Closing Remarks

Despite a rather large clock speed deficit, the second-gen Ryzen CPUs aren't often that far behind their Intel rivals in application benchmarks and here we see why when comparing them clock-for-clock at 4GHz. In applications such as Cinebench R15 we see that the single core performance is down just 3% but where SMT is well leveraged AMD can be up to 4% faster.

We found that AMD was 3% slower in the Corona benchmark but much the same for our Excel, V-Ray and video editing tests. Then while it was 15% slower in HandBrake it was also 8% faster for the PCMark 10 gaming physics test. Of course, there is still the matter of gaming and I bet a few AMD fans were hoping we could put most of the gaming performance deficit down to clock speed. Sadly though, that's not the case.

One issue here is how AMD connects its cores, or rather its CCX modules. Intel's Ring Bus is very low latency and always sees resources accessed via the shortest path. However, as you add more cores the rings grow in size and more rings are required to connect all the cores and this sees efficiency go out the window. There's more to it, but basically Intel needed a better method for connecting large amounts of cores, like 28 of them for example. At this core count, the Mesh interconnect architecture is superior.

However, for 6, 8 and 10-core CPUs we already know it to be an inferior solution and this is why the Core i7-7800X, 7820X and 7900X all get smoked by the 8700K when it comes to gaming. The 8700K has an average latency of about 40ns between cores whereas the 7800X is more like 70 to 80ns.

The Ryzen CPUs are a little more complex as cores within a CCX module have a similar core to core latency to that of the 8700K and this is regardless of the DDR4 memory speed. However once you exit the CCX, core to core latency increases to around 110ns and that's with DDR4-3200 memory. The CCX to CCX core latency is reduced with faster memory as AMD's Infinity Fabric is tied to the memory clock rate and lower latency DRAM also helps immensely.

Another issue is the games themselves since almost all gaming titles are designed to work on just a few cores and we're only just starting to see some effort made toward breaking tasks up into pieces so they can be run in parallel. Prior to Ryzen's release, games were designed and almost exclusively optimized for Intel CPUs. That's slowly starting to change now, so Ryzen's gaming performance will improve, just don't expect it to be wiping the floor with Ring Bus CPUs anytime soon.

Shopping Shortcuts:

In terms of IPC performance, AMD has certainly closed the gap. The improved cache latency has also really helped and there are several benefits to buying a second-gen Ryzen CPU over a Coffee Lake CPU, so it's going to be exciting to watch the battle unfold in 2018 and beyond.