[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-09-12 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 5: Verified+1


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 5
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 
Gerrit-Comment-Date: Wed, 12 Sep 2018 21:10:53 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-09-12 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has submitted this change and it was merged. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..

IMPALA-7426: Use Mann-Whitney U to compare benchmarks

The Mann-Whitney test can be used to compare samples taken from
non-normal distributions, and so can more accurately reflect
performance changes than a T-test. This patch does not remove t-tests
from the benchmark reporting, it just supplements them by including
the Mann-Whitney test result as well.

Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Reviewed-on: http://gerrit.cloudera.org:8080/11194
Reviewed-by: Jim Apple 
Tested-by: Impala Public Jenkins 
---
M tests/benchmark/report_benchmark_results.py
M tests/util/calculation_util.py
2 files changed, 85 insertions(+), 37 deletions(-)

Approvals:
  Jim Apple: Looks good to me, approved
  Impala Public Jenkins: Verified

--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: merged
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 6
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-09-12 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 5:

Build started: https://jenkins.impala.io/job/gerrit-verify-dryrun/3149/ 
DRY_RUN=false


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 5
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 
Gerrit-Comment-Date: Wed, 12 Sep 2018 17:47:22 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-09-12 Thread Jim Apple (Code Review)
Jim Apple has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 5: Code-Review+2

Carry Philip's +2 on rebase


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 5
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 
Gerrit-Comment-Date: Wed, 12 Sep 2018 17:46:58 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-09-10 Thread Philip Zeyliger (Code Review)
Philip Zeyliger has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 4: Code-Review+2

Thanks!


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 4
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 
Gerrit-Comment-Date: Mon, 10 Sep 2018 21:57:18 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-31 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 3:

Build Failed

https://jenkins.impala.io/job/gerrit-code-review-checks/562/ : Initial code 
review checks failed. See linked job for details on the failure.


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 3
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 
Gerrit-Comment-Date: Sat, 01 Sep 2018 02:51:36 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-31 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 3:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/11194/3/tests/benchmark/report_benchmark_results.py
File tests/benchmark/report_benchmark_results.py:

http://gerrit.cloudera.org:8080/#/c/11194/3/tests/benchmark/report_benchmark_results.py@573
PS3, Line 573: a
flake8: W503 line break before binary operator



--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 3
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 
Gerrit-Comment-Date: Sat, 01 Sep 2018 02:33:09 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-31 Thread Jim Apple (Code Review)
Hello Philip Zeyliger, Impala Public Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/11194

to look at the new patch set (#3).

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..

IMPALA-7426: Use Mann-Whitney U to compare benchmarks

The Mann-Whitney test can be used to compare samples taken from
non-normal distributions, and so can more accurately reflect
performance changes than a T-test. This patch does not remove t-tests
from the benchmark reporting, it just supplements them by including
the Mann-Whitney test result as well.

Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
---
M tests/benchmark/report_benchmark_results.py
M tests/util/calculation_util.py
2 files changed, 85 insertions(+), 37 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/94/11194/3
--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 3
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Reviewer: Philip Zeyliger 


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-30 Thread Philip Zeyliger (Code Review)
Philip Zeyliger has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 2:

(7 comments)

http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py
File tests/benchmark/report_benchmark_results.py:

http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@a568
PS2, Line 568:
is removing num_clients intentional?


http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@129
PS2, Line 129:  dest="max_percent_change_threshold", 
default=float("inf"),
Is the default change intentional?


http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@394
PS2, Line 394:   if not options.hive_results:
 : try:
 :   save_runtime_diffs(results, ref_results, 
self.perf_change, self.zval, self.tval)
 : except Exception as e:
 :   LOG.error('Could not generate an html diff: 
{0}'.format(e))
This code used to never run because it was indented incorrectly? It looks ok to 
me, but I'm not clear how it ever ran before?


http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@563
PS2, Line 563:  'Base StdDev(%)', 'Iters', 
'Median Diff(%)', 'Zval',
Should this be "MW Z-Val" or something to indicate what kind of zvalue this is?


http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@1037
PS2, Line 1037: prefix = ('reg' if zval >= 0 and tval >=
This is different than __check_perf_change_significance's implementation. Does 
that make sense?


http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@1037
PS2, Line 1037: prefix = ('reg' if zval >= 0 and tval >=
  :   0 else 'imp' if zval <= 0 and tval <= 0 else 
'???')
I think this would be more legible as multiple lines. (I have a deep hatred of 
python's non-ternary, but when it's nested, I can't make heads or tails of it.)


http://gerrit.cloudera.org:8080/#/c/11194/2/tests/util/calculation_util.py
File tests/util/calculation_util.py:

http://gerrit.cloudera.org:8080/#/c/11194/2/tests/util/calculation_util.py@74
PS2, Line 74:   Calculates the Mann-Whitney U Test Z value for the given 
samples and reference.
Have you compared this to 
https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.mannwhitneyu.html?
 Do we have a significant number of sample (at least 20 per that  page) to use 
this?

