As I stated in my “Decompression test” post, the new algorithm I use is faster in decompression than LZMA + usually it compress better, sometimes significantly. But there is a downside – it’s PAINFULLY slow at compression, 6-7 times slower than LZMA.
Here’s the sample of Black Ops 3 (hehe) data compressed:
Compressed 284,478 files, 83,612,223,393 => 32,521,343,528 bytes. Ratio 38.90%
Compression time: cpu 40860.64 sec/real 140086.91 sec = 29%. Speed 0.60 mB/s
Compression time: cpu 40860.64 sec/real 140086.91 sec = 29%. Speed 0.60 mB/s
As you can see, it took almost 39 hours to compress 77.9 GB. Yuck. But for end-user it’s much better due to above-mentioned pros, so now this algo (not public) is becoming my main, replacing the widespread LZMA. Faster installation + better compression = magic combo!
