[graph-tool-tickets] [graph-tool] Page: Performance changed

graph-tool webmaster at skewed.de
Tue Mar 5 22:46:36 UTC 2013


 * The user 'tiago peixoto' has changed the page: Performance.
 * Diff link: <URL:http://graph-tool.skewed.de/wiki/Performance?action=diff&version=5> 

Index: Performance
==============================================================================
--- Performance (version: 4)
+++ Performance (version: 5)
@@ -14,23 +14,32 @@
 The functions are called several times, and the average time take per
 function call is computed. The results can be seen below.
 
+{{{
+#!html
+<div style="align:center">
+}}}
 ||=Algorithm           =||= graph-tool (openmp) =||= graph-tool =||=   igraph   =||=  NetworkX  =||
 ||Shortest paths        ||        0.0107 s       ||    0.0098 s  ||    0.0099 s  ||              ||
 ||!PageRank             ||        0.201 s        ||    0.479  s  ||    0.832 s   ||   34.26 s    ||
 ||k-core                ||        0.0207 s       ||    0.0212 s  ||    0.0564 s  ||   0.9586 s   ||
 ||Minimum spanning tree ||        0.0305 s       ||    0.0304 s  ||    0.0413 s  ||   0.413  s   ||
 ||Betweenness           ||        ---            ||              ||              ||              ||
+{{{
+#!html
+</div>
+}}}
 
 In the second column are shown the results for graph-tool with openmp
 enabled using 4 threads on a [https://en.wikipedia.org/wiki/List_of_Intel_Core_i7_microprocessors#Sandy_Bridge_microarchitecture Intel Core i7 2620M] CPU with 4 cores, and
 the third column shows the results using only one core.
 
-The overall conclusion is that igraph delivers comparable performance to
+The overall conclusion is that igraph delivers performance comparable to
 graph-tool, being only marginally worse (~ 2x slower) on !PageRank and
 k-core. Of course, this difference increases when openmp is enabled by
 an additional factor of 2 on algorithms which run in parallel (!PageRank
 and Betweenness). On the other hand, NetworkX comes at a distant third
-with running times in the order of 10 to 170 times slower than graph-tool.
+with running times in the order of 10 to 170 times slower than
+graph-tool.
 
 All these libraries have their own advantages and
 disadvantages. Graph-tool's performance comes at the cost of large

--
Page URL: <URL:http://graph-tool.skewed.de/wiki/Performance>
graph-tool <URL:http://graph-tool.skewed.de/>
An efficient python module for graph analysis and manipulation.
-------------- next part --------------
 * The user 'tiago peixoto' has changed the page: Performance.
 * Diff link: <URL:http://graph-tool.skewed.de/wiki/Performance?action=diff&version=5> 

Index: Performance
==============================================================================
--- Performance (version: 4)
+++ Performance (version: 5)
@@ -14,23 +14,32 @@
 The functions are called several times, and the average time take per
 function call is computed. The results can be seen below.
 
+{{{
+#!html
+<div style="align:center">
+}}}
 ||=Algorithm           =||= graph-tool (openmp) =||= graph-tool =||=   igraph   =||=  NetworkX  =||
 ||Shortest paths        ||        0.0107 s       ||    0.0098 s  ||    0.0099 s  ||              ||
 ||!PageRank             ||        0.201 s        ||    0.479  s  ||    0.832 s   ||   34.26 s    ||
 ||k-core                ||        0.0207 s       ||    0.0212 s  ||    0.0564 s  ||   0.9586 s   ||
 ||Minimum spanning tree ||        0.0305 s       ||    0.0304 s  ||    0.0413 s  ||   0.413  s   ||
 ||Betweenness           ||        ---            ||              ||              ||              ||
+{{{
+#!html
+</div>
+}}}
 
 In the second column are shown the results for graph-tool with openmp
 enabled using 4 threads on a [https://en.wikipedia.org/wiki/List_of_Intel_Core_i7_microprocessors#Sandy_Bridge_microarchitecture Intel Core i7 2620M] CPU with 4 cores, and
 the third column shows the results using only one core.
 
-The overall conclusion is that igraph delivers comparable performance to
+The overall conclusion is that igraph delivers performance comparable to
 graph-tool, being only marginally worse (~ 2x slower) on !PageRank and
 k-core. Of course, this difference increases when openmp is enabled by
 an additional factor of 2 on algorithms which run in parallel (!PageRank
 and Betweenness). On the other hand, NetworkX comes at a distant third
-with running times in the order of 10 to 170 times slower than graph-tool.
+with running times in the order of 10 to 170 times slower than
+graph-tool.
 
 All these libraries have their own advantages and
 disadvantages. Graph-tool's performance comes at the cost of large

--
Page URL: <URL:http://graph-tool.skewed.de/wiki/Performance>
graph-tool <URL:http://graph-tool.skewed.de/>
An efficient python module for graph analysis and manipulation.


More information about the graph-tool-tickets mailing list