After installing python-scipy, I ran the following. Of note, the U calculation 
agrees (good). The scipy thing gives a p-value and we are returning a z-value 
and then taking its absolute value and looking at a threshold. Are we doing the 
right thing?

You might want to update the pydoc here to explain what to do with one of these 
values.



$cat test.py
import math

def calculate_mwu(samples, ref_samples):
  """
  Calculates the Mann-Whitney U Test Z value for the given samples and 
reference.
  """
  tag_a = [(s, 'A') for s in samples]
  tab_b = [(s, 'B') for s in ref_samples]
  ab = tag_a + tab_b
  ab.sort()
  # Assume no ties
  u = 0
  count_b = 0
  for v in ab:
if v[1] == 'A':
  u += count_b
else:
  count_b += 1
  # u is normally distributed with the following mean and standard deviation:
  mean = len(samples) * len(ref_samples) / 2.0
  stddev = math.sqrt(len(samples) * len(ref_samples) * (1 + len(ab)) / 12.0)
  print u, mean, stddev
  return (u - mean) / stddev

from scipy.stats import mannwhitneyu

a = range(30)
b = [ z + 0.5 for z in a ]

print mannwhitneyu(a, b)
print mannwhitneyu(a, b, use_continuity=False)
print mannwhitneyu(a, b, alternative="two-sided")
print mannwhitneyu(a, b, alternative="less")
print mannwhitneyu(a, b, alternative="less",use_continuity=False)
print mannwhitneyu(a, b, alternative="two-sided",use_continuity=False)

print calculate_mwu(a, b)
print 1 - abs(calculate_mwu(a, b))
09:28:24 pannier cmf [master !x?*] ~/src/cmf
$python test.py
MannwhitneyuResult(statistic=435.0, pvalue=0.8302552839111963)
MannwhitneyuResult(statistic=435.0, pvalue=0.82449575165477107)
MannwhitneyuResult(statistic=435.0, pvalue=0.8302552839111963)
MannwhitneyuResult(statistic=435.0, pvalue=0.41512764195559815)
MannwhitneyuResult(statistic=435.0, pvalue=0.41224787582738553)
MannwhitneyuResult(statistic=435.0, pvalue=0.82449575165477107)
435 450.0 67.6387462923
-0.221766381286
435 450.0 67.6387462923
0.778233618714



--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194

[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-12 Thread Jim Apple (Code Review)
Jim Apple has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 2:

(9 comments)

http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py
File tests/benchmark/report_benchmark_results.py:

http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@121
PS1, Line 121:
> flake8: E502 the backslash is redundant between brackets
Done


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@380
PS1, Line 380:
> flake8: E703 statement ends with a semicolon
Done


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: ,
> flake8: F821 undefined name 'ref_results'
This whole section was, even in HEAD, indented 4 spaces, which puts it at the 
class level, not the method level. I indented in 6 spaces now, but something 
was already fishy here.


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@573
PS1, Line 573: a
> flake8: W503 line break before binary operator
This is now flipped?

https://github.com/python/peps/commit/c59c4376ad233a62ca4b3a6060c81368bd21e85b

I'll send another patch.


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@1016
PS1, Line 1016:
> flake8: E302 expected 2 blank lines, found 1
Done


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@1036
PS1, Line 1036:
> flake8: E501 line too long (94 > 90 characters)
Done


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py
File tests/util/calculation_util.py:

http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py@71
PS1, Line 71:
> flake8: E302 expected 2 blank lines, found 1
Done


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py@75
PS1, Line 75:
> flake8: E231 missing whitespace after ','
Done


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py@76
PS1, Line 76: ,
> flake8: E231 missing whitespace after ','
Done



--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 2
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Reviewer: Jim Apple 
Gerrit-Comment-Date: Sun, 12 Aug 2018 22:08:07 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-12 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 2:

Build Successful

https://jenkins.impala.io/job/gerrit-code-review-checks/300/ : Initial code 
review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun 
to run full precommit tests.


