I have an Intel Core i7 6700HQ with a GeForce GTX 980M! Or an AMD A10-7870K with a R9 380X! But what does this all mean and how do you even begin to make sense of this? This is what this post is going to be about: making sense of these model numbers.
An example part can look like this:
The parts in bold are what’s most important to look at. Right off this tells me it’s a high-end processor in its class (i7), that’s of the 6th generation “Core” processor, and HQ means it mobile, quad-core part.
In general, the higher the “i” number and generation number, the more it performs and features it has. The letters are a different story, but Intel is consistent with them (this isn’t a comprehensive, it just has the ones you’re most likely going to encounter):
- [No Letter] - A desktop class part with no special features.
- H - Mobile part.
- K - Multipler is unlocked, allowing for easier overclocking.
- M - Mobile part.
- Q - Quad core mobile part. Note: This can trip you up if you aren’t careful. If you think a Core i7 is always a quad core part, you’ll be in for a surprise if you get a laptop with a Core i7-6500U as it’s a dual-core part.
- R - Contains the Iris Pro GPU
- S - Low power desktop part. Seems to be fallen out of use in Skylake (6th generation), due to the no letter desktop part having the same power rating as the S part.
- T - Lower power desktop part.
- U - Low power part, usually for Ultrabooks or 2-in-1's
- X - Extreme part, the highest-end on the consumer spectrum.
- Y - Ultra low power part, usually for tablets
The 700 is what Intel refers to as the “SKU” number. This breaks down further and worse, can be confusing. The number in the hundreds digit (7) is yet another tier level, the higher the better. However, the tens and ones digits have no clear pattern. A higher number can mean worse specs in one lineup, while a lower number can mean worse specs in another. The difference is usually too small to be appreciable.
In other words, determining what you’re getting in an Intel processor requires research. Bad move Intel, bad move. Thankfully Intel does catalog everything, so a Google search will get you what you want.
Intel simplified their motherboard naming scheme since the Core i# parts, thankfully. Today it’s a letter followed by a number, and there are fewer combinations to choose from. Regarding numbers, it’s important to pay attention to the second number from the left.
The difference between each chipset level is the number of SATA and USB ports are available and how many expansions are allowed on the chipset itself. Intel uses PCI Express lanes to hook up to peripherals on the motherboard, so lower end chipsets have less of these.
Note that all Intel processors (except the ones based off the Atom line) have at least 16 PCI Express lanes in the processor itself for a video card. The processor connects to the chipset via another bus called DMI.
The following is a list from best to worst, using only the letter and the second number for simplicity:
H1 -> B5 -> Q5 -> H7 -> Z7 -> X9.
A caveat with the X chipset family, they’re on a different development path than the rest. This means that the other chipsets may get more features since they’re refreshed more often.
At the moment, AMD has two kinds of processor lineups: the FX processors and the APUs.
The FX processors are pretty easy to figure out.
The “FX” just designates what lineup it’s in. The number in the thousands place, an 8 in this case, means how many cores there are except if it’s a 9, then it’s still 8 cores. The last numbers are just a performance indicator. Generally the higher this number, the better the performance. In fact, taken as a whole, the higher the number, the better it performs. Pretty easy, huh?
The APUs are where things get a little tricky... So here’s a few examples
Athlon X2 450
AMD separates their APUs into actual APUs and APUs without a GPU (or just the CPU part). The Athlon in this case is an APU based CPU.
A quick way to tell if the APU has a GPU is if the model number has four digits, not if it has an A# before the model number. In this example, the Sempron 3850 has a GPU in it, even though it doesn’t have an A# prefix. The Athlon X2 450 does not have a GPU.
If it has a four digit model number, the thousands place is a generation indicator, but only within that family. This makes it a headache sometimes to keep track of. For example, the A6-3650 does not come from the same processor family as the Sempron 3850. The A6-3650 is a Llano core while the Sempron 3850 is a Kabini core. But, you can still use the “higher the number the better performance” generalization as long as you stick with the same processor type. i.e., compare only the A series with other A series, Semprons with other Semprons, Athlons with other Atlons.
The letters at the end mean either an unlocked processor (K), mobile (M), or a second version (B). The B parts are annoying, because it can be unlocked if it was replacing a K processor or not if the processor didn’t have a K.
