With the incredible market demand around DDR5 memory and significantly elevated pricing on the more premium DDR5 memory modules, as part of the AMD Ryzen 7 9850X3D launch there’s been some communication that thanks to 2nd Gen AMD 3D V-Cache, using lower memory speeds like DDR5-4800 can be suitable without much of an impact to the gaming performance. But what about for Linux gaming? And other workloads with the Ryzen 7 9850X3D? Complementing yesterday’s Linux review of the Ryzen 7 9850X3D are benchmarks of DDR5-4800 vs. DDR5-6000 performance with Ubuntu Linux and this new 3D V-Cache 8-core / 16-thread desktop processor.
The AMD Ryzen 7 9850X3D Linux review was done with the GSKILL DDR5-6000 EXPO memory along with the other tested processors while today’s article is evaluating the performance of the 9850X3D across 300+ workloads when paired running at DDR5-4800 speeds.
AMD’s messaging with the Ryzen 7 9850X3D with DDR5-4800 is that for gaming there can be minimal performance impact compared to using DDR5-6000. They cited 2 x 16GB DDR5-4800 DIMMs costing around $400 to 2 x 16GB DDR5-6000 DIMMs for around ~$470 USD. The prices are very volatile still and so the testing in this article is focusing solely on the performance impact and thus the performance-per-dollar/value being an exercise for the reader depending upon your local memory pricing and whether you’d plan to re-use existing DDR5 memory in your possession or not.
These benchmarks are mainly done for reference purposes in seeing the performance difference of DDR5-4800 vs. DDR5-6000 not only for gaming but other workloads to help you decide in the cost/benefits of opting for DDR5-6000+ as opposed to potentially going for lower-cost DDR5-4800 DIMMs or re-using some existing DIMMs if you are considering an upgrade to the new AMD 3D V-Cache Zen 5 processors.
The testing with the AMD Ryzen 7 9850X3D on the ASRock X870E Taichi was done while running Ubuntu 25.10 with the Linux 6.17 kernel. Thanks to AMD for supplying the review hardware for making this independent benchmarking comparison possible.
