Initial data on AMD Zen5 x265 performance

[benwaggoner]

There has been much speculation about how the new AMD desktop chips on the Zen5 architecture would perform with AVX512, as they have new full-width AVX512 support.

Ars technical has an initial look: https://arstechnica.com/gadgets/2024...-speed-boosts/

The higher-end 13th and 14th-generation Core chips from Intel manage to keep pace with the 9950X's multi-core performance, mostly thanks to the sheer number of cores, an advantage of Intel's heterogenous approach to its desktop processors. But there are outliers in either direction; 3DMark Time Spy and Shadow of the Tomb Raider both run quite a bit faster on Intel's cores, but our H.265 video encoding benchmark is a lot faster on the Ryzen 9000 chips thanks to their beefed-up AVX512 support (the 9000-series CPUs are disproportionately faster than the 7000-series CPUs here, too).

AMD has the best fps, at about half again better fps/watt.

[Selur]

"Initial data on AMD Zen3 x265 performance" -> "Initial data on AMD Zen5 x265 performance"

[rwill]

Heh..

And several Review Sites and YouTube Channels are bashing AMD because "ZeN5 iS nOt FaStEr ThAn ZeN4 iN GAMES". Well, looks like playtime is over.

[excellentswordfight]

Techpowerup does x265 with preset slow at crf20:



Do note that to non of these tests are with avx512-activated, and that most review-sites does tests at medium or faster.

But both anandtech and phoronix did test avx512 performance in other software, and there is a clear uplift from zen4, and the avx512 perforamnce in general is good (do note that there is huge amount of your mileage might vary when it comes to the performance gain of using avx512 depending on the load, with the right load, like different types of simulation you can get like 10x the performance), My guestimate is that activating avx512 will give you another 10% performance for x265 with zen5.

[Z2697]

Originally Posted by excellentswordfight

Techpowerup does x265 with preset slow at crf20:



Do note that to non of these tests are with avx512-activated, and that most review-sites does tests at medium or faster.

But both anandtech and phoronix did test avx512 performance in other software, and there is a clear uplift from zen4, and the avx512 perforamnce in general is good (do note that there is huge amount of your mileage might vary when it comes to the performance gain of using avx512 depending on the load, with the right load, like different types of simulation you can get like 10x the performance), My guestimate is that activating avx512 will give you another 10% performance for x265 with zen5.

I'm not suggesting the data is wrong, but how could 14900K that much faster than 13900K? They are basically same chips aren't they?
And how could 14700K not much faster than 13700K (I mean, 2 seconds faster is significant but shouldn't it be better than this?) considering it's the only chip that has a real advantage over 13th gen (4 more E-cores) in the desktop 14th gen lineup?