[SOLVED] 4x8 vs 2x16 and about latency (CL)

Yakifuji

Commendable
May 3, 2020
10
0
1,510
I'm looking at some parts to build a computer and I'm confused about some things.

A 3000mhz (16-18-18-36) 4x8 is better than 3000mhz (16-18-18-36) 2x16 ??? It's for a B450M Aorus Elite and R5 3600 build.

The way people use to calculate absolute latency (I think that's the name) is fair? I used this website here, the first calculator you find on Google.
I found two Kingston DDR4, one is 3733MHz CL19 and other is 3000MHz CL15, the calculator shows that the first one latency is 10.17 and the last one is 10. Why would they do this if the absolute latency is almost the same? Am I missing something?
 
Solution
You should also take into consideration the memory clock speed. If your not familiar with it look into Ryzen's Infinity Fabric. It's the interlink used to communicate between cores of the Ryzen CPU and is clocked at the memory base frequency. So 1500mhz vs 1866mhz between your to selections (DDR memory runs at double the clock frequency). If all else is equal I'd go for the higher memory clock speed. The only considerations is 1866 is pushing the max speed you can run the infinity fabric at. A weaker Ryzen CPU might not handle that speed.
You should also take into consideration the memory clock speed. If your not familiar with it look into Ryzen's Infinity Fabric. It's the interlink used to communicate between cores of the Ryzen CPU and is clocked at the memory base frequency. So 1500mhz vs 1866mhz between your to selections (DDR memory runs at double the clock frequency). If all else is equal I'd go for the higher memory clock speed. The only considerations is 1866 is pushing the max speed you can run the infinity fabric at. A weaker Ryzen CPU might not handle that speed.
 
  • Like
Reactions: white.a.drew
Solution
I'm looking at some parts to build a computer and I'm confused about some things.

A 3000mhz (16-18-18-36) 4x8 is better than 3000mhz (16-18-18-36) 2x16 ??? It's for a B450M Aorus Elite and R5 3600 build.

The way people use to calculate absolute latency (I think that's the name) is fair? I used this website here, the first calculator you find on Google.
I found two Kingston DDR4, one is 3733MHz CL19 and other is 3000MHz CL15, the calculator shows that the first one latency is 10.17 and the last one is 10. Why would they do this if the absolute latency is almost the same? Am I missing something?
It depends on the use of the computer it generally doesn't matter some people don't care at all.... others if it's specifically designed for specific things will use one or the other. The more sticks there are the longer you need to wait for the memory response.