jyapayne/Nim
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

faster CIs (#13803)

Browse source 
* ttables: smaller table, 5x speedup

* thavlak: less iterations, less loops; 30% speedup

* tasyncclosestall: shorter timeout; 35% speedup

* gcleak4: less iterations, 2x speedup

* ttimes: remove deprecated stuff

* tdangerisrelease: remove cpp backend, 3x speedup

* tfrexp1: smaller range, 2x speedup

* trtree: fix warnings, less iterations, 6x speedup

* tasyncawait_cyclebreaker: smaller swarm size; 2x speedup

* trealloc: smaller number of iterations; 10x speedup

* towned_binary_tree: less iterations, 4x speedup

* tclosure: remove unused code, less iterations; 2x speedup

* twaitany: less durations; 1.4x speedup

* tasync_misc: less iterations, 2x speedup

* t8535: smaller sleep, 1.5x speedup

* tmanyjoin: smaller sleep, 2x speedup

* t12221: shorter sleeps, removed two slower tests; 1.6x speedup

* tfuturestream: smaller sleep; 1.5x speedup

* growobjcrash: less iterations; 2x speedup

* ttryrecv: smaller sleep; 1.5x speedup

* treusetvar: less threads; 2x speedup

* delete tthreadanalysis2, basically a duplicate of tthreadanalysis

* t7758: less iterations, 1.5x speedup

* tasyncawait: smaller swarm, less messages; 1.5x speedup

* tjsandnativeasync: smaller sleep, 1.5x speedup

* tpendingcheck: smaller sleep, 1.5x speedup

* remove rodfiles test category

* move tseq from its own category to 'collections' category

* remove unneeded tests and helpers from 'assert' category

* stdlib: merge tbitops2 into tbitops

* remove 'trepr2' from 'stdlib' cat

* merge 'tstreams' into one file

* remove 'tinefficient_const_table' from 'ccbugs' cat

* merge 'tcollections_to_string' into 'tcollections'

* tblocking_channel: smaller sleep, small speedup

* tconvexhull: less iterartions; 1.2x speedup

* merge 'tdeepcopy2' into 'tdeepcopy'

* merge 'tdisjoint_slice2' into 'tdisjoint_slice1'

* tmissing_deepcopy: smaller sequence

* tsendtwice: smaller arrays; 5x speedup

* remove 'tindexerrorformatbounds'

* disable multimethod tests

* remove 'gc:none' and 'refc' without 'd:useRealtimeGC' from gc tests

* koch.nim: bootstrap just with '-d:release', no need for 'csource'

* add github workflow for documentation

* testament: no need for 8 sub-second decimals
This commit is contained in:
Miran 2020-03-30 13:18:12 +02:00 • committed by GitHub
commit 8088633250
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
72 changed files with 532 additions and 1082 deletions
Download patch file Download diff file Expand all files Collapse all files

7
tests/parallel/tconvexhull.nim
View file

@ -51,12 +51,11 @@ proc convex_hull[T](points: var seq[T], cmp: proc(x, y: T): int {.closure.}) : s
ul[k] = spawn half[T](points, k == 0)
result = concat(^ul[0], ^ul[1])
var s = map(toSeq(0..99999), proc(x: int): Point = (float(x div 1000), float(x mod 1000)))
var s = map(toSeq(0..9999), proc(x: int): Point = (float(x div 100), float(x mod 100)))
# On some runs, this pool size reduction will set the "shutdown" attribute on the
# worker thread that executes our spawned task, before we can read the flowvars.
setMaxPoolSize 2
#echo convex_hull[Point](s, cmpPoint)
for i in 0..5:
for i in 0..2:
doAssert convex_hull[Point](s, cmpPoint) ==
@[(0.0, 0.0), (99.0, 0.0), (99.0, 999.0), (0.0, 999.0)]
@[(0.0, 0.0), (99.0, 0.0), (99.0, 99.0), (0.0, 99.0)]