Update doc

This commit is contained in:
Patrick Pelissier committed 2023-04-08 11:32:15 +02:00
1 parent bafd307247
commit e74f80cc59
1 file changed
+8 -10
+8 -10
View File
@@ -457,33 +457,31 @@ Performance
-----------
M\*LIB performance is compared to the one of GNU C++ STL (v10.2) in the following graphs.
Each graph is presented in non-logarithmic and logarithmic scale to better realize.
Each graph is presented first in linear scale and then in logarithmic scale to better realize the differences.
M\*LIB is on par with the STL or even faster.
All used bench codes are available in the bench directory.
The results for several different libraries are also available [in a separate page](https://github.com/P-p-H-d/mlib/wiki/performance).
#### Singly List
![Singly List performance](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-list.png)
![Singly List performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-list.png)
![Singly List performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-list-log.png)
#### Array
![Array performance](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-array.png)
![Array performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-array.png)
![Array performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-array-log.png)
#### Unordered Map
![Unordered Map performance](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-umap.png)
![Unordered Map performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-umap.png)
![Unordered Map performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-umap-log.png)
#### Ordered Set
![Ordered Set performance](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-oset.png)
![Ordered Set performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-oset.png)
#### Bench source
All used bench codes are available in the bench directory.
The results for several libraries are available [in a separate page](https://github.com/P-p-H-d/mlib/wiki/performance).
![Ordered Set performance - Log Scale](https://raw.githubusercontent.com/P-p-H-d/mlib/master/doc/bench-oset-log.png)