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 2
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Comment-Date: Sun, 12 Aug 2018 18:11:32 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-12 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 2:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py
File tests/benchmark/report_benchmark_results.py:

http://gerrit.cloudera.org:8080/#/c/11194/2/tests/benchmark/report_benchmark_results.py@573
PS2, Line 573: a
flake8: W503 line break before binary operator



--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 2
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Comment-Date: Sun, 12 Aug 2018 17:39:53 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-12 Thread Jim Apple (Code Review)
Hello Impala Public Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/11194

to look at the new patch set (#2).

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..

IMPALA-7426: Use Mann-Whitney U to compare benchmarks

The Mann-Whitney test can be used to compare samples taken from
non-normal distributions, and so can more accurately reflect
performance changes than a T-test. This patch does not remove t-tests
from the benchmark reporting, it just supplements them by including
the Mann-Whitney test result as well.

Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
---
M tests/benchmark/report_benchmark_results.py
M tests/util/calculation_util.py
2 files changed, 82 insertions(+), 37 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/94/11194/2
--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 2
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-11 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 1:

Build Successful

https://jenkins.impala.io/job/gerrit-code-review-checks/298/ : Initial code 
review checks passed. Use gerrit-verify-dryrun-external or gerrit-verify-dryrun 
to run full precommit tests.


--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 1
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Comment-Date: Sat, 11 Aug 2018 18:34:17 +
Gerrit-HasComments: No


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-11 Thread Impala Public Jenkins (Code Review)
Impala Public Jenkins has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11194 )

Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..


Patch Set 1:

(14 comments)

http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py
File tests/benchmark/report_benchmark_results.py:

http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@121
PS1, Line 121: \
flake8: E502 the backslash is redundant between brackets


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@380
PS1, Line 380: ;
flake8: E703 statement ends with a semicolon


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: s
flake8: F821 undefined name 'save_runtime_diffs'


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: r
flake8: F821 undefined name 'results'


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: r
flake8: F821 undefined name 'ref_results'


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: s
flake8: F821 undefined name 'self'


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: s
flake8: F821 undefined name 'self'


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@396
PS1, Line 396: s
flake8: F821 undefined name 'self'


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@573
PS1, Line 573: a
flake8: W503 line break before binary operator


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@1016
PS1, Line 1016: def save_runtime_diffs(results, ref_results, 
change_significant, zval, tval):
flake8: E302 expected 2 blank lines, found 1


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/benchmark/report_benchmark_results.py@1036
PS1, Line 1036: ?
flake8: E501 line too long (94 > 90 characters)


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py
File tests/util/calculation_util.py:

http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py@71
PS1, Line 71: def calculate_mwu(samples, ref_samples):
flake8: E302 expected 2 blank lines, found 1


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py@75
PS1, Line 75: ,
flake8: E231 missing whitespace after ','


http://gerrit.cloudera.org:8080/#/c/11194/1/tests/util/calculation_util.py@76
PS1, Line 76: ,
flake8: E231 missing whitespace after ','



--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 1
Gerrit-Owner: Jim Apple 
Gerrit-Reviewer: Impala Public Jenkins 
Gerrit-Comment-Date: Sat, 11 Aug 2018 18:01:08 +
Gerrit-HasComments: Yes


[Impala-ASF-CR] IMPALA-7426: Use Mann-Whitney U to compare benchmarks

2018-08-11 Thread Jim Apple (Code Review)
Jim Apple has uploaded this change for review. ( 
http://gerrit.cloudera.org:8080/11194


Change subject: IMPALA-7426: Use Mann-Whitney U to compare benchmarks
..

IMPALA-7426: Use Mann-Whitney U to compare benchmarks

The Mann-Whitney test can be used to compare samples taken from
non-normal distributions, and so can more accurately reflect
performance changes than a T-test. This patch does not remove t-tests
from the benchmark reporting, it just supplements them by including
the Mann-Whitney test result as well.

Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
---
M tests/benchmark/report_benchmark_results.py
M tests/util/calculation_util.py
2 files changed, 75 insertions(+), 33 deletions(-)



  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/94/11194/1
--
To view, visit http://gerrit.cloudera.org:8080/11194
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newchange
Gerrit-Change-Id: I8d6631ebeba1422b832def5cd68537624f672fa0
Gerrit-Change-Number: 11194
Gerrit-PatchSet: 1
Gerrit-Owner: Jim Apple