AMD's laptop processor numbers, decoded digit by digit

AMD publishes the rule for its mobile processor numbers openly, which is more than most naming schemes manage, and the rule is genuinely useful once you know it. The catch is that one of the digits encodes the year the part was brought to market and a different digit encodes the architecture inside it. Those two can be several years apart in the same processor, and that gap is where buyers lose money.

Read the digits in place order

A four-digit Ryzen mobile number is read as: model year, then tier, then architecture generation, then a feature or SKU digit. The suffix letters after the digits describe the power class — the same job Intel's suffix does.

What this means concretely: a processor sold in a laptop launched this year can carry an architecture digit pointing at a design two or three generations old. The part is new stock. The silicon design is not. Both facts are true, both are disclosed in the number, and a listing that only prints the tier ("Ryzen 7") hides the one that matters.

The digit that predicts performance

It is the architecture digit, not the model year digit and not the tier. Two chips with the same tier and the same model year but different architecture digits are different products with different performance, different integrated graphics and different platform features. Two chips with the same architecture digit and different model years are, broadly, the same design.

So the comparison procedure is: line the candidate processors up, write out their four digits, and compare the third digit first. If it differs, that is your primary difference and everything else is secondary. If it matches, you are comparing siblings and the tier digit is meaningful.

The suffix letters

The letters at the end place the chip in a power class — the ultra-low-power class for fanless and very thin machines, the mainstream mobile class, and the high-power class for large chassis. As with Intel, the manufacturer publishes a default power figure and a configurable range for each part on its own product pages, and laptop makers choose where in that range to configure the chip. Two laptops with identical processor numbers can be configured at different points in that range; the machine with the bigger cooling system will generally be configured higher, and it will be faster in anything sustained.

The configurable range is published. It is worth looking at, because its width tells you how much of the outcome is in the laptop maker's hands rather than the chip's: a part with a wide configurable range is a part whose real performance depends heavily on the chassis.

Integrated graphics follow the architecture digit

AMD's integrated graphics have been the strongest part of its mobile line for several generations, and they track the architecture digit rather than the model year. A chip with an older architecture digit carries the older graphics architecture with it, whatever year it was sold in. If integrated graphics matter to you — which they do if you play anything at all on a machine without a discrete chip — this is the line to check.

As with every integrated graphics part, the memory configuration caps it. A single memory module halves the bandwidth available and the effect on graphics is large and immediate. See memory channels for the mechanism and how to check what a laptop actually ships.

What the number does not tell you

• Whether the memory is upgradeable. That is a chassis decision, not a processor one, and on many current designs the memory is on the package. See how to tell before you buy.
• What the machine will sustain. Chassis and configuration, not the part number.
• Which platform features are enabled. Laptop makers can and do omit features the chip supports — a storage slot wired with fewer lanes, a port without display output. Those are in the laptop's specification, not the processor's.

A worked reading

Take any Ryzen mobile number from a listing you are considering. Split it into its four digits and its suffix. Say out loud: "brought to market in [first digit] year, [second digit] tier, [third digit] architecture, [suffix] power class." Then look up that exact string on AMD's product pages and write down the default power and the configurable range. Now do the same for the competing machine. You will usually find that the difference you were about to pay for is in the tier digit, and the difference that will actually affect you is in the architecture digit or the power class — and those two often point the other way.