⏱ 8 min read  ·  ✅ Updated Oct 2026

CAS latency, usually written CL, is the number of memory clock cycles a RAM module waits between receiving a read command and starting to return data. A lower CL means fewer cycles of waiting, but cycles get shorter as memory speed rises, so CL only means something when you compare it with the data rate.

That is why a DDR5 kit labeled CL30 can respond about as quickly as a DDR4 kit labeled CL16, even though the DDR5 number looks nearly twice as large. This guide gives the conversion formula, a comparison table across common kits, buying rules by platform, and a checklist for confirming the right profile is actually active. The information is drawn from published memory standards, manufacturer documentation and buyer reviews.

What CAS latency means

Memory is a grid of rows and columns. To read a value, the controller first opens a row, then picks the column, then waits for the data to appear. CAS latency covers the wait between the column command and the data. It is one of several timings, but it is the one printed on the box because it is the easiest to compare at a glance.

A typical kit label reads something like “6000 CL30” or “3200 CL16.” The first number is the data rate in megatransfers per second, and the second is CAS latency in clock cycles. If a listing shows only a speed and no CL, check the manufacturer’s specification page, since the same speed is sold with different timings at different prices.

Clock cycles versus real time

Memory transfers data twice per clock cycle, which is what “double data rate” means. So the actual clock runs at half the advertised data rate. A kit sold as 6000 has a clock around 3000 megahertz, and each cycle lasts about one third of a nanosecond. The conversion is short.

True latency in nanoseconds = CL x 2000 / data rate (in MT/s).

The factor of 2000 comes from converting megahertz to nanoseconds and accounting for the double rate. Run it for any kit with a calculator, and you can compare DDR4 and DDR5 fairly. Note this tells you CAS delay only, not total memory behavior, because other timings, the memory controller and the processor’s design all add to the real wait.

True latency table: DDR4 versus DDR5

Kit Calculation True CAS latency Comment
DDR4-2666 CL19 19 x 2000 / 2666 About 14.3 ns Older, entry level
DDR4-3200 CL16 16 x 2000 / 3200 10.0 ns Common baseline
DDR4-3600 CL16 16 x 2000 / 3600 About 8.9 ns Strong DDR4 pick
DDR4-3600 CL18 18 x 2000 / 3600 10.0 ns Same delay as 3200 CL16, more bandwidth
DDR5-4800 CL40 40 x 2000 / 4800 About 16.7 ns Base-speed DDR5
DDR5-5600 CL36 36 x 2000 / 5600 About 12.9 ns Mainstream
DDR5-6000 CL36 36 x 2000 / 6000 12.0 ns Popular and widely available
DDR5-6000 CL30 30 x 2000 / 6000 10.0 ns Often sold at a premium
DDR5-6400 CL32 32 x 2000 / 6400 10.0 ns Fast, check board support
DDR5-7200 CL34 34 x 2000 / 7200 About 9.4 ns High end, platform dependent

Look at how DDR4-3200 CL16, DDR4-3600 CL18, DDR5-6000 CL30 and DDR5-6400 CL32 all land at 10.0 nanoseconds. The CL number alone would have told you very different stories. What changes across those kits is bandwidth: the faster data rate moves more data per second once the wait is over. The real gain for DDR5 is usually that bandwidth and capacity, not a drop in CAS time.

Why DDR5 numbers look worse than DDR4

New buyers often see a DDR5 label such as CL36 next to a DDR4 label such as CL16 and assume the older memory is twice as quick. The conversion shows why that reading fails. DDR5 runs at far higher data rates, so each clock cycle is shorter, and a higher cycle count ends up costing about the same time. DDR5 also changes the module’s internal layout, with two independent channels per module and on-module power regulation, which helps bandwidth and efficiency but does not shrink CAS delay by itself. So the sensible approach is to compare nanoseconds for responsiveness and data rate for throughput, then check your motherboard’s supported speeds.

Worked example: two kits, one budget

Assumptions: you are choosing between two 32 GB DDR5 kits for a new build. Kit A is rated 6000 CL36. Kit B is rated 6000 CL30 and sits in a higher price tier. Both use the same data rate, so the comparison is straightforward.

