RAM Speed Calculator
Enter a DDR4 or DDR5 kit's speed and CAS latency to get its true latency in nanoseconds and peak bandwidth in GB/s, compare it with DDR4-3200 CL16, and see how it rates for gaming and work.
What the RAM Speed Calculator Works Out
Memory kits are sold by speed (for example DDR5-6000) and timings (for example CL30). Neither number alone tells you how quick the memory is: a higher speed with a looser CL can respond no faster than a slower kit. This calculator turns the two numbers into true latency in nanoseconds and peak bandwidth in GB/s, compares them with a common DDR4-3200 CL16 dual-channel setup, and gives a rough gaming and productivity rating that also considers capacity.
It works from the figures you enter. It doesn't detect your hardware, measure your system or predict frame rates.
How to Use It
- Find your kit's speed and CL. They're in the product name or on the module label, e.g. "DDR4-3600 CL16" or "6000 MT/s 30-38-38-96". The speed is in MT/s (megatransfers per second), often written as MHz.
- Enter total capacity: all modules together, e.g. 32 for two 16GB sticks.
- Choose the channel count. Most desktops are dual channel when two (or four) matching modules are installed in the slots the motherboard manual recommends; one module is single channel. Quad channel applies to workstation platforms.
- Pick your main use and click Calculate, or choose a preset.
The Formulas
true latency (ns) = CL × 2000 ÷ speed (MT/s) peak bandwidth (GB/s) = speed (MT/s) × channels × 8 ÷ 1000DDR memory transfers data twice per clock cycle, so the clock is half the rated speed: DDR4-3200 runs at 1,600 MHz. CL is counted in clock cycles, and one cycle at 1,600 MHz lasts 0.625 ns, so CL16 is 16 × 0.625 = 10 ns.
Each DDR4 channel, and each DDR5 module (as two 32-bit subchannels), is 64 bits (8 bytes) wide, which gives the 8 in the bandwidth formula. DDR4-3200 in dual channel is 3,200 × 2 × 8 ÷ 1000 = 51.2 GB/s.
| Kit | Channels | True latency | Peak bandwidth | Note |
|---|---|---|---|---|
| DDR4-2133 CL15 | Single | 14.06 ns | 17.06 GB/s | DDR4 JEDEC base speed, one module |
| DDR4-3200 CL22 | Dual | 13.75 ns | 51.2 GB/s | JEDEC DDR4-3200, typical without XMP |
| DDR4-3200 CL16 | Dual | 10.00 ns | 51.2 GB/s | Common XMP kit; the comparison baseline |
| DDR4-3600 CL16 | Dual | 8.89 ns | 57.6 GB/s | Common XMP kit |
| DDR5-4800 CL40 | Dual | 16.67 ns | 76.8 GB/s | JEDEC DDR5-4800, typical without XMP/EXPO |
| DDR5-6000 CL30 | Dual | 10.00 ns | 96.0 GB/s | Common EXPO/XMP kit |
| DDR5-8000 CL38 | Dual | 9.50 ns | 128.0 GB/s | High-end XMP kit |
DDR5-6000 CL30 and DDR4-3200 CL16 have the same 10 ns CAS latency, but the DDR5 kit has 1.875 times the bandwidth. DDR4-3600 CL16 has lower latency than either.
How the Ratings and Scores Work
The scores are Orbit2x's own rough weighting, meant for comparing kits with each other. They're not benchmark results.
- Gaming score: 60% latency, 30% bandwidth, 10% capacity. Games tend to be more sensitive to latency when the CPU is the bottleneck.
- Productivity score: 50% bandwidth, 40% capacity, 10% latency. Rendering, encoding, compiling and virtual machines lean on bandwidth and capacity.
- Mixed: the average of the two.
Each factor is scored in bands, for example:
- Latency: 8 ns or less is rated excellent, up to 10 ns very good, up to 12 ns good, up to 15 ns average.
- Bandwidth is rated on separate scales for DDR4 and DDR5, since DDR5 platforms start at higher bandwidth.
- Capacity is rated against the use: 16GB is very good and 32GB excellent for gaming; productivity needs 64GB for the top rating.
Practical Tips
- Enable XMP or EXPO. Kits rated above JEDEC speeds start at a JEDEC default (often DDR4-2133/2400/3200 or DDR5-4800) until you turn on the profile in the BIOS. Check the speed your system reports against the speed on the box.
- Use matched pairs for dual channel. A single module halves peak bandwidth compared with two in the right slots.
- Check what your platform supports. The CPU's supported memory speed and the motherboard's memory support list (QVL) say which kits are expected to run at their rated speed.
- DDR4 and DDR5 aren't interchangeable. They use different slots, so moving to DDR5 means a motherboard and CPU that support it.
- Prefer one matched kit to mixing kits. Mixed modules usually run at the slowest common settings and may be less stable.
- Running out of memory matters more than speed. If you regularly use most of your RAM, more capacity helps more than a faster kit.
Limitations
- "True latency" here is CAS latency only. Latency measured by memory benchmarks is several times higher, because it includes other timings and the memory controller.
- Bandwidth is the theoretical peak. Measured throughput is lower and depends on the CPU and memory controller.
- Secondary timings (tRCD, tRP, tRAS and others), memory rank and the CPU's memory architecture also affect performance and aren't modelled.
- How much memory speed changes frame rates depends heavily on the game, the CPU, the GPU and the resolution.
How Your Input Is Handled
The numbers you enter are sent to Orbit2x's server, which calculates the results and sends them back. They aren't saved. See the privacy policy for how the site handles data generally.
Frequently Asked Questions
Is RAM speed in MHz or MT/s?
The number in a kit's name, such as 3200, is the data rate in MT/s, though it's often labelled MHz. The actual clock is half that: 1,600 MHz for DDR4-3200. CPU-Z reports the clock as DRAM frequency, which is why it shows half the rated number; Windows Task Manager shows the data rate.
Is a higher speed always faster?
Not on its own. DDR4-4000 CL19 has 9.5 ns true latency, slightly worse than DDR4-3600 CL16 at 8.89 ns, though it has more bandwidth. Compare both figures, then check that your platform runs the faster kit at its rated speed.
Is DDR5-6000 CL30 better than DDR4-3600 CL16?
DDR5-6000 CL30 has 10 ns true latency against 8.89 ns, and 96 GB/s peak bandwidth against 57.6 GB/s in dual channel, about two-thirds more. Which matters more depends on the workload, and DDR5 needs a DDR5 motherboard and CPU.
Will faster RAM increase my FPS?
It can when the CPU is the limit, typically at high frame rates or lower resolutions. When the graphics card is the limit, for example at 4K with high settings, memory speed usually makes little difference. Benchmarks of your specific game and CPU are the best guide.
Why does my PC show a lower speed than my RAM's rating?
Usually because the XMP or EXPO profile isn't enabled, so the modules run at their JEDEC default. If they're enabled and the speed is still lower, the motherboard or CPU may not support that speed with your number of modules.