These are chat archives for boostorg/hana

21st
Jan 2016
Jason Rice
@ricejasonf
Jan 21 2016 00:15
@ldionne I made a new PR to the develop branch. Github seems to have tacked on all of those missing comments from the other branch. :P
Perhaps I could snatch that improved benchmark and put it in there.
Jason Rice
@ricejasonf
Jan 21 2016 00:53
git pull --rebase :heart_eyes_cat:
Jason Rice
@ricejasonf
Jan 21 2016 03:14
I haven't run your improved at_key.lookup benchmark against the canonical map, but the hash version appears to run a little faster than the version of map in benchmarks/map
      "name": "hana::map",
      "data": [[1, 0.44933838400174864], [6, 0.9713527849962702], [11, 1.447662008999032], [16, 1.9539198130078148], [21, 2.5651180559943896], [26, 2.9488706910051405], [31, 3.4967271969944704], [36, 4.089561719010817], [41, 4.627005657006521], [46, 5.077234721989953], [50, 5.543645400000969], [75, 8.234895086003235], [100, 10.817610883997986], [125, 13.418156313011423], [150, 15.854205777999596], [175, 18.82086028800404], [200, 21.132252871000674]]

      "name": "hana::map",
      "data": [[1, 0.3815776410046965], [6, 0.3848808329930762], [11, 0.41526929300744087], [16, 0.433267189990147], [21, 0.47254635399440303], [26, 0.49840208899695426], [31, 0.5164234940020833], [36, 0.55401539299055], [41, 0.5615368240105454], [46, 0.6031375279999338], [50, 0.6096746969997184], [75, 0.7392608359950827], [100, 0.8715285969956312], [125, 0.9752421349985525], [150, 1.1177485280059045], [175, 1.3036450960062211], [200, 1.348750895995181]]
Jason Rice
@ricejasonf
Jan 21 2016 03:26
:D
Louis Dionne
@ldionne
Jan 21 2016 05:01
Well that’s pretty damn good!
Louis Dionne
@ldionne
Jan 21 2016 23:55
@ricejasonf Did you mean to put the two comments you made on #118 at #242 instead?
Jason Rice
@ricejasonf
Jan 21 2016 23:56
perhaps I should have, but they are both related
Louis Dionne
@ldionne
Jan 21 2016 23:57
Yes, they are. I’d rather we keep the comments on the PR, since that’s where my comment was.
The comment about hash collisions, I mean.
Jason Rice
@ricejasonf
Jan 21 2016 23:57
sorry about that
Louis Dionne
@ldionne
Jan 21 2016 23:58
No problem, but can you please re-comment on #242 and I’ll answer there?
Jason Rice
@ricejasonf
Jan 21 2016 23:58
ok
Louis Dionne
@ldionne
Jan 21 2016 23:58
That looks picky, but it’ll be a great help if someone ever looks at these PRs again. Otherwise it’s difficult to follow what happened and why.
Jason Rice
@ricejasonf
Jan 21 2016 23:59
right