Does DDR5 RAM's CAS Latency Really Matter for 4K Gaming?

[DarkShadowFaisal]

Hi,

Recently i just upgraded my Gaming Rig & i only use my Desktop PC for 4K Gaming (No Streaming).

Never play: COD Warzone, PUBG, CS, Apex, Valorent types of Games.

Mostly Playing: FIFA and other story Based Games like Tomb Raider, Far Cry, Assassin Creed, Battlefield, TC Breakpoints, RDR etc.

 

PC Spec:

Motherboard: MSI MPG Z690 Carbon Wifi

CPU: Intel Core i5 13600KF

Cooling: Custom Liquid Cooling Loop

RAM: G.Skill Trident Z5 RGB DDR5-5600 CL30-36-36-89 (16GB Single Stick)

GPU: ASUS ROG Strix LC RTX 3080 Ti

Display: Gigabyte M32U (144Hz IPS 4K)

Samsung 970 EVO Plus NVMe M.2 SSD

PSU: ASUS ROG Thor 850W Platinum

OS: Windows 10

 

I want to change my DDR5 RAM, currently i am using 1 single stick of 16GB 5600MHz (CL 30 may be). I want to change to 8GB 2 stick 5200MHz (CL 40).

I want to know that in DDR5, is there any performance difference issue in 4K Gaming if i switch to CL 30 to CL 40

Also want to know is there any performance difference issue in 4K Gaming if i switch to 5600Mhz to 5200Mhz

Be informed that i only use this setup for 4K Gaming (No Streaming), so it doesn't matter if there is any performance difference issue in other cases like video editing, rendering, Streaming etc. as i don't do those things.

[RONOTHAN##]

CAS Latency is measured in clock cycles, so CL30 by itself means absolutely nothing. CL30 at 6000MT/s will end up with the same theoretical latency as say CL34 at 6800Mt/s or CL28 at 5600MT/s (though practically the 6800MT/s kit will have the lowest actual latency for reasons I'll get to later). 

 

With DDR5 in particular, primary timings only matter if you plan on doing RAM overclocking. If you aren't, they can be pretty much ignored as they have such a minimal impact on actual RAM performance. What matters with DDR5 is the actual frequency, with the only timings that have a noticeable impact on performance being the sub timings (not the ones advertised). 

 

That said, there's a problem with your plan. The sticks you're planning swapping to are 1Rx16 organized (it's the only way 8GB DDR5 can be made), which in general has 60% the memory bandwidth as the 1Rx8 organized stick you're currently using. Having 2 sticks with dual channel does double the memory bandwidth available, but the frequency and timing disadvantage of that new kit would mean that performance practically speaking will be near identical. This just isn't worth the time or effort to do. If you want to do a RAM upgrade, buy a second stick and upgrade to 32GB, that way you get a proper dual channel setup with proper 1Rx8 DIMMs which should be the best overall performance. 

[Somerandomtechyboi]

1 hour ago, DarkShadowFaisal said:

RAM: G.Skill Trident Z5 RGB DDR5-5600 CL30-36-36-89 (16GB Single Stick)

Hynix m die bin, apparently gskill has a 50/50 chance of hynix a die on their m die bins so check thaiphoon burner for ic

 

Simple as buying either hynix m or a die depending on what thaiphoon burner says for your current rams ic, theres only 2 ic choices with 5600c30 bin which is hynix m or a die so you wont be mixing your decent rams with shittier ics like 16gbit ddr5 b die or god forbid 16gbit micron a(ss) die, and the worst that can happen is mixing hynix m and a die which doesnt really matter unless your current kit is a die but still doesnt really matter that much anyways cause 7200 ish ddr5 should ride out the entire ddr5 generation, if you have a die and get another a die now youve got 8000+ ddr5 so you are guaranteed set for the entire ddr5 gen with those 2 sticks

 

And no cl dont do shit for performance if that wasnt obvious enough already (i mean why would 3600c18 be better than 3200c16 if cas is looser?)