What makes a processor the best? Some would say it's the price tag, and if money were no object this might hold weight. Others suggest that it's the number of cores or overclockability that determine how high a particular central processing unit (CPU) ranks in a price/performance chart.
In reality, it's simply a matter of taste, only made worst by the plethora of other variables to be taken into consideration. Which is why this lists contains a bit of everything; older processors, AMD APUs and even the odd Pentium.
You would be disappointed after handing over a small fortune to build a machine used exclusively for cranking out documents. At the same time, it would be an equally misguided decision to skimp on the CPU in your gaming rig just to save money.
In the end, it's not about achieving the best performance possible, but to do so and still meet a personalized budget. Certainly, if funds were limitless, we would all be rocking supercomputers.
Fortunately, there are plenty of less expensive options from both AMD and Intel focused on everything from word processing to gaming. With AMD’s latest 14-nanometer Zen CPUs all but released, Intel may have the upper hand in performance and value, but that’s subject to change very soon.
What's more, you'll have to take into consideration your PC’s other components. Pairing the best graphics card with a cheap processor would result in a severe bottleneck. Whether you're building a budget PC or a premium-priced server machine, you'll want to keep your parts consistent.
Overclocking, too, is worth your consideration. It's completely understandable if you're reluctant to push your chip beyond its stock speeds. But, if it piques your interest down the line, it’s safer to plan ahead and buy an unlocked CPU than to not.
With this list, we're focused purely on readily available desktop and server processors; laptop CPUs, obsolete sockets and non-x86 parts will be ignored. So without further ado, here are our picks of the top 10 best processors you can buy right now for your desktop PC.
If you are an AMD enthusiast (or just prefer rooting for the underdog), these are interesting times indeed. AMD is about to launch a series of processors based on a new architecture (Zen) which will obliterate the current generation of CPUs. So, of course, prices are falling accordingly. The A8-7670K remains one of the rare bright spots in AMD's lineup despite being more than two years old.
It is built on a newer 28nm manufacturing process which kind-of explains why it has a 95W TDP – thermal design power, or a part's share of your power supply's available Watts – despite a relatively high base and turbo clock speed (3.6GHz and 3.9GHz). Its graphics performance is where it shines thanks to an onboard GPU that is slightly more powerful than the Radeon R7 240 GPU (six compute units, 384 shader cores, 757MHz GPU clock speed).
Read the full review: AMD A8-7670K
There are many little-known secrets in the world of computer hardware. One of the best-kept is that, occasionally, data centers – globally operated by some of the biggest tech companies in the world – dump their processors as they upgrade to newer models. In effect, you probably can’t help but wonder what they do with all of their used-up goods.
In most cases, they wind up on eBay or Amazon from second-hand sellers. From there, you can purchase these older, “slower” CPUs for a fraction of their going rate. The Sandy-Bridge E5-2670 v1 is one of them; you can nab it for one-tenth of what it would have cost to buy new. Snag a pair of Xeon twins and you’ll be ready to assemble a hardy workstation rig, complete with a heart-melting 16 cores, 32 threads and 40MB cache.
If you want to do some heavy lifting but don’t want to blow your savings on a piece of silicon, then check out this chip. The Intel Core i3-6100 is the cheapest Core processor based on the new Skylake architecture, and you don’t have to fork out a fortune for it.
True, you’ll want to pair it with a motherboard with a decent chipset (Z710) in order to run faster memory (2.66GHz), but that isn’t necessary. It is not a K-model, and there are two SKUs, the 6100 (higher TDP and higher clock speed) and the 6100T (lower TDP, lower clock speeds) so make sure you choose the right one.
Using a 14nm node, it reaches 3.7GHz with a 65W TDP; its dual-core/4-thread configuration should make for a decent gaming rig, and the 4K-capable Intel HD 530 GPU is clocked at 350MHz. Oh and it should make a fairly good overclocker as well.
At the other end of the spectrum is the Sempron 3850, one of AMD’s cheapest quad-core processors. It sports a Kabini core and is built on a 28nm process, which explains why its TDP only reaches 25W, almost one seventh of the FX-9590.
Obviously, the fact that it runs at only 1.3GHz also helps a lot. Add in the fact that it comes with an integrated AMD Radeon HD 8280 GPU (basic, but decent) and you get something that’s better than most Baytrail-based systems at least. The best part though has to be the price; it is cheap especially, as it includes the heat sink and the fan; that means that you can envisage getting a motherboard bundle for less than Intel’s cheapest CPU. A shame that it has only one memory channel though.