On a side note, the Althon X2 uses the same naming conventions that AMD used in the past. The X2 denotes it’s a dual-core processor. An Athlon X4 would be a quad core processor.
AMD motherboard chipsets are again, separated into two groups for the FX processors and APUs.
The FX processors are compatible with 800 and 900 series chipsets. The 800 ones lack SLI while only the 990FX can do it. Otherwise much like Intel, the difference is mostly between how many and what kinds of peripherals are available.
The APUs use the A-series chipsets. There’s no discernible difference between the model numbers unfortunately, but higher numbers generally mean better expansion and port options.
Unfortunately like the processors, the numbering scheme for Intel’s GPUs are all over the place. For example, the HD 5100 and HD 5200 GPUs are better than the HD 5300 and HD 5500.
However, higher end GPUs are paired with higher end processors usually. Starting with Skylake, or the 6000 series processors, Intel started using a three digit model number scheme.
NVIDIA Video Cards
NVIDIA simplified its naming schema around 2010, but has since introduced a few hiccups that I’ll explain in a bit. So let’s start with an example (note that you’ll most likely be buying a GeForce card, so I won’t prefix the model names with that):
GTX 780 Ti
Right off the bat you’ll notice that all of them start with “GTX”. Basically the “GTX” part or anything like that is useless. Originally NVIDIA wanted to use the letters G, T, and X to do a quick categorization of the GPU’s performance. G would be lowest, GT would be midrange, and GTX would be high end. The problem is that NVIDIA doesn’t offer anything lower than GTX anymore to consumers and applied it to the entire range.
For three of these GPUs, there’s a three number moniker (soon to be four with the GTX 1080). The hundreds digit is the generation of that part. The tens and ones are the performance tier. For some reason, NVIDIA has been allergic to model numbers not divisible by 10. If there’s a GPU that’s slightly better than one GPU but not enough to beat out the next tier above it, it’s give the “Ti” badge.
If the GPU is a mobile version, it will have an M at the end of the number, e.g., GTX 980M.
The Titan cards are something special, they are the highest tier in the generation. Unfortunately NVIDIA doesn’t say which generation it belongs to. But if you’re getting a Titan card, you probably know what you’re doing.
While you can generally use the “higher number is better” assumption, it’s not always the case. For example, the GTX 670 and GTX 760 have about the same performance, as does the GTX 980 and the GTX 780 Ti. In some sense, you can add the numbers together and get a result that way.
AMD Video Cards
AMD simplified the naming schema back in 2007 with the introduction of the HD 3000 series. There was no extra suffix except for the X2 moniker for dual-GPU cards. But in 2013, someone in AMD’s marketing decided to go “eh, screw that, let’s give our customers an edgy name!” at the expense of simplicity. Now, we have something like this:
Like NVIDIA did with the G, T, and X letters, the R# designation is supposed to mean what performance tier the card is in. You can ignore this because
- It can be derived from the model number much like the GeForce cards
- AMD contradicted itself. There’s an R9 370 and an R7 370 with the R9 370 being worse than the R7 370.
The number in the hundreds place means the generation. The next two numbers are the performance tiers. If there’s a suffix it means the card is better than the base model. So far there’s only E and X for better and best respectively.
Also like NVIDIA, they created an even higher tier of cards, the Fury series. And like its kind of counterpart-but-not-really the Titan, there’s no generation number to go by... but then again there’s only one generation of Fury, so far.
You can also use the “add all the numbers up” to get a performance comparison.
RAM “model numbers” is based on the type of RAM and its speed. There are two ways of doing this:
- PC# - #####
Usually the DDR spec is used. The DDR spec is what speed the RAM is running at in MHz. For example, DDR3-1,600 is third generation DDR memory running at 1,600 MHz. DDR4-2,666 is fourth generation DDR memory running at 2,666 MHz.
The PC spec is how much data in megabytes per second is transferred. The number is times eight the speed it operates at. So for example, DDR3-1,600 memory is the same as PC3-12,800 and DDR4-2,666 is the same as PC4-21,000 (technically 21,328, which is probably why they don’t use the PC spec as much)
RAM also has a latency component to it, usually a series of four numbers separated by dashes. For the most part, ignore these. While lower is better, this only applies to comparing models with the same speed rating, as latency is the number of clock cycles.