Ryzen 3000 AES

L

Link

Gawd
Joined
Jul 28, 2002
Messages
699
While browsing the Geekbench, I've noticed much lower Multicore AES score from Ryzen 3700x (@4.30mhz) than from 2700x (@4.33mhz) https://browser.geekbench.com/v4/cpu/compare/13730078?baseline=13741736

Anyone can shed some light on this?
Annotation 2019-07-02 135215.jpg
 
Last edited:
There are several places where the performance decreases in which I think we will have to wait for official reviews.
 
The memory latency makes me think a subtiming is loose on the 3700x particularly command rate 2T would drop performance like that.
 
PliotronX said:
The memory latency makes me think a subtiming is loose on the 3700x particularly command rate 2T would drop performance like that.
Click to expand...
The memory latency alone cannot be the cause of the big difference. The single core AES scores are almost identical.
Maybe the benchmark is not using all the cores?
 
drescherjm said:
There are several places where the performance decreases in which I think we will have to wait for official reviews.
Click to expand...
All of those are very subtle and memory latency would explain them, even AES single-core. Except the AES multicore, but does this multicore AES performance translate to gaming?

Looks odd, as if the program can't use all of the 3700X cores for this AES test.
 
Link said:
The memory latency alone cannot be the cause of the big difference. The single core AES scores are almost identical.
Maybe the benchmark is not using all the cores?
Click to expand...
Good point, in that case I'll just scrutinize that bench this weekend for curiosity :)
 
memory latency would explain them
Click to expand...

I was thinking of that before I posted my first comment however I am not 100% sure I believe the result. It could be true because of the way the IO die works but not sure.
 
You must log in or register to reply here.
Back
Top