AMD is often the top choice when you’re looking for value in a processor, but soon, it may take the crown of top performance from Intel—at least in the short term. Consider AMD when building your next PC.
AMD made a big splash this spring with the introduction of its Ryzen 3000 desktop CPUs and the accompanying X570 chipset. This duo starts shipping July 7, 2019, with promises of zippy PCIe 4.0 transfer rates, and a killer value proposition in terms of cost, core count, and power usage.
Value has always been AMD’s advantage over Intel, with its Zen, Zen+, and now Zen 2 architecture. We won’t know for sure how well the new Ryzen 3000 processors perform until independent benchmarks and tests appear. Nevertheless, it sure looks like Ryzen 3000 will be impressive.
Intel, meanwhile, isn’t making a move on new desktop processors any time soon (with, perhaps, one exception), strengthening the rather convincing argument to consider AMD for your next desktop build.
AMD vs. Intel: The Struggle Is Real
AMD owned the Computex trade show in May when the company introduced its Ryzen 3000 desktop processors, which are based on the Zen 2 architecture and the new X570 motherboard chipset. The new CPUs use a 7nm (nanometer) process, with a wide range of core and thread counts at lower heat generation (TDP) and, presumably, lower power usage than previous models.
At E3, AMD followed up its Computex triumph by introducing yet another Ryzen 3000 processor, the 16-core Ryzen 9 3950X. Before the Ryzen 3000, you would only find 16-core chips at the enthusiast level, requiring high-end motherboards at a high-end price.
AMD’s 16-core chip, by comparison, has a sticker price of $749. That’s still expensive, but Intel’s 16-core chip (the Core i9-9960X) is more than double that price. Perhaps that’s not quite a fair comparison, as the Intel chip is overkill for most people. It supports a whopping 44 PCIe 3.0 lanes compared to 24 PCIe 4.0 lanes in the new AMD chip, and Intel’s CPU can handle a boatload of memory.
Then again, that’s the point. The AMD 16-core chip is a mainstream CPU that fits into mainstream boards. That’s something Intel doesn’t have. If Intel intends to provide a more affordable response to Ryzen 3000, we won’t see it for a while. The next generation of Intel CPUs, called Ice Lake, are headed to notebooks around the end of the year, but there’s been no word on when the next round of desktop CPUs will appear.
AMD’s Value Proposition
AMD’s new processors are offering a lot of value over its previous generation parts and Intel’s current desktop processors. Let’s take a simple example with the $329 Ryzen 7 3700X and its predecessor, the Ryzen 7 2700X, currently selling for about $280. The newer processor has the same core and thread count as the older version, and it offers around the same clock speeds. But the newer CPU has a bigger total cache at 36MB, compared to around 21MB for the 2700X. This suggests the 3700X will be better with heavy workloads, such as video processing. A CPU’s cache is like its onboard memory. It lets the processor access instructions faster than fetching it from the system’s memory.
AMD also rates the 3700X at a TDP of 65W, compared to 105W for the 2700X. That means the newer CPU generates less heat and should consume less power, as well—not a bad upgrade for a $50 price hike.
The same goes for the cheaper Ryzen 5 3600 and its cousin, the Ryzen 5 2600. Here, we have the same core and thread count (6 and 12), but the 3600 is slightly faster, has a bigger cache, and supports the faster PCIe 4.0. If you find a good sale, you can get the Ryzen 5 2600 for about $145-$150, while the Ryzen 5 3600 has an MSRP of $200. Again, it’s a nice little bump in specs for about $50 more.
Okay, fine. Big surprise that AMD’s newer chips are better than its old ones. What about value versus Intel?
Let’s compare the 3700X to Intel’s popular Core i9-9900K. Both processors have eight cores and 16 threads, and both have the same base clock of 3.6GHz. Intel’s boost on the 9900K is much better at 5.0 GHz versus 4.4 GHz on the 3700X. AMD’s CPU has a 36MB cache over the 9900K’s 16MB. The 3700X also has a lower TDP at 65W, versus 95W for the 9900K. Presumably, that means the 3700X is sucking up less power, but given that TDP isn’t a standardized measurement, we’ll only know how close they are when we see some real-world testing.
The real kicker here is pricing. The AMD 3700X, with its MSRP of $329, is just scads cheaper than Intel’s $485-$490 Core i9-9900K. Given Intel’s boost clock and the 9900K’s favorability as a top gaming CPU, the 3700X probably won’t beat the 9900K in performance. Just how shy it will be of the 9900K isn’t yet clear. However, even going a step up to the eight-core, 16-thread Ryzen 7 3800X—which reportedly did beat out the 9900K in early (and anonymous) benchmark leaks—you’re still saving about $85 over Intel. That may not seem like much, but when you start adding up costs for a new PC, that lower price starts to matter.
These CPUs don’t have integrated graphics, like Intel’s do. But, if you’re looking for serious performance, you’re going to get a discrete GPU for your desktop PC anyway.
Zen 2’s Value Caveat
We’ve established that these Zen 2 processors sound great when it comes to value, but there’s one big caveat. If you want these Ryzen 3000 chips to support PCIe 4.0, you need to buy an X570 motherboard.
These motherboards are expected to be quite expensive for a few reasons. They have a more expensive chipset, are built on higher quality PCB, and require some serious cooling design, with fans, heat sinks, and so on.
That may put a damper on the bargain pricing of these new Ryzen CPUs for now. In early 2020, it may be a different story if the newer version of the traditionally cheaper Ryzen motherboards (expected to be called B550) roll out. For now, a fresh motherboard to go with that new Ryzen CPU is going to cost you.
The alternative, then, is to use a Ryzen 3000 CPU with a cheaper X470 board. You’ll still get the processor performance, but it means losing PCIe 4.0 for PCIe 3.0.
PCIe 4.0: A Big Leap, Too Soon?
Check out our article on why PCIe 4.0 matters to understand the advantages of the new standard in detail. In short, PCIe 4.0 is twice as fast as PCIe 3.0. For gaming, that doesn’t really matter right now, as PCIe 3.0 offers more than enough bandwidth.
The big advantage for PCIe 4.0 in these early days is that it promises to make NVMe drives quite a bit faster. PCIe 4.0 NVMe drives promise read speeds close to 5,000 Megabytes per second, while top NVMe drives right now hit around 3,500 MBps.
Unless NVMe speeds really matter to you, the value of PCIe 4.0 probably isn’t worth it in terms of value at the moment. Again, with the new AMD CPUs, we’d recommend looking for a well-priced X470 board to house it and wait for PCIe 4.0 to become more important than it is now before shelling out for an X570 board.
Intel’s Interim Response?
Overall, AMD is looking strong for the near future. Intel does have one desktop CPU up its sleeve, but it might not be much different from what we’ve already seen.
Intel introduced the Core i9-9900KS in late May during Computex. This processor will have a base clock of 4.0 GHz—up from the 9900K’s 3.6 GHz—and the same boost of 5.0 GHz. The difference, however, is that Intel says the boost on the 9900KS will affect every core. In other words, all eight cores will boost to 5.0 GHz, whereas on Intel’s other CPUs, the boost typically only affects a single core, with the others operating below that higher clock.
Poking around reviews and forum comments about the 9900K, you’ll find that when Intel’s Multi-Core Enhancement (MCE) feature is active, it often boosts all cores to 5GHz.
If the 9900KS surprises the world with a jaw-dropping performance over the 9900K—and its price isn’t out of reach for most people—then Intel might have a convincing final statement for 2019. If not, the immediate future looks like it’s all about AMD.