Intel’s chief performance strategist, Ryan Shrout has posted a very interesting article on his medium account. He tested the Intel Core i9-9900K at 1080p in four, six and eight-core modes in an attempt to analyze core-scaling for PC games. Based on the results he claims a number of things. Perhaps the most interesting claims are that adding extra cores doesn’t improve performance and “8-cores is the optimal spot for performance scaling in modern PC gaming”. Posting this just before AMD releases its first 16-Core gaming CPU in the Ryzen 9 3950X seems like a sorry attempt to disparage the competition.
“adding core-count just because you can… doesn’t result in better performance.”Ryan Shrout, Chief Performance Strategeist, Intel
Using an Intel i9-9900K, Shrout tested it in 4, 6 and 8-core configurations, locking it to 4GHz. Among the six games tested, there is a 20-30% performance increase in frames at 1080p. The switch to 8 cores, however, doesn’t show much improvement at all. He also tested the i9–9980XE to similar results. To say that the testing is inconclusive would be an understatement. Coming to a conclusion without even bothering to test the Ryzen 3000 processors is quite an oversight.
Moreover, Intel has been marketing their flagship processors primarily for streaming. So, Shrout’s argument that cores don’t improve performance holds even lesser weight with regards to his own Company’s offerings. Not to mention that with the Ryzen processors being significantly cheaper, they outperform Intel’s offerings in price-to-performance alone.
“For gaming, 8-cores is the optimal spot for performance scaling in modern PC gaming,” says Shrout, “and at 8-cores we get the best frequency scaling out of our Coffee Lake products and 14nm process technology today. And clock speed is what feeds the hungry primary threads of game engines today!”Ryan Shrout, Chief Performance Strategeist, Intel
Intel’s latest flagship, the 9900KS runs at 5GHz on all cores. As per Shrout’s results, this would mean that the KS grossly outperforms every other performance on the market, right? Tests and benchmarks say otherwise. In fact, there’s barely any difference. It is technically the fastest gaming processor around, that doesn’t really matter in gaming or other tasks. While he does make a valid point about frequency and IPC being important factors in today’s games, the Comet-Lake processors are proof that iterating on the same architecture over a period of 4 years and increase in subsequent clock speeds doesn’t translate to better performance.
With the gaming industry pushing 4K and higher resolutions, the impact of CPU is slowly diminishing. Mostly gaming performance is GPU-intensive. You can opt for a cheap CPU with a decent GPU, and there won’t be a significant difference in most games. Moreover, with the testing being limited to 1080p, there’s barely any conclusions to be made and it’s disappointing to see Intel’s Chief Performance Strategist trying to downplay the competition in such a silly way. Maybe if Intel had been more focused on new technologies, it wouldn’t have to resort to such methods.