Step Kit A Kit B
True CAS latency 36 x 2000 / 6000 = 12.0 ns 30 x 2000 / 6000 = 10.0 ns
Difference 2.0 ns on CAS delay, about 17 percent lower for Kit B
What it means in a game Usually a small effect, since CAS is one part of total memory delay and the GPU often limits frame rate
When Kit B makes sense CPU-limited play at high refresh rates, or a small price gap

If Kit B costs noticeably more, a better use of the money may be a stronger graphics card or more storage. If the price gap is small, take the lower CL. The same logic applies to every pair of kits: convert, compare the nanoseconds, then ask whether the gap is worth the premium.

Buying rules by platform

The processor’s memory controller sets what works well, so match the kit to the platform instead of buying the biggest number.

Platform type Rule of thumb Why
Recent AMD desktop on DDR5 Look at kits rated around 6000 with a low CL Many builders find this speed pairs well with the controller and fabric clock, though results vary by board and chip
Recent Intel desktop on DDR5 Higher speeds are often supported, check the board’s list Controllers vary by generation, and some boards struggle with very fast kits
DDR4 builds Choose 3200 to 3600 with CL16 to CL18 Best value region for the generation
Four modules instead of two Expect lower stable speeds More load on the memory controller
Mixed or leftover modules Avoid when possible The slowest timing sets the pace and stability is harder to guarantee

These are guidelines, not guarantees. Check your motherboard maker’s memory support list for the exact kit, and consider two larger modules rather than four small ones, which is easier on the controller.

How to verify and apply the correct profile

Buying a fast kit does nothing if the system still runs at the default speed. Follow these steps after installing the memory.

  1. Seat the modules in the recommended slots. For two modules, the board manual usually names the slots, often the second and fourth from the processor.
  2. Enter the BIOS and find the memory profile option. On many boards it is labeled XMP, DOCP or EXPO, depending on the vendor.
  3. Enable the profile and save. The system will restart, sometimes more than once while it trains the memory.
  4. Return to the BIOS or open a system information tool. Confirm the reported speed matches the kit label and that timings match the profile.
  5. Run a memory stress test. Use a recognized memory-testing tool for a full pass to catch instability early.
  6. Update the BIOS if training fails. New firmware often improves memory compatibility.

Hardware verification checklist

  • The kit’s speed and CL on the label match the product listing you intended to buy.
  • Your motherboard lists the memory type, capacity and speed you are installing.
  • Both modules are fully clicked into place with the retention clips closed.
  • The profile is enabled, and the reported speed is not the base default.
  • Task Manager or a system tool shows the full capacity, with no missing module.
  • No crashes, blue screens or failed boots during the first few days of use.

When to stop tweaking and ask for support

If the system fails to boot after enabling a profile, let it cycle and reset to defaults, or use the board’s clear-CMOS method from the manual. Drop to a lower speed or looser timings and see if stability returns. If random crashes continue at default settings, a faulty module is possible, so test each stick alone, then contact the memory maker or retailer. Do not raise voltages beyond the kit’s rating without understanding the risk, because excess voltage can shorten component life.

Related guides

Frequently asked questions

Is lower CAS latency always better?

Only when the data rate is the same. A CL30 kit at a high speed can have the same real delay as a CL16 kit at a much lower speed, because the clock cycles are shorter. Compare kits by converting to nanoseconds, then also weigh capacity, stability on your motherboard and price. A lower CL number alone does not decide anything.

Does CAS latency matter more than RAM speed?

Neither wins alone. Speed sets how often data moves, and CAS latency sets how many cycles you wait for the first piece. The useful comparison is true latency plus bandwidth together. For most gaming PCs the difference between sensible kits is small, so avoid paying a large premium chasing a lower number.

What are the other numbers in a timing like 30-38-38-96?

The first is CAS latency. The next two are the row-to-column delay and the row precharge time, and the last is the row active time. Together they describe how long the memory takes to open, read and close data rows. Kit makers print them on the label, and tuning tools can show them, but CAS is usually the headline figure.

Why is my RAM running slower than the speed on the box?

Most kits start at a safe default speed until you enable a stored profile in the BIOS, called XMP on many boards and EXPO on some AMD boards. Until you turn that on, memory runs at a lower standard speed with looser timings. Boards and processors can also limit the speed when four modules are installed.

Is it safe to enable XMP or EXPO?

These profiles are set by the kit maker and are widely used, but they count as running memory beyond the base specification and a few boards or processors may not hold them stably. If you see crashes or boot failures after turning one on, lower the speed or return to defaults and check the motherboard’s supported memory list or BIOS updates.

Explore Our Guides & Free Tools