Xeon E5 V1 to E5 V2 performance increase with x264

Stereodude

Active Member
Feb 21, 2016
412
65
28
USA
I thought this information might help someone. I've been running a x264 2nd pass encode with x264 build 2525 (x64) re-encoding a 1920x1080 Blu-ray with the following command line switches:
Code:
--bitrate 18438 --preset veryslow --tune film --bluray-compat --vbv-maxrate 40000 --vbv-bufsize 30000 --level 4.1 --keyint 24 --open-gop --slices 4 --colorprim "bt709" --transfer "bt709" --colormatrix "bt709" --sar 1:1
E5-2687W v2 is 25.68% faster than a E5-2670 v1
E5-2689 v1 is 1.21% faster than a E5-2670 v1
E5-2690 v2 is 38.62% faster than a E5-2670 v1
E5-2690 v2 is 87.58% faster than a i7-4770k @4.2gHz
E5-2687W v2 is 70.07% faster than a i7-4770k @4.2gHz
E5-2690 v2 is 10.30% faster than a E5-2687W v2
 
  • Like
Reactions: Jeggs101

Jeggs101

Well-Known Member
Dec 29, 2010
1,492
229
63
If there was some stock footage, it'd be cool to come up with a benchmark for this. All of us here have lots of different kinds of CPUs.
 

Stereodude

Active Member
Feb 21, 2016
412
65
28
USA
If there was some stock footage, it'd be cool to come up with a benchmark for this. All of us here have lots of different kinds of CPUs.
There are some "stock" video clips we could use (see here), but they tend to be rather short which makes them hard to get an accurate measure of performance. Unless people want to download a near 50gB file to use as a test clip, like Big Buck Bunny, which is a longer & more suitable length for using as a benchmark.

There are also considerations of threading restrictions. To maintain the best quality for a 1080p video x264 should not be allow to spawn more than 27 threads. Of course there are command line switches for x264 to control the threading, but it starts to get complicated because some people will likely argue that encodes using different thread counts shouldn't be compared because they may not have exactly the same encode quality.