AMD Ryzen 5 1600X
Taking AMD’s new hexa-core hotness for a spin
WHEN AMD PULLED the wraps off its radical new Ryzen CPU, it was all about those eight awesome cores. Count ’em, Intel. The first Ryzen 7 chips blew the market for $300plus CPUs wide open. But not everybody can unload that kind of money on a processor, especially a brand new design that also requires a motherboard upgrade. Enter the Ryzen 5 1600X. On paper, it might just be the ultimate balancing act between performance bang and efficient deployment of your hard-earned buck.
Try this for a compelling proposition. The 1600X packs six cores, and supports 12 software threads, while a similarlypriced Intel processor, the Core i5-7600K, has four cores and just four threads. And as a consequence of being the top sixcore model, it also clocks in at 3.6GHz nominal and 4.0GHz turbo. In other words, it looks like a very nicely balanced chip. Plenty of cores and threads for powerful parallelism, plus high clocks to provide solid single-threaded performance. And all for $249. It’s the mid-range CPU that we’ve all been waiting for, right?
For the most part, that’s an affirmative. When it comes to outright multithreaded performance, it pops a cap in the head of the Intel Core i5. Take Cinebench: The Core i5-7600K manages 663 points. The Ryzen 1600X? A massive 1,223 points, or nearly double the Core i5’s capability. As youknow-who would say, that’s ’uge.
Admittedly, the Core i5 retains some dignity in the single-threaded version of Cinebench, with 179 points compared to the 1600X’s 159 points. However, the AMD chip looks like the better trade-off at first glance. Elsewhere, if the results aren’t quite so spectacular, AMD still chalks up some decent wins. It crunches highquality video encoding at 22fps to the i5’s 16fps, and it motors through the Fry Render benchmark in just 3 minutes and 46 seconds—that’s fully two minutes faster than the 7600K.
PARALLEL LINES Put simply, if you want a CPU for content creation, or really any workload that majors on parallelism, this isn’t even a race. The Intel competition is hideously, hopelessly outclassed. That’s true even taking into account the fact that the Ryzen CPU is a little more power-hungry. An extra 10 watts at the wall is surely worth it.
Where the Ryzen proposition is a little less compelling is in the games arena. It looks competitive in our benchmark numbers—when it comes to average frame rates, the 1600X delivers on that 4GHz promise. The subjective experience, however, tells a slightly different story. TotalWar:Attila is a great case study here. Running on the Intel processor, it’s super-smooth, whether you’re up in the sky, lording it over your troops, or down in the thick of the action. But with the Ryzen CPU at the same settings and using the same video card, there’s noticeable judder almost everywhere. It’s not nearly as smooth, and the difference is far greater than the modest average frame rate gap.
As for the reasons why, there are numerous candidates. For starters, we used an Nvidia GPU, and there are at least indications that the Nvidia driver is poorly optimized for the Ryzen architecture. There are also question marks surrounding Ryzen’s basic architecture, which is composed of a pair of quad-core modules. Long story short, some latency is involved in communicating between the two quadcore modules, and that can require careful management by both OS and application to avoid performance penalties. Ryzen is brand new, so most software has yet to be tweaked to take account of such issues.
Of course, most software runs OK without any modifications. It’s just a pity that the main class of application that seems to be particularly sensitive to Ryzen’s architectural nuances just so happens to be gaming. On a final note, like other Ryzen processors, the 1600X has pretty much zero overclocking headroom. AMD has these things running pretty ragged. An additional 100MHz was the best we could achieve.