Please paste more here.įor each one we've noted uname -a, gcc -version and vips -version. When we say CPUs - these days that really means "cores". (real-all-cpu-time / real-1-cpus-time) in other words speedup using all cores Time is lowest real time (wall clock time) in seconds, Speedup is Results summary ProcessorĪRM A7 quad core Raspberry Pi 2 B (32 bit)
Feel free to email them to for inclusion here.
On intel CPUs -O2 is slightly faster than -O3.Īnd see results for your system. Most CPUs now clock faster in this case, which skews the results. The speed-up in the table is the speed-up compared to running one thread (on one core). Also that Hyperthreading is not very useful for float-intensive tasks such as this. Interesting things to learn from these include the fact that a slightly slower CPU with more cores is better (also less complex to keep cool!) - for example the i7 6700K is slower than the 6 core 3930K. Without increasing its workload and scaling. In 2020 the time taken has reduced to less than 0.5s so it may become harder to measure from now You could make im_benchmark() a lot faster very easily if that was your aim. No attempt was made to make it quick (there was no point) This was originally processing images from a remote server over a 100 You can see theĮxact sequence of operations the benchmark performs in the source Processed, resized, cropped and sharpened.
In this benchmark images from a 10k by 10k studio digital camera are colour Some detail on application and performance, in the PARSEC Tecnical There's a description of the benchmark, including To generate images for The National Gallery's Print on DemandĪs one of their tests. VIPS (from version 7.11.12) includes a benchmark adapted from the system used We have a separate set of benchmarksĬomparing VIPS to other image processing systems on the Speed-and-memory-use page. This benchmark is useful for testing the VIPS multi-threading system, for comparing generations of processorsĪnd for testing for performance regressions or