aboutsummaryrefslogtreecommitdiff
path: root/bench.py
Commit message (Collapse)AuthorAgeFilesLines
* [bench] exit with status 1 on failureFlorian Fischer2021-04-071-1/+8
|
* [bench] remove 'all' allocator default targetFlorian Fischer2021-04-071-2/+1
|
* add simple integration test with a new dummy benchmark and the system ↵Florian Fischer2020-07-191-4/+6
| | | | default allocator
* [bench.py] don't summarize for each allocatorFlorian Fischer2020-07-191-2/+2
|
* [bench.py] import AVAIL_BENCHMARKS from allocbench.benchmarkFlorian Fischer2020-07-141-3/+3
|
* fixup! [bench.py] use util.set_verbosityFlorian Fischer2020-07-141-1/+0
|
* [bench.py] use util.set_verbosityFlorian Fischer2020-07-141-9/+3
|
* remove globalvars and pass allocators to Benchmark.runFlorian Fischer2020-07-141-3/+2
|
* [logging] call getLogger only with the basename of the module filesFlorian Fischer2020-07-131-2/+2
|
* [refactoring] fix or ignore many pylint hintsFlorian Fischer2020-07-131-0/+1
| | | | | | | | | * Add missing docstrings * Remove unused imports * fix line-to-long * fix some typos * ignore global-statement * ignore wrong-import-order for scripts under /scripts
* replace print_* functions from allocbench.util if loggingFlorian Fischer2020-07-131-29/+34
|
* move VERBOSITY from globalvars to utilFlorian Fischer2020-07-131-2/+3
|
* [globalvars] remove list of available benchmarks from globalvars.BENCHMARK ↵Florian Fischer2020-07-111-2/+2
| | | | to benchmark.AVAIL_BENCHMARKS
* replace directory string constants in globalvars with Path objects from ↵Florian Fischer2020-07-111-18/+21
| | | | allocbench.directory
* [pylint] disable some warningsFlorian Fischer2020-06-021-3/+3
|
* [benchmarks] reintroduce prepareFlorian Fischer2020-06-021-8/+7
| | | | | | Always running the possibly very expensive preparation code when a benchmark object is created is not reasonable when benchmarks should only be summarized.
* [benchmark.py] add function to retrieve Benchmark instance from a moduleFlorian Fischer2020-06-021-10/+8
|
* bench.py: fix benchmark checkFlorian Fischer2020-06-021-1/+1
|
* make benchmark definitions more pythonicFlorian Fischer2020-06-021-14/+15
| | | | | | Don't export a singleton object in the definiton. An object is now created when the Benchmark should be executed, thus we don't need a prepare method this can be done in __init__
* make the whole project more python idiomaticFlorian Fischer2020-06-021-41/+34
| | | | | | | | | * rename src directory to allocbench * make global variable names UPPERCASE * format a lot of code using yapf * use lowercase ld_preload and ld_library_path as Allocator members * name expected Errors 'err' and don't raise a new Exception * disable some pylint messages
* implement analysis of specific allocator and benchmark combinationsFlorian Fischer2020-02-091-40/+8
| | | | A allocator implementation is analyzed by prepending chattymalloc.so.
* use argparse version action to print allocbench versionFlorian Fischer2020-01-271-5/+2
|
* add error message for unknown or misspelled benchmarksFlorian Fischer2020-01-231-0/+5
|
* readd option to summarize after benchmark runFlorian Fischer2019-12-281-4/+10
|
* use sys.exit instead of exit and other small code improvementsFlorian Fischer2019-12-141-8/+7
|
* move allocbench_version to facter and make the facts a global variable of ↵Florian Fischer2019-12-141-5/+6
| | | | facter.py
* yapf format and remove old hacky output codeFlorian Fischer2019-12-111-21/+47
|
* change result format from pickle to jsonFlorian Fischer2019-11-231-2/+2
|
* always print path to result directoryFlorian Fischer2019-11-221-1/+1
|
* let facter handle load/store of global factsFlorian Fischer2019-11-221-3/+1
|
* remove left over status file at exitFlorian Fischer2019-11-131-0/+4
| | | | | If writing the status file fails because of not sufficient space left successive runs can fail because of a broken status file.
* remove unused chattyparser importFlorian Fischer2019-11-061-1/+0
|
* don't measure analyse runsFlorian Fischer2019-09-211-0/+4
|
* store start, end and duration for every benchmarkFlorian Fischer2019-09-141-0/+5
|
* don't exit if we fail to prepare a benchmarkFlorian Fischer2019-09-111-1/+7
| | | | | benchmark.prepare should raise an Exception if it encounters an unrecoverable, intolerable error.
* abort if there are no allocators to benchmarkFlorian Fischer2019-09-111-0/+4
|
* add --version flag to executablesFlorian Fischer2019-08-301-1/+5
|
* make all default allocator targetFlorian Fischer2019-08-281-1/+2
|
* simplify bench.py codeFlorian Fischer2019-08-281-180/+105
| | | | | | | | | | | bench.py does only what it's name say: run benchmarks. Summarizing results is done with summarize.py. Merging two saves is done using merge.py. Misc: * -vdebug is now -vvv * license header is added * replace logger calls with print_error
* remove -vdebug flagFlorian Fischer2019-08-261-6/+2
|
* only plot chattymalloc results if nosum is not setFlorian Fischer2019-08-251-1/+1
|
* add result directory to trace fileFlorian Fischer2019-08-251-1/+1
|
* fix chattymalloc nameFlorian Fischer2019-08-251-1/+1
|
* s/analyse/analyzeFlorian Fischer2019-08-251-16/+16
|
* fix analyse chattymalloc analyzingFlorian Fischer2019-08-251-1/+1
|
* plot traces produced by chattymallocFlorian Fischer2019-08-241-1/+7
|
* support chattymalloc as analyze fallback if malt is not availableFlorian Fischer2019-08-221-33/+32
| | | | | | bench.py saves the benchmark specific result directory in the object member "result_dir". This member is used by the analyze allocators to know where to write the trace files to.
* remove duplicate check if we really should run the benchmarkFlorian Fischer2019-08-221-2/+2
|
* load benchmarks using importlib instead of evalFlorian Fischer2019-08-121-1/+5
|
* add requirement check for matplotlib and python version >= 3.6Florian Fischer2019-08-121-5/+16
|
generated by cgit v1.2.3 (git 2.25.1) at 2026-04-25 12:23:30 +0000