Updated | Rampant Ryzen 9 7900X further exposes UserBenchmark bias as AMD's 12-core Zen 4 chip rips up Raptor Lake rivals
Update September 16: Another Ryzen 9 7900X UserBenchmark run has been recorded, with this second test resulting in a huge bench of 143%. This is in contrast with the 135% recorded by the sample discussed below, and it now leaves the upcoming Zen 4 part on a stunning average of 139% based on two samples.
It’s best to make this point straight from the offset: UserBenchmark has gained considerable notoriety in the benchmarking community because of its clear and proven bias toward Intel parts. However, the controversial site is still a useful tool for spotting new and upcoming chips being tested, and now we have a first sighting (via APISAK) of the AMD Ryzen 9 7900X on the synthetic benchmark. The 12-core Zen 4 part has made an astonishing run that left it with an average bench score of 135%, although as the site fairly points out the lack of samples means “this result should be interpreted with caution”. The familiar withering criticism and attack on “Advanced Marketing Devices” is undoubtedly in preparation.
This result can assist with comparing the Ryzen 9 7900X with a variety of rivals and stablemates. For instance, the Ryzen 5 7600X currently tops the average bench chart on 117% while the Intel Core i9-1200KS is second on 115%. So, this particular run from the 12-core Zen 4 chip already places it an incredible 18% ahead of its fellow Ryzen 7000 processor. However, Zen 4’s main competition is not Intel’s 12th-gen Alder Lake chips but the upcoming 13th-gen Raptor Lake family. Interestingly, a couple of Raptor Lake engineering samples have been spotted on UserBenchmark, with one of them producing an also impressive average bench of 134%. But the outstanding results awarded to the 8-core and 16-core Raptor Lake chips (one potentially being an i7-13700 or i7-13700K) also show just how confounding the site’s scoring method is. A Core i7 nipping at the heels of a Ryzen 9 part looks better for Team Blue than for Team Red.
The Ryzen 9 7900X, tested with an ASRock X670E Pro RS motherboard, managed these individual section percentages for an overall average bench of 135%: normal (1-core and 2-core): 150%; heavy (4-core and 8-core): 185%; server (64-core): 223%. A look back at the best-performing Raptor Lake ES on 134% shows these marks: normal: 139%; heavy: 164%; server: 117%. It’s well known that UserBenchmark appears to not care so much about multi-core supremacy when producing its benchmark scores, but the AMD Ryzen 9 7900X is still way above the Raptor Lake sample in normal and heavy but only gains 1% in average bench score. There’s even more damning evidence to pour over that shows something is awry here.
The two Raptor Lake ES parts that we previously reported on had a 4% difference in average bench (134% vs 130%); however, the differences in normal and heavy scores were small while the lower-scoring 16-core part destroyed the 8-core part in the server score. The differences between the admittedly promising Raptor Lake chips were: 5% in normal, 5% in heavy, and 113% in server. The differences between the 134%-scoring Intel chip and the Ryzen 9 7900X were: 11% in normal, 21% in heavy, and 106% in server in favor of the AMD Zen 4 part. But somehow, the Ryzen 7000 processor is only +1% ahead of the Raptor Lake engineering sample (see table below). Despite UserBenchmark’s obvious attempts at hampering AMD’s Zen 4 offerings, it’s clear to see that the Ryzen 9 7900X may force a rethink for the much-vilified site.
|Ryzen 9 7900X||8-core Raptor Lake ES||16-core Raptor Lake ES|