Discussion around the Framework mission of building products that last longer by making them upgradeable, customizable, and repairable. Consumer electronics can be better for you and for the environment.
CL40 is better than CL46. CL stands for CAS Latency, or Column Access Strobe Latency: lower is better.
That being said, I have yet to see any head-to-head comparison between 5600 CL46 and 5600 CL40 kits. My assumption is that it would be minor in most cases. The primary performance benefit to higher-frequency (and therefore higher-bandwidth) RAM is iGPU performance. I wouldn't expect that to be terribly latency-sensitive.
Out of interest do we know what CAS the Framework branded RAM is?
CL46 I believe.
Cl 46 is worse than cl 40 you what the LOWEST latency possible. In terms of speed not much of a difference.
The CL primary timing has very little direct perf changes. The other primary timings have some slight perf changes but not by a whole lot.
The main perf differences will come from secondary/tertiary timings which will be auto trained by the motherboard (or set manually of the BIOS supports manual tuning). CL has some influence on certain timings but the differences are still minor for what will be auto trained. The timings set by the auto training will be dependent on the memory IC type, binning of the IC, PCB quality of the DIMMs, voltage being supplied to the DIMMs, DRAM frequency, CPU (IMC [integrated memory controller], some other misc stuff), and slightly by the primary timings.
The XMP values tend to allow one to infer what IC is being used and the bin of it (if said IC is even binned), which is what usually tends to make the biggest differences in what the timings will be. But the advertised primary timings themselves tend not to be the reason for perf differences. Note that better XMP values won't always mean a better DIMM. Possible to have high quality ICs in a low tiered product.
In the desktop space, all of those secondary and tertiary timings being auto-set to something reasonable depends on the specific RAM kit being on that motherboard model's QVL. I have no idea if something similar is at play here, but if so, it could explain people getting poor results with anything other than 5600 MHz RAM similar to or better than what Framework sells.
~ 2,14ns difference from the CAS alone. Other timings are also shorter.
Given that total memory latency is often above 80ns not that much of a difference. And that assumes that AMD supports CL40 5600 Jedec. At least those Dimms should also support CL46 5600 Jedec, so that they should still work either way...
Would you mind passing the source on this? If so I'm not even gonna price track the CL40 kit.
It is 6 CL better