I'm comparing 32GB DDR5-6000 kits with the same specifications except for CAS latency. The CL30 kit costs about 3,299.99 RON (around $696), so I'm wondering whether CL32, CL34, or CL36 would noticeably reduce gaming performance. How do the latency and real-world differences compare, especially with a Ryzen 7 9800X3D, RTX 5070 Ti, B850 motherboard, and high-end AIO? I mainly play competitive Fortnite. Also, would a good 850W power supply be sufficient instead of a 1000W model?
3 Answers
A quality 850W ATX 3.0 or ATX 3.1 power supply should be enough for a 9800X3D and RTX 5070 Ti, assuming the manufacturer’s recommendations are met. A good 1000W unit is also fine and gives more headroom, but capacity alone does not determine quality. Look for a reputable Gold-rated model with the proper modern GPU power connector rather than choosing solely by wattage.
For high-refresh competitive gaming, CL30 can provide a small advantage when the system is CPU-limited, but it won’t transform performance. If CL36 is substantially cheaper, it is still perfectly usable. Check the full timings and whether the kit supports AMD EXPO rather than focusing only on the CAS number.
At DDR5-6000, the approximate first-word latency is CL30: 10.0 ns, CL32: 10.67 ns, CL34: 11.33 ns, and CL36: 12.0 ns. In actual games, the difference is usually tiny. With a 9800X3D, CL30 versus CL36 will often be difficult to measure, particularly at 1440p or 4K. In very high-FPS, CPU-limited competitive games such as Fortnite, you might see a small difference in average FPS or 1% lows, but generally only a few percent at most. Buy the cheaper kit unless the price difference is very small.
Tighter secondary timings can improve 1% lows in some games, but tuning them manually takes effort and the premium for CL30 usually isn’t worthwhile. A decent kit with an EXPO profile is the more sensible choice.
I mainly play competitive Fortnite, so I’m more likely to be CPU-limited than someone playing single-player games at 4K.