Note that the Athlon X2 340, a different beast altogether (a different socket and no GPU), is AMD’s next cheapest processor, costing about 10 percent more.
Its predecessor, the Pentium G3258, was one of the best options at the lower end of the market, and the G4400 is a shoe-in replacement offering Skylake architecture on a budget. The laws of supply and demand means that it is actually retailing for cheaper than the G3258 or the slower Celeron parts.
This chip sports 3MB of L2 cache, hits 3.3GHz on its dual-core, two-thread setup with a TDP of 54W. It doesn’t have an unlocked multiplier like the G3258 though but its improved processor graphics (HD510) can support DirectX 12, 4K resolutions at 60Hz over DisplayPort which makes it great for a home theatre PC.
Note that there is a slower version (G4400TE, with a much lower TDP), which is great if you want an even quieter model to build your rig on.
This is Skylake, Intel’s sixth Core generation. The i7-6700K, which cost just under $345 (£290, about AU$463), is the company’s most powerful Skylake model set to replace the Broadwell-based desktop processors in the short term.
Here we’ve got a pretty powerful processor boasting four cores, eight threads, 8MB cache, a base clock speed of 4GHz, a turbo-boost of 4.2GHz and an Intel HD Graphics 530 subsystem inside. Overclocking is what may get some of us excited, however, as it’s the distinguishing feature of the “K” models such as itself.
Pair that with a decent 100-series chipset, an oversized HSF and a couple of overclocker-friendly DDR4 memory modules, and watch it fly. And, although you’ll want to pay close attention to that 91W TDP, 5.0GHz isn’t a lofty goal with the 6700K.
Read the full review: Intel Core i7-6700K
There is a good reason why the Intel Core i5-4690K is among the best-selling processors on Amazon.
This Devil’s Canyon part is one of the most, if not the most affordable K-series processor from Intel’s Core range at $239 (£182, about AU$321) and as such can overclock fairly easily with modest efforts. It has a base frequency of 3.5GHz with many users reporting being able to hit 25% increase in speed using a decent aftermarket HSF.
The 4690K doesn’t come with hyper-threading, but for the price it wasn’t expected. The processor, bilt on the 22nm fabrication process, packs 6MB of L2 cache, an 88W TDP and even an Intel HD Graphics 4600 onboard GPU.
Meet the AMD FX-8320E; this is one of the cheapest eight-core processors on the market and costs a smidgen under $110 (£108, about AU$148) on Amazon.
Built on a mature 32nm node, it’s clear why the FX-8320E has such a high TDP (95W). Then again, maybe it’s not an unusual spec given the 3.2GHz clock speed. Plus, when needed, it can even boost all the way to 4GHz.
But don’t get your hopes too high, though. On most tasks, the FX-8320E will be outperformed even by a modest Haswell Core i3. Where it truly shines is when you throw multi-threaded jobs (encryption, encoding etc) at it, where it can beat even the more expensive Core i5 parts. What’s more, many users have been able to overclock the chip easily using a non-stock heatsink fan, some all the way up to 4.8GHz.
With AMD ceding a bit of the limelight the past few years, Intel has gone back to releasing products that require a new socket on a quasi-yearly basis – great for sales, not great for customers. The Core i7-5820K, a Broadwell component, is no exception. It uses yet another socket aimed at the high-performance and server market.
What makes this little processor worth listing here is that even though it has all six cores intact, it isn’t priced outrageously. At $389 (around £313, AU$507), it’s really not a bad deal. Add in 15MB of cache, 12 threads and 28 PCI Express lanes, and you get a compelling compute solution. Shame about the TDP, a whopping 140W for a part that’s clocked at 3.3GHz.
Like most other AMD CPUs in this list, this is an old CPU which managed to remain relevant by combining decent performance with a superbly low price. Since this is an Athlon model, it doesn’t come with a GPU. It does however offer an unlocked clock multiplier which means that it is ideal if you want to overclock it.
It has all the necessary ingredients to be at the core of a great entry level computer. It is cheap, has four cores clocked at 3.6GHz (4GHz turbo) and, thanks to its 28nm manufacturing process, has a rather reasonable 95W TDP. Expect it to roughly match a Skylake-based Core i3-6100T CPUs albeit with a far higher TDP (three times) but almost half the price.
Gabe Carey has also contributed to this article