Kaby Lake x264 Benchmarks (Desktop)

[ezcapper]

Anandtech has benchmarked the Kaby Lake desktop processors x264 encoding speed so you can compare it to other processors. It's interesting that both the i3 and i5 underperform the i5-3470. That is probably a good sign for those anticipating Ryzen.

They use the hd x264 3.03 benchmark. I very much could be wrong, but I believe it is done by encoding a 720p "I am Legend" movie trailer with commands similar to those below:
--quiet --pass 2 --bitrate 3959 --stats "3.stats" --level 4.1 --keyint 24 --min-keyint 1 --ref 3 --mixed-refs --bframes 3 --weightb --direct auto --subme 6 --analyse p8x8,b8x8,i4x4,i8x8 --8x8dct --ipratio 1.1 --pbratio 1.1 --vbv-bufsize 30000 --vbv-maxrate 38000 --qcomp 0.5 --me umh --threads auto --thread-input --sar 1:1 --output "../run3-720p.mp4" "test-720p.avs" --mvrange 511 --aud 2>&1

benchmark results:

[sneaker]

It's interesting that both the i3 and i5 underperform AMD's $90 fx-6300

What do you mean by that? Not a single Kaby Lake was slower than the fx-6300 in the test.

Aside from that the hd x264 3.03 is from 2009. Probably lacking AVX, AVX2 and other optimizations.

[sophisticles]

I have seen reviews where the new I7 7700k is actually slower, clock for clock, than the i7 6700k; and I think this is done on purpose...

My theory: Intel expects AMD is hammer them with Ryzen from a price-performance standpoint, while I don't believe the initial claims of a sub $300 8C/16T Ryzen cpu, I do believe the claims of $250 6C/12T part. Intel has also expressed it's plans to bring their own hexacore main stream cpu's to market in 2018, I think Intel knows that Ryzen will be IPC competitive with Broadwell and to counter will release Coffee Lake in 2017 instead of 2018. If they do plan on doing this, Intel would want to show a significant performance gain between a hexacore Coffee Lake and it's "previous" gen quad core high end parts, namely Kaby Lake. The only way to do that is if Kaby Lake is no faster, and in some cases slower, than Skylake.

The reviews bear this out.

[ezcapper]

It's interesting that both the i3 and i5 underperform AMD's $90 fx-6300

What do you mean by that? Not a single Kaby Lake was slower than the fx-6300 in the test.

Aside from that the hd x264 3.03 is from 2009. Probably lacking AVX, AVX2 and other optimizations.

my mistake, I misread the fx-6300 and 6350 result. I will change the original post

[ezcapper]

I have seen reviews where the new I7 7700k is actually slower, clock for clock, than the i7 6700k; and I think this is done on purpose...

My theory: Intel expects AMD is hammer them with Ryzen from a price-performance standpoint, while I don't believe the initial claims of a sub $300 8C/16T Ryzen cpu, I do believe the claims of $250 6C/12T part. Intel has also expressed it's plans to bring their own hexacore main stream cpu's to market in 2018, I think Intel knows that Ryzen will be IPC competitive with Broadwell and to counter will release Coffee Lake in 2017 instead of 2018. If they do plan on doing this, Intel would want to show a significant performance gain between a hexacore Coffee Lake and it's "previous" gen quad core high end parts, namely Kaby Lake. The only way to do that is if Kaby Lake is no faster, and in some cases slower, than Skylake.

The reviews bear this out.

I agree with your theory. At the lower end, there are some great deals with kaby lake, like the sub $50 celeron g3930 and the hyperthreaded sub $90 pentium g4600, but it looks like they are setting things up for coffee lake to counter ryzen