AMD took to its blog today to provide clarity regarding the gaming performance of its Zen 5 architecture, particularly in light of the mixed reviews following the launch of the Ryzen 9000 series. The company acknowledged the discrepancies between its advertised performance metrics and those reported by reviewers, attributing much of the variance to differences in testing methodologies, hardware configurations, and software settings. AMD has reached out to reviewers to investigate these inconsistencies and has subsequently updated its internal benchmark suite, introducing a revised list of game titles alongside new performance projections.
AMD’s Revised Zen 5 Ryzen 9000 Performance Projections Against Intel
In its initial assessments, AMD claimed that the Ryzen 9000 processors were, on average, 6% faster than their Intel counterparts. However, after further testing, the company now states that the processors generally perform at parity with Intel chips when both are tested under optimized settings. Despite this adjustment, AMD has maintained its performance improvement projections for the Ryzen 9000 series over the previous generation, the Ryzen 7000 models.
AMD's original benchmark suite included a mix of esports, AAA, and older titles, which skewed towards games that were less CPU-intensive. To better reflect modern gaming performance, AMD has added several new titles to its benchmark suite. The company also noted that instability issues with Intel’s 13th- and 14th-Gen Core processors may have impacted initial comparisons, as many reviews utilized Intel’s conservative power settings during testing. With the release of a microcode patch from Intel, AMD has updated its performance data to reflect a more accurate comparison.
AMD’s Ryzen 9000 Over Ryzen 7000 Projections
AMD has made slight adjustments to its projections for performance gains with the Ryzen 9000 series over the Ryzen 7000 series. The initial claim of a 9% improvement in 1080p gaming has been revised to a more conservative estimate of 5-8%. This range indicates that performance gains may vary depending on the specific processors being compared. Reviewers have reported varying results, with some noting a 12% improvement while others observed minimal gains.
Admin Account Gives Varying Performance Boosts
Another factor influencing performance metrics is the use of an Admin account during testing. AMD's benchmarks were generated using an Admin account, which allows for optimizations that enhance performance by 2-3%. This feature, however, is not available in standard user accounts. AMD is collaborating with Microsoft to make this optimization available as an optional update for standard users.
The Impact of VBS
AMD also highlighted the impact of Virtualized-Based Security (VBS) on performance. While VBS is enabled by default in Windows to enhance security, it can reduce gaming performance by approximately 5%. The effects of VBS can differ based on the architecture of the processors being tested, potentially leading to inconsistent benchmark results if one processor handles virtualization better than another.
AMD Will Fix Its Chipset Drivers
While AMD's blog did not address ongoing issues with its chipset drivers, the company confirmed that a fix is in development. The current chipset driver employs a core parking technique that can inadvertently reduce performance when switching between different processors. AMD aims to resolve this issue to ensure consistent performance across various processor comparisons.