When buying your next laptop, all-in-one, or branded PC, you might notice that you have choices when it comes to the amount of RAM and storage, but you can't do much about the CPU that the manufacturer decides to go with for each model because of the complexities of design and pricing. It's good to understand exactly what you're getting before you spend your money, since it's usually impossible to upgrade a CPU somewhere down the line.
The only time that you really have to put a lot of thought into choosing a CPU is when building a desktop PC for yourself. In either case, people generally understand that newer generations, higher speeds, and more cores are better. There's a lot more complexity to the CPU selection process than that, though.
It's important to make sure your needs will be met, not just now but for the next several years, and that your money is being spent wisely. Here's our complete guide to understanding CPU specifications and features, as well as the names and numbers that today's manufacturers use.
There are only two CPU manufacturers in the world: Intel and AMD. Many people only know of Intel and its Celeron, Pentium, and Core processor lines, and that isn't a big surprise since this company has had the lion's share of the market for decades now. The other player is AMD, which has always had to fight for recognition and is only just emerging from a years-long period when its products were too weak to be taken seriously.
AMD's Athlon and Ryzen brands are familiar to DIY PC builders, but are virtually absent from the laptop market because it's hard for buyers to trust a relatively unknown name. Those who remember AMD from the 1990s and early 2000s associate it with chips that get very hot, which is unfortunate because that isn't true anymore.
The first thing you need to know is that CPUs for desktop PCs and laptops are quite different, even though they're sold under the same names. Laptops have constraints in terms of the amount of power that a CPU can draw and the amount of heat any component is allowed to generate inside such a small body. We're going to look at both AMD and Intel's product lineups one by one.
At the very low end, Intel has its Celeron and Pentium Silver CPUs. These are based on derivatives of the old Intel Atom processors, and they are designed to be affordable. You'll find them in low-end desktops as well as laptops, and they usually run without fans. They're enough to surf the Web and get basic work done with. You should be able to stream HD videos, but you can forget about playing games, and your experience will not be very smooth.
The latest Intel Celeron and Pentium Silver processors are codenamed ‘Gemini Lake' but there's no easy way to make sure you're getting one of these just by looking at the model numbers. You can search through Intel's online database of CPU specifications to confirm details.
Slighty higher up the ladder, we have the Pentium Gold series which uses a cut-down version of Intel's higher-powered Core architecture. Specifications are weak, in terms of the number of cores and the clock speeds you get, but these chips are not too underpowered for students and home users with light workloads.
The meat of Intel's lineup is in its Core i3, Core i5, Core i7, and Core i9 series. These tiers represent steps in terms of the amount of processing power you get. Generally, the lower tiers have fewer CPU cores and lower speeds, while the higher-end ones have all the bells and whistles including support for overclocking.
As of April 2019, you get two or four cores with the low end Core i3 options, and up to eight cores with Hyper-Threading (where a core can process two instructions at once) with the top-end Core i9 models. If you want to multitask between lots of apps at once, play intense 3D games, edit videos or work with 3D animation, you should go as high up the ladder as your budget will allow.
Intel also segments its CPUs by the type of devices they will end up in, and you'll notice this in the letters that follow the four-digit model numbers. Within the Core i3, Core i5, Core i7, and Core i9 families, there are huge differences in power, and this is all based on thermal conditions expressed in TDP (Thermal Design Power).
There's the Y-series for ultra-slim fanless tablets and laptops, with sub-5W TDP ratings, then the U-series for Ultrabooks with a 15W rating. Mainstream laptops get the H-series which is typically around 45W. Desktop CPUs range from 35W to 95W and most do not have letter suffixes on their model numbers.
The TDP of a processor lets you know how far it can be pushed before it exceeds the capabilities of the heatsink or fan attached to it. A higher TDP means that a CPU can run at higher speeds for longer. Once the heat limit is hit, the speed is reduced no matter whether there's still work to be done or not – this is commonly known as throttling. The slimmer and lighter a laptop is, the less hefty its cooling apparatus can be, so you get CPUs with lower TDPs that can't handle as much pressure as high-end ones.
An additional K at the end of a model number, such as the Core i9-9900K, means that a particular CPU model is unlocked and can be forced to run much faster than its rated speeds — although this will also push it past its rated TDP. Desktop models with an F at the end lack integrated graphics capabilities but are otherwise no different from their non-F counterparts.
The next kind of segmentation to understand is Intel's generations. In previous years, there were simple progressions from one generation to the next, bringing new improvements to both the design and manufacturing process of CPUs across the board. Sometimes, generations are rolled out segment by segment. Of late, however, Intel has been struggling to progress from one generation to the next, and so it has somewhat blurred the lines defining them.
Each generation used to refer to a single architecture, but that's a little more complicated now. We currently have 8th Gen chips in some segments and 9th Gen chips in others. The codename ‘Coffee Lake' has been used for 8th Gen and 9th Gen chips across the desktop and mainstream laptop categories, but 8th Gen U-series Ultrabooks are now codenamed ‘Whiskey Lake' while the super-slim Y-series category gets ‘Amber Lake'.
