Open BenCasses opened 7 years ago
Do I want to stat the log file before opening it and give up if size is above some threshold to avoid parsing huge stack trace logs?
@BenCasses I have been trying to run this plugin this morning and I cannot get it to print any results.
$ pav table
Pavilion Version -> 1.1.14
Running from -> /users/dejager/swlt/Pavilion/PAV
Logging to -> /tmp/dejager/pav.log
Invoke command: -> table
test-name result start-time end-time
I have tried fiddling with the start time and fields option but it should not matter because I ran the tests Thursday.
Here is output from get_results (I used baseline to shorten it).
$ pav get_results -bl
Pavilion Version -> 1.1.14
Running from -> /users/dejager/swlt/Pavilion/PAV
Logging to -> /tmp/dejager/pav.log
...
hpl_intel_1x.25x900(10):
Gflops = 1008.86 Gflops (entries - 750, min - 951.96, max - 1040.08, stDev - 13.68
Time = 3.32 seconds (entries - 750, min - 3.22, max - 3.51, stDev - 0.04
hpl_intel_1x.25x900(256):
Gflops = 1007.03 Gflops (entries - 2500, min - 933.54, max - 1039.25, stDev - 14.53
Time = 3.32 seconds (entries - 2500, min - 3.22, max - 3.58, stDev - 0.05
First guess is that its rejecting all results, run it with a "-v" flag and It should tell you why, for each .log file, it is rejecting.
new plugin : table prints a configurable table of results either looking from cwd or from the results location of a test suite -experimentally runs faster than get_results -a few fields are offered : name, nodes, result, etc., or the user can request any line starting with as a field
-start and end times are both intervals here : filter for tests that "started" (on the path) within the start interval and "ended" within the end interval ( tag or "cleanup" line). One-sided intervals are assumed to start 15 days ago and end now.
-no perl or subprocess