For the unconscious, IPC (instructions per cycle) provides a good indicator of how fast a processor is, and both a high IPC with a high operating frequency is the best combination for maximum performance. This is the case with Intel's 8th generation Coffee Lake CPUs, and although AMD lags far behind in terms of frequencies, the company seems to have really achieved Intel's IPC performance. This is probably why so many of you asked about this type of test.
To see how far AMD has progressed here, we will limit as many variables as possible while keeping things as realistic as possible. The first and most obvious step is to remove the core frequency from the equation. To do this, we locked all CPU cores at 4 GHz. Any type of boost technology has been disabled and the cores cannot exceed 4 GHz.
The second generation Ryzen CPUs were tested on the Asrock X470 Taichi Ultimate and the Coffee Lake CPUs on the Asrock Z370 Taichi. Both configurations used the same G.Skill FlareX DDR4-3200 memory with the memory profile & # 39; Xtreme & # 39; and the same MSI GTX 1080 Ti Gaming X Trio.
We can say in advance that this article doesn't buy advice in any way, but we only test on the science of it.
The Coffee Lake CPUs have a clear advantage in terms of clock speed. For actual performance information, see our latest Ryzen 5 2600, 2600X, and 2700X reviews.
For this test, we recorded results for the Intel Core i7-8700K, 8600K, Ryzen 7 2700X, 2600X and Ryzen 7 1800X together with the 1600X. Now 1600X, 2600X and 8700K all have the same CPU resources: 6 cores with 12 threads.
The 1800X and 2700X models have the advantage of being 8-core / 16-thread CPUs, while the 8600K is at a disadvantage because it is a 6-core / 6-thread CPU. So please keep this in mind when you continue. Let's get to the results.
Starting with the test of the permanent memory bandwidth, we see that the Ryzen CPUs of the first and second generation are similar with a bandwidth of around 39 GB / s. With exactly the same memory, the Coffee Lake CPUs are limited to around 33 GB / s. This corresponds to a bandwidth reduction of 15% compared to the Ryzen CPUs.
If we move on to Cinebench R15, we see that the 2600X is 4% higher than the 1600X for the multithread test and 3% higher for the single thread score. Then when we look at the 8700K, we see that it is 4% faster than the 2600X for the single-thread score, but 4% slower for the multi-thread score.
As you might have expected, the Ryzen CPUs with 8 cores and 16 threads clock-by-clock easily outperform the multithreaded score of the 8700K. I just took it in because I had the results. Depending on your needs, I could update this test with the Core i7-7820X, for example.
Next we have PCMark 10 video editing results and this is a test with lighter threads, although we previously noticed a notable difference between the 1600X and 1800X. As a result, we see a solid 10% jump from the 1600X to the 2600X, making AMD equivalent to Intel in terms of IPC performance for this test.
As with Cinebench R15, SMT appears more efficient than Intel's HT technology at maximum load. Here the 1600X was 3.5% faster than the 8700K, while the 2600X was 8% faster, and that's a remarkable margin right there.