog.c:7:9 : note: recursive call\n 7 | silly(j);\n | ^
~~~\nprog.c:9:9: note: recursive call\n 9 | silly(j + 1);\n | ^~~~\ncc1: all warnings being treated as err ors\n', 'stdout': '', 'stderr': ''}Infinite recursion (stack error) on C
Jobe result: Compile error
Compiler output:
prog.c: In function ‘silly’:
prog.c:4:6: error: infinite recursion detected [-Werror=infinite-recursion]
4 | void silly(int i) {
| ^~~~~
prog.c:7:9: note: recursive call
7 | silly(j);
| ^
~~~ prog.c:9:9: note: recursive call 9 | silly(j + 1); | ^~~~ cc1: all warnings being treated as errorsWe don't understand why it completes quickly in CLI, but times out in moodle. Any suggestions or tips?
Regards Björn and Sebastian DSV Stockholm university