The first digit of each Intel CPU's four-digit naming scheme tells you which generation you're getting, so a Core i5-9600 is better than the Core i5-8600 but not necessarily better than a Core i7-8700. Newer platforms give you newer features such as faster USB and Wi-Fi, better security, and support for future upgrades. In rare cases, if you can get a good deal on an older high-end CPU, you should check online to make sure of its exact specifications. Don't go too far back, though - at this point, it's best to look beyond the 7th Gen and earlier models.
For those with high-end professional workloads such as 3D content creation, Intel offers its Core X-series CPUs based on the much older Skylake-X architecture, but with up to 18 cores. These are priced out of reach for most people, and are not really relevant for home users and gamers.
AMD spent the better part of a decade trying to catch up to Intel, but its CPUs just weren't good enough and not even the deepest of discounts made them worthwhile. All of that changed in 2017 with the launch of the AMD Ryzen series, an effort that was started from scratch. AMD's low prices and higher core counts have helped win many fans over, and in fact it was competition from AMD that spurred Intel to work harder and give buyers much more bang for their buck than were getting before.
This company still sells its legacy Athlon and FX-series chips, which should be avoided. The new Athlon 200GE series however is based on the new Zen architecture that underpins the Ryzen series, and these are well regarded for their low cost and relatively powerful integrated graphics capabilities.
The AMD Ryzen 3, Ryzen 5 and Ryzen 7 sub-brands mirror Intel's Core i3, Core i5 and Core i7 in a not-so-subtle move to establish their competitive positioning. Expectedly, you get more cores and higher speeds as you go up each level. The company has launched a wave of second-gen parts over the past year with model numbers in the 2xxx range for desktop CPUs, but confusingly, 3xxx models using the same architecture do exist for laptops.
Despite a resurgence with desktop DIY enthusiasts, AMD's Ryzen chips are only found in a handful of laptops. This might grow as the company continues to prove its capabilities, but if you're looking to buy a laptop, chances are you won't see Ryzen-based options in the market at all. On the other hand, the company's strategy of offering more cores and lower prices has earned it strong recommendations for creative professionals and gamers whose software can benefit from running multiple instructions in parallel.
In stark contrast to Intel's lineup, most AMD Ryzen CPUs do not have integrated graphics capabilities, which means they aren't suitable for low-cost PCs. The exceptions to this rule are the Ryzen 3 2200G and Ryzen 5 2400G – that G suffix is of course for graphics. These solutions are actually just as powerful as today's most entry-level discrete GPUs, so you could actually save some money or build now with a plan to upgrade later.
AMD has also decided not to restrict features such as overclocking to only its higher-priced offerings. There are some CPU models with an X suffix, such as the Ryzen 7 2700X, and these can be pushed further if you have a good enough cooling solution.
The Ryzen 3, Ryzen 5, and Ryzen 7 models are all priced very competitively, and AMD will refresh the whole lineup later this year with a major new architecture and manufacturing process that could give it even more of an advantage.
If you need to go way beyond eight cores, the high-end Ryzen Threadripper series offers between 12 and 32 cores. Prices are of course much higher, but still extremely reasonable compared to Intel's high-end desktop Core X-series.
Some people place a lot of importance on the manufacturing process of a CPU, which is expressed in terms of transistor size. As a general rule, the lower you go, the faster a CPU can run and the less heat it generates, but each step downwards involves tremendous cost and complexity for Intel and AMD. Intel has been stuck at 14nm for a while, and its struggles to transition to 10nm have been documented thoroughly, but we could see 10nm chips launching this year or in early 2020. AMD is currently at 12nm but will debut its 7nm chips in just a few months.
In practice, you can't compare processes between companies or even between different kinds of chips. Intel has refined its 14nm technology several times now, and its products are still very competitive in terms of performance. This should not be a reason to choose one brand over another.
When choosing your CPU, it's important to consider what motherboard, cooler, RAM, and power supply you will be using. Adding up the total costs of these components will also help you choose between Intel and AMD platforms. All current Ryzen CPUs come with highly capable coolers, but Intel assumes you'd prefer to choose your own. If you're overclocking, you'll need high-end components that are guaranteed to be stable under heavy load.
Speaking of cost, Intel is struggling to meet demand in the market due to manufacturing difficulties at the moment. As a result of this, there have been huge price spikes for many models, and several others are simply not available. AMD therefore enjoys quite an advantage in several categories where things would otherwise have been a lot more even between the two competitors.
Finally, we did say that there are only two CPU manufacturers in the world, but there is one emerging competitor — Qualcomm. This company's Snapdragon chips can be found inside millions of smartphones, and as they get more and more powerful, Qualcomm has been eyeing the PC market. Laptops featuring Snapdragon processors could arrive in India this year, with the promise of 4G connectivity and multi-day battery life.
What CPU do you think is the best for your demands and budget? Do let us know in the comments section.