It looks like someone over at UserBenchmark is being “sponsored” by Intel. The popular benchmark platform recently updated their CPU benchmark score weightages. Single-core performance now has a much higher impact on the score than earlier. Naturally, this puts AMD’s Ryzen chips at a disadvantage.
At launch, the first-gen Ryzen parts were quite a bit behind their Kaby Lake/Skylake counterparts in terms of IPC and raw single-threaded performance. This was offset by the fact that Ryzen offered more cores and more threads versus the comparable Intel Core part at every pricepoint. With Ryzen 3000 or Zen 2, cumulative increases to IPC and higher clocks courtesy of the 7nm process mean that Ryzen 3000 parts are now at least somewhat competitive in single-core performance.
When the multi-core performance was more important in Userbenchmark, Ryzen parts ruled the roost because they offered comparable single and quad-core performance to their Intel rivals, while being significantly better in the multi-core segemnt.
In the latest update to the Userbenchmark algorithm, multicore performance’s weightage has been brought down to just 2 percent, as spotted by this Redditor. This is especially problematic because over the years, game developing has been steadily shifting from single-threaded designs to a greater emphasis on multi-core scaling. There seems to be quite a bit of backlash on the interwebs to this update. We’ll see what, if anything, Usermark does about this.