I'm back my friends
one thing is bothering me the specs of the i7 14700k or any other CPU I did look at does say :Memory Types Up to DDR5 5600 MT/s
Manufactures producing 7200mhz ram and what not all above the 5600mhz so that triggered me to do some research and what i have found out is that ram over the CPU specs just gets slowed down to the max speed of the CPU that would be 5600 MHz...so why do people use it and Manufactures produce it where there is no use for it?
Sure you can use it. Of course, DDR5-7200 is getting up there, it becomes a bit difficult to run (even for a kit like 2x 16 GB on Intel 14th gen), but for example up to DDR5-6000~6400 should not be that big of a deal with 2x 16 GB. Enable XMP and see if it works.
See
here on what factors it depends, and at
this particular point in that video, it explains how and why Intel and AMD come up with these low numbers. In reality, there's speeds that are generally accepted to work well, there's a "sweet spot", and then there's speeds and configurations that are likely to cause trouble. On Intel (13th and 14th gen), all of these are slightly higher than on AM5.
If you prioritize problem-avoidance over everything else, use up to DDR5-5600. But DDR5-6000 is more of a sweet spot. DDR5-6400 and above can be hit or miss sometimes, especially on AM5, but on Intel it should still be doable with two modules of 2x 16 GB (which is the optimal RAM kit with DDR5 for the vast majority of people). Two bigger modules, like 2x 32 GB, and you would probably do good to restrict yourself to DDR5-6000.
You have to realize two things about the DDR5 speed that is listed:
1) The CPU manufacturer is low-balling the number.
2) The board manufacturer is listing a high marketing number.
The optimum between problem-free operation and nice performance is somewhere in between, therefore. It also highly depends on things like your CPU's memory controller, the motherboard model, the RAM module size (12/24 GB modules = easier to run vs. 32/48 GB modules = harder to run). And you should always
avoid using four modules.
Timings only mean something in relation to the speed. Certain timings at a certain RAM speed result in an absolute latency of certain nanoseconds. Also see
here. So you don't compare timings directly, they cannot be compared at different clock speeds. They have to be calculated into the absolute latency in nanoseconds to be comparable.
The "Row-Active-Time (tRAS) 32" on the website for the second kit is of course nonsense. Kingston don't list the tRAS in the specs. The timings all depend on each other. For tRAS, the minimum should be tRCD + tRP, and often times, a couple more clock cycles are added. So with tRCD 36+ tRP 36 on that kit, you'd have a minimum tRAS of 72, but the common value would be 76.
The timings are interesting insofar as they can tell you what kind of memory ICs (chips) are used. For the tightest timings (especially low tRCD & tRP) they have to use high-end ICs, for looser timings they can use cheaper ICs. The better the ICs, the more you can tune it later, because especially the secondary and tertiary timings are still often quite loose, even if the primary timings are tight. This aids stability of course.
Tuning the RAM is a whole science. You can spend/waste days on optimizing all the timings if you really want perfection, and often it's not really that worth it (this is coming from a guy who has indeed spent days on it at some point). The main timing to optimize especially on DDR5 is called tREFI (Refresh Interval), which is the time interval between so-called refresh cycles for the RAM, which you want as infrequent as possible (while not being so infrequent that you have data loss). Because during a refresh cycle, no other data transfers can happen. So, the higher you can set tREFI too, the more time there is for data transfers in between and the better the performance. But theoretically it's possible to set this so high that some RAM cells "forget" part of their content because it didn't get refreshed in time. And this is difficult to test for too. So normally you'd set tREFI to some well-performing, but reasonably safe value, like 30000 or 32000. This alone would achieve more than tinkering with most of the secondary and tertiary timings.
The first 3 numbers are like 2 ns apart, but the 4e is a huge gap
As mentioned, the timings are not given in nanoseconds, but in clock cycles. Only when combined with the DDR speed, you get absolute latencies (not timings) in nanoseconds.
Example for CAS Latency (same for other timings):
DDR5-5600 CL28 = 10 ns latency
DDR5-5600 CL30 = 10.71 ns
DDR5-5600 CL36 = 12.86 ns
DDR5-6000 CL30 = 10 ns
DDR5-6000 CL32 = 10.67 ns
DDR5-6000 CL36 = 12 ns
DDR5-6000 CL38 = 12.67 ns
But DDR5-6000 has a higher bandwidth.
