Commit graph

56 commits

Author SHA1 Message Date
jaffee
a03ecf92e9 fix rebase issues 2016-12-19 15:12:56 -06:00
jaffee
c720ddc28e few more json fixes 2016-12-19 15:10:35 -06:00
jaffee
6bb0d21e65 define json for all benchmarks 2016-12-19 15:10:35 -06:00
jaffee
293621516e have bagent add metadata to results
need this to differentiate multiple agents' output when it's all going to stdout.
2016-12-19 15:10:35 -06:00
Matthew Jaffee
e211dbd54a Merge pull request #199 from alanbernstein/benchmarks-readable
Use Duration's string representation in JSON output
2016-12-19 14:50:53 -06:00
Matthew Jaffee
58f697d814 Merge pull request #200 from alanbernstein/benchmarks-zipf
Benchmarks zipf
2016-12-19 14:21:05 -06:00
Alan Bernstein
9ae006908a Improve comment formatting 2016-12-19 13:56:06 -06:00
Matthew Jaffee
bde6f03c81 Merge pull request #208 from jaffee/171-cluster-creation
#171 cluster creation
2016-12-19 11:12:38 -06:00
Matthew Jaffee
d8f5b7d78b Merge pull request #174 from jaffee/fix-errgroup-dep
copy errgroup code to bench and remove giant blob of dependencies
2016-12-16 13:55:09 -06:00
Alan Bernstein
44a738afde Add more prettify tests 2016-12-13 14:21:49 -06:00
Alan Bernstein
2968c15a37 Improve comment formatting and update usage 2016-12-13 13:25:31 -06:00
jaffee
6829f7cdd0 create - start cluster over ssh when given hosts 2016-12-13 10:14:09 -06:00
Alan Bernstein
356f1cfd17 Add test for prettify 2016-12-12 22:17:56 -06:00
Alan Bernstein
da3bb99f0a Generalize Prettify function 2016-12-12 22:10:07 -06:00
Alan Bernstein
d08ffd20e2 Move prettify to own file 2016-12-12 16:57:21 -06:00
Alan Bernstein
cbd7b2ff37 Add comments 2016-12-12 15:55:52 -06:00
Alan Bernstein
741ee7c867 Move stuff to Init 2016-12-12 15:55:32 -06:00
Alan Bernstein
bc4c32e027 Put tabs in usage string 2016-12-12 15:53:31 -06:00
Alan Bernstein
30436bd573 Update usage message 2016-12-08 19:18:34 -06:00
Alan Bernstein
4de72e27dd Add zipf benchmark with efficient random ID permutations 2016-12-08 19:14:24 -06:00
Alan Bernstein
a6cda39866 Add comments 2016-12-08 15:29:05 -06:00
Alan Bernstein
7dfb53d08e Use Duration's string representation in JSON output 2016-12-08 15:06:02 -06:00
jaffee
53171da640 thread context through Benchmarks
Also fix a few ignored errors, an issue with pilosactl usage string, a broken
test, and a mistaken import.
2016-12-08 11:02:14 -06:00
tgruben
6e71afc391 Merge pull request #191 from jaffee/166-single-slice
166 single slice
2016-12-08 10:31:22 -06:00
Alan Bernstein
060f06c1dc Merge stats updates 2016-12-07 17:30:08 -06:00
Alan Bernstein
75d282a478 Move sqrt out of Add() 2016-12-07 12:45:14 -06:00
Alan Bernstein
51faa59ada Add standard deviation to stats 2016-12-07 12:44:52 -06:00
jaffee
4c67cb30c0 make results nicer 2016-12-07 12:09:18 -06:00
jaffee
b4bb9cf70d add slice height benchmark 2016-12-06 17:20:38 -06:00
jaffee
af1d35affc import benchmark sort by profile id after bitmap id 2016-12-06 16:11:00 -06:00
jaffee
838941b26a wire up import benchmark, fix bug with ImportCommand.Host
also add some config metadata to results in order to distinguish them
2016-12-06 10:16:28 -06:00
jaffee
e5edf3e432 add import benchmark 2016-12-05 17:37:33 -06:00
jaffee
859a459ff6 copy errgroup code to bench and remove giant blob of dependencies 2016-12-01 16:58:56 -07:00
jaffee
fe28a095a5 add option to save all stats 2016-12-01 15:46:41 -07:00
jaffee
1d0a314287 rename imports umbel->pilosa 2016-11-28 15:45:18 -07:00
jaffee
a3b69f9066 rename bench commands and args
lower case and dash separated
2016-11-28 15:45:18 -07:00
jaffee
ea88941cff rename QueryMaker -> QueryGenerator 2016-11-28 15:45:18 -07:00
jaffee
406efcaca3 fixup imports in multidb 2016-11-28 15:45:18 -07:00
jaffee
6be7fb9bc2 use errgroup to simplify parallel benchmark Init 2016-11-28 15:45:18 -07:00
jaffee
aa2b4949d8 add random query benchmark 2016-11-28 15:45:18 -07:00
jaffee
861aee9a94 add some benchmark helpers for randomly generating queries 2016-11-28 15:45:18 -07:00
jaffee
9bf213af94 add fine grained stats within benchmarks 2016-11-28 15:45:17 -07:00
jaffee
65ac486b4c refactor Init and provide ClientType option
Benchmark implementations can now use 'HasClient' as a component which provides
an Init method which sets up the pilosa client in a configurable way.
Implementations must still expose the ClientType argument through ConsumeFlags
and describe it in Usage()
2016-11-28 15:45:17 -07:00
jaffee
2f4171bba8 add agentNum to benchmark Init as well as run 2016-11-28 15:45:17 -07:00
jaffee
0dc8548f45 split benchmarks into separate files 2016-11-28 15:45:17 -07:00
jaffee
256d516726 fix broken context stuff after rebase 2016-11-28 15:45:17 -07:00
jaffee
86d408310c add MultiDBSetBits benchmark 2016-11-28 15:45:17 -07:00
jaffee
553180870b don't repeat benchmark docs in 3 places
make Usage() the single place where arguments are described.
2016-11-28 15:45:17 -07:00
jaffee
cb85882289 add RandomSetBits benchmark demonstrating seeds 2016-11-28 15:45:17 -07:00
jaffee
c24c60fdba add timing info to benchmark results 2016-11-28 15:45:17 -07:00