RAM Frequency Calculator

Convert between real clock (MHz), effective transfer rate (MT/s) and clock cycle time — and end the 3000 vs 6000 confusion.

Updated 11 Nov 2026 · 6 min read

Why CPU-Z and Task Manager disagree

CPU-Z reports the real clock (e.g. 3000MHz for a DDR5-6000 kit). Task Manager reports the effective transfer rate (6000 MT/s). Both are correct — they are measuring the same thing in different units. Our lower-speed guide explains when a low number is a real problem (JEDEC fallback) versus just units.

This is the single most confusing thing about RAM speeds, and it comes from the name itself: DDR means Double Data Rate. The memory bus transfers data twice per clock cycle — once on the rising edge of the clock signal and once on the falling edge. So a memory clock running at 3000 MHz completes 6000 million transfers per second. CPU-Z shows you the clock (3000 MHz); Task Manager shows you the transfers (6000 MT/s); the box in the shop shows you 6000 and usually labels it ‘MHz’, which is technically wrong but universal.

How to use the calculator

Enter either number — the real clock in MHz or the effective rate in MT/s — and the calculator gives you the other, plus the duration of one clock cycle in nanoseconds. The two conversions are:

Cycle time is the third output, and it is the bridge between cycle-based timings and real time. A CAS latency of 30 cycles means 30 × the cycle time in nanoseconds — which is exactly how the latency calculator works. At 3000 MHz, one cycle lasts 0.333ns, so CL30 is 10.0ns. At 1600 MHz, one cycle lasts 0.625ns, so CL16 is also 10.0ns. Same real latency, different numbers on the box.

Worked examples

Example 1 — CPU-Z shows 3000 MHz. Effective rate = 3000 × 2 = 6000 MT/s. Cycle time = 1000 ÷ 3000 = 0.333ns. Your DDR5-6000 kit is running at full speed — the 3000 figure is the clock, not a problem.

Example 2 — CPU-Z shows 1200 MHz on a DDR4-3200 kit. Effective rate = 1200 × 2 = 2400 MT/s, cycle time = 1000 ÷ 1200 = 0.833ns. The kit is running at 2400 MT/s instead of its rated 3200 — a genuine JEDEC fallback, almost certainly because XMP is off in the BIOS. This is a real problem to fix, not a units confusion: Task Manager will also show 2400, confirming it. By contrast, CPU-Z showing 1600 MHz on the same kit means 3200 MT/s and a cycle time of 0.625ns — everything as expected.

Example 3 — the common speeds, converted.

Kit (as sold)	Real clock	Effective rate	Cycle time
DDR4-2400	1200 MHz	2400 MT/s	0.833 ns
DDR4-3200	1600 MHz	3200 MT/s	0.625 ns
DDR4-3600	1800 MHz	3600 MT/s	0.556 ns
DDR5-4800	2400 MHz	4800 MT/s	0.417 ns
DDR5-5600	2800 MHz	5600 MT/s	0.357 ns
DDR5-6000	3000 MHz	6000 MT/s	0.333 ns
DDR5-6400	3200 MHz	6400 MT/s	0.313 ns

What cycle time is used for

Cycle time is the bridge between cycle-based timings and real time: timing in ns = cycles × cycle time. Latency and timing calculators do this automatically.

Common mistakes

• Thinking CPU-Z’s number means half speed. CPU-Z shows the real clock. Double it to get the effective rate, and compare that with the box — not the raw CPU-Z figure.
• Trusting ‘MHz’ on the box literally. Manufacturers print the MT/s figure but label it MHz. Treat any RAM ‘MHz’ number on packaging or in a shop listing as MT/s, and halve it to get the clock.
• Mixing up MHz and MT/s when entering speeds into other calculators. The latency, bandwidth and timing calculators all expect MT/s. If you enter the CPU-Z clock figure (3000) where they expect the transfer rate (6000), every result will be wrong by a factor of two.
• Assuming a low Task Manager number is a units issue. Task Manager already shows the effective rate. If it reads 2400 on a 3200 kit, that is a genuine JEDEC fallback — check XMP/EXPO in the BIOS rather than reaching for this calculator.

Frequently asked questions

Is ‘MHz’ on RAM boxes accurate?

Manufacturers print the MT/s figure but label it MHz — technically wrong, universally done. Treat any RAM ‘MHz’ number as MT/s.

What is clock cycle time used for?

It is the bridge between cycle-based timings and real time: timing in ns = cycles × cycle time. Latency and timing calculators do this automatically.

CPU-Z shows half the speed on the box. Is my RAM faulty?

No — CPU-Z shows the real clock, which is half the effective rate by design. Double the CPU-Z figure: if it matches the box, the RAM is running correctly.

Does DDR5’s higher ‘MHz’ mean it is twice as fast as DDR4?

Not twice as fast overall. DDR5’s higher transfer rate means more bandwidth, but latency in nanoseconds is similar. Bandwidth roughly doubles across the generational jump; responsiveness does not.

Why do some tools show ‘MHz’ and others ‘MT/s’?

Older tools and marketing predate the MT/s convention, and the industry never cleaned it up. Modern specifications use MT/s because it is unambiguous; consumer packaging still says MHz because shoppers recognise it.