1403de85a2
Include the cost of shifts in insert/remove benchmarks ( #268 )
...
Thanks!
2021-10-09 17:08:35 -07:00
c76756b473
Deprecate and hide ExtendFromSlice trait
...
This trait is only needed for internal benchmarking and should not have
been public. It will be removed in version 2.0.
2020-04-22 13:50:48 -07:00
aabaee0a82
bench and optimize from_slice
...
This adds "bench_from_slice(_vec)(_small)" benches, renaming the previous
versions to "bench_from_iter*" (because they were using `From<&[T]>`, which
calls `FromIterator` internally) and optimizes the `from_slice` constructor
considerably.
Before:
```
test bench_from_slice ... bench: 42 ns/iter (+/- 0)
test bench_from_slice_small ... bench: 12 ns/iter (+/- 0)
```
After:
```
test bench_from_slice ... bench: 27 ns/iter (+/- 0)
test bench_from_slice_small ... bench: 8 ns/iter (+/- 0)
```
2018-08-22 22:14:23 +02:00
759549e59f
Add return type to remove_noinline
2018-05-22 19:37:41 +02:00
5c3a66cb9e
Roll in improvements from #94
2018-05-22 13:17:43 +02:00
cd002392ed
Add benchmarks for smallvec when it's on the stack
2018-05-18 11:50:43 +02:00
65a42f3a1c
Add benchmarks for Vec to compare to SmallVec
2018-05-17 17:27:10 +02:00
ae82dc2ec2
Add smallvec! macro
2018-04-06 00:40:55 -04:00
82ee158e68
Add a benchmark for from_slice
...
Because it seems like good practice.
2017-03-14 12:10:18 -07:00
2a8859106c
Add bench insert_from_slice
2017-01-24 23:35:00 -08:00
24f52bd44c
Add benchmark for extend_from_slice
2017-01-24 23:34:59 -08:00
15c41c87a1
Implement insert_slice for SmallVec
...
Add benchmark for insert_many
2016-09-21 00:33:27 -07:00
59f5ea6057
Use a #noinline wrapper from the benchmarks to get consistent results
2016-09-12 00:48:00 -07:00
0e963993b7
Move benchmarks to benches module
...
This ensures that the benchmarks are in a separate crate and linked
against the smallvec dynamic library rather than being compiled
together.
2016-09-12 00:31:48 -07:00
Ben Kimock