Benchamarks

Jan 9, 2011 at 5:20 PM

Your benchmarks are completely wrong and misleading, sorry :)

For example, it does not take "Original CLZF C#" 8 seconds to compress 150 KB. That would give a speed of 10 KB/sec. But in reality it compresses at something like 30 MB/s.

That's a factor 3000 wrong.

Also, "QuickLZ C#" is alot faster at compression than "Original CLZF C#" but 2-3 times slower in your benchmark.

Finally, SharpZipLib does not decompress at just 10 KB/s, but at more than 100 MB/s. A factor 10000 wrong.

I think your timer resolution might be too coarse. Try executing each compression/decompression in a loop multiple times or use high res timer.