Back

RAM Latency Calculator

Convert CL, tRCD, tRP, and tRAS timings into true latency in nanoseconds and compare DDR4 and DDR5 kits side by side.

MT/s
clk
clk
clk
clk

Timings are in memory clock cycles. True latency in nanoseconds is derived from the data rate.

True Latency

First-word latency (CL)

10.00 ns

Time from a read command to the first bit of data arriving.

I/O bus clock

3000 MHz

Cycle time

0.333 ns

tRCD delay

12.67 ns

tRP delay

12.67 ns

tRAS

26.00 ns

tRC (tRP + tRAS)

38.67 ns

Peak bandwidth per channel48.0 GB/s

Assumes one 64-bit channel. Dual channel doubles this, quad channel quadruples it.

How much RAM do you need?

This calculator compares RAM speed and timings. If you came here to find out how much RAM to buy, here's a quick answer for common uses.

Office and browsing8-16 GB
Windows 11 and Microsoft 365 each need at least 4 GB, and Copilot+ PCs start at 16 GB.
Gaming16-32 GB
16 GB is the common floor for modern games. Go for 32 GB if you keep Discord or a browser open while you play, and a few games already recommend it.
Video editing32 GB for 4K
On Windows, Adobe Premiere recommends 16 GB for HD and 32 GB or more for 4K. DaVinci Resolve asks for 16 GB, or 32 GB with Fusion.

Running AI models locally? What you can run depends on VRAM or unified memory more than on system RAM. The LLM VRAM Calculator shows which models fit.

How to use this tool

  1. Enter the memory data rate

    Type the speed in MT/s as printed on the kit (e.g. DDR5-6000 gets 6000, DDR4-3600 gets 3600). The calculator uses the data rate (not the internal clock), so enter the big number on the box.

  2. Enter the CAS Latency (CL)

    Type the CL timing, typically a small number like 30, 36, or 40 for DDR5 or 16 to 18 for DDR4. CL is the headline timing for first-access latency.

  3. Read the true latency for CL

    The calculator returns true latency in nanoseconds using the formula (CL × 2000) ÷ MT/s. Kits with the same MT/s can land at different ns figures depending on CL, and ns lets you compare kits across generations.

  4. Compare a second kit side-by-side

    Use Compare mode to enter a second memory kit. The tool compares each timing and peak bandwidth so you can see whether a pricier low-CL kit is actually faster than a cheaper kit with a higher data rate.

About this tool

The RAM Latency Calculator converts memory timings into true latency in nanoseconds, which lets you compare kits across DDR3, DDR4, and DDR5 on equal footing. Two DDR5 kits can advertise the same MT/s but land at different latencies depending on their CL rating, and a fast DDR4 kit can sometimes undercut a slow DDR5 one on first-access speed.

Use it when you're shopping for memory and trying to decide between two kits with different speeds and timings, or when you're tuning BIOS and want to see what dropping from CL 30 to CL 28 does at the same data rate: at DDR5-6000, true latency goes from 10.0 ns to 9.3 ns. Enter the speed in MT/s and CAS Latency, and the calculator returns true latency for CL, tRCD, tRP, and tRAS in ns.

Formula

True latency (ns) = (timing × 2000) ÷ data rate.

The factor of 2,000 is because DDR (Double Data Rate) transfers twice per clock cycle, so a kit labeled DDR5-6000 has an actual 3000 MHz internal clock and a 0.333 ns cycle time. Multiply the CL by that cycle time to get CAS latency in nanoseconds.

When to use it

Use it to compare memory kits with different speed and CL combinations before buying, check whether a pricier low-latency kit is actually faster than a cheaper kit with a higher data rate, or tune XMP/EXPO profiles manually. To see common kits ranked in one list, use the CAS latency chart.

Pair with the Memory & Storage Latency Visualizer to see how memory latency compares against L1/L2/L3 cache and NVMe, and the Power Cost Estimator to weigh running cost across different platform configurations.

Why true latency matters more than MT/s

Manufacturers print the data rate in MT/s in large type on the box, but data rate is only half the picture.

A 7200 MT/s kit might sound faster than a 6000 MT/s kit. With loose timings such as CL40, though, it has a true latency of 11.1 ns, so it takes longer to answer the CPU's first request than a tight 6000 MT/s CL30 kit at 10.0 ns.

In latency-sensitive workloads like high frame rate gaming or large database queries, the time it takes to fetch the first piece of data often matters more than the raw bandwidth available once the transfer starts.

Popular memory kits

Pre-computed true latency and bandwidth for common DDR4 and DDR5 kits.

Memory generation comparisons that put raw timings into context.

Frequently asked questions

What's the true latency of DDR5-6000 CL30?
True latency = (CL × 2000) ÷ speed = (30 × 2000) ÷ 6000 = 10 ns. That's low for a mainstream kit. DDR5-6000 kits with stock style timings usually sit around 15-16 ns, for example CL46 is 15.3 ns and CL48 is 16 ns.
How do CL and MT/s affect RAM speed?
Data rate in MT/s determines how much bandwidth you get, and CL determines how long you wait for the first bit. Higher MT/s plus lower CL means more throughput and faster access. A kit with twice the data rate and equal CL halves the true latency in nanoseconds.
Is DDR5-6000 CL30 better than DDR4-3600 CL16?
On bandwidth, yes: DDR5-6000 has nearly 67% more, which usually matters more for modern CPUs, especially AMD Ryzen 7000/9000 and Intel 12th-gen+. On latency, DDR4 is slightly lower: DDR4-3600 CL16 has a true latency of 8.9 ns against 10 ns for DDR5-6000 CL30.
What are tRCD, tRP, and tRAS?
Secondary timings. tRCD is the delay to activate a row before accessing a column. tRP is the time to close a row before opening another, and tRAS is the minimum time a row must stay open. All three are measured in cycles, and converting to ns uses the same formula as CL.
Does faster RAM help gaming?
On AMD Ryzen and modern Intel platforms, yes, but the gain is usually small. It shows up mostly in CPU-bound 1080p titles. At higher resolutions the GPU becomes the bottleneck and RAM speed matters less.
What is JEDEC vs XMP/EXPO?
JEDEC is the default SPD profile every DIMM boots with, with conservative timings guaranteed to work anywhere. XMP (Intel) and EXPO (AMD) are overclocking profiles stored on the DIMM that you enable in BIOS to unlock the speeds printed on the heatspreader.
What is a good RAM latency for gaming?
For a modern gaming PC, a true latency around 10 ns is a good target (for example DDR5-6000 CL30 or DDR4-3200 CL16). Anything under 12 ns performs well in most titles. Below 9 ns, the gains shrink compared with the extra cost of the kit.
Should I prioritize higher MT/s or lower CL?
MT/s first, up to the data rate your CPU's memory controller handles best, then lower CL. For example, AMD Ryzen 7000 processors perform best around 6000 MT/s because that's about the fastest the memory controller runs 1:1 with the memory clock on most chips. Once you hit that data rate, spend any remaining budget on the lowest CL you can get at that speed.

Use this tool on your own site

Drop the RAM Latency Calculator into your blog, docs, or CMS with one iframe snippet. Free, no signup. See all embeddable tools