4286ffd7a9
Documentation, copyright headers, etc., for open source releasing.
2019-02-25 15:28:29 -05:00
a557a28c5c
Create benchmarks for i32, i128, u32768.
2018-12-28 23:18:47 -05:00
fa353df41a
Factored benchmarks apart in a reasonable way.
2018-12-27 01:41:34 -05:00
227e2c548e
Write some docs.
2018-12-26 16:55:48 -05:00
c95c2769c4
Don't cfg(test) inside of cfg(test).
2018-12-16 15:18:09 -05:00
eecd629a7d
Bump up to Rust 2018.
2018-12-16 15:14:37 -05:00
99b96b390b
Add a README before forwarding to Google IARC.
2018-05-01 23:24:01 -04:00
8883b58cf3
Some benchmark outputs I was caching locally. Might be useful to reference in the future.
2018-04-29 12:31:34 -04:00
fe87772edc
Use Criterion for benchmarking.
2018-01-20 02:10:59 -05:00
d431f0bbfb
Run benchmarks over a wider variety of heap sizes.
2018-01-13 14:22:22 -05:00
bf8ac989c7
Get rid of unused unstable feature.
2018-01-13 14:18:56 -05:00
a66fe36d34
Use raw pointers to make fewer offset-based jumps.
2018-01-13 14:18:24 -05:00
2d1892fdb7
Recursive version of sift_down with some raw pointers to get maybe more faster.
2018-01-12 02:31:50 -05:00
2e43baa5fc
Messing around with trying to make things faster.
...
Unsafe array access is in, probably doesn't help.
Storing valence inline is in, probably doesn't help with speed. (Might help with
memory.)
Using bools instead of EdgeValence probably helps.
2018-01-12 00:42:15 -05:00
92b08ce3c0
Switch to Vec<bool> from BitVec to get speedup almost to parity with BinaryHeap.
2018-01-11 23:35:16 -05:00
6f82a8d099
Ditch insertion buffer to get # compares better than BinaryHeap.
2018-01-11 23:19:03 -05:00
d1d260b519
Unsafe version of sift_up with fewer swaps but same # of compares.
2018-01-11 23:10:44 -05:00
5e01209a4b
Slightly more efficient update of self.max_in_buffer in push.
2018-01-08 21:05:46 -05:00
cdcaaaa7b4
Add benchmarks for just insertion.
2018-01-08 21:05:33 -05:00
33b273e837
Improve timing vs. builtin so that we are only slower by a factor of 2.
...
Still doing more comparisons than builtin.
2018-01-07 22:51:30 -05:00
0c47621262
Move benchmarks to a standalone module that counts comparisons.
2018-01-07 15:45:51 -05:00
4fd58136ae
Fix bugs from checking wrong buffer length.
2018-01-04 11:34:14 -05:00
d57c951b86
Added a buffer for bulk inserts.
2017-12-29 22:12:44 -05:00
8815d9d996
Basic impl of a weak heap. About 2.5x slower than std::collections::BinaryHeap.
2017-12-29 19:09:48 -05:00
75759360f7
Hello, world.
2017-12-29 11:29:31 -05:00
