There has been a contentious discussion surrounding memory speeds and AMD CPU buses on the AM5 platform. It appears that the widely promoted DDR5-4400 at a 1:1 ratio is not the “magic formula” for Ryzen 9000 CPUs as suggested by some sources, particularly from China. Performance data now indicates that this configuration actually results in fewer frames per second (FPS) compared to faster RAM setups. Could this be another fabricated tale?
Tests conducted with the Ryzen 7 9800X3D revealed that while the DDR5-4400 1:1 configuration performs adequately, especially with its 3D V-Cache, it lags behind in average FPS when compared to DDR5-5600 JEDEC, DDR5-6200 EXPO, DDR5-6400 1:1, and DDR5-8200 2:1 in the tested games.
Ryzen 9000’s DDR5-4400 1:1: Just Another Tall Tale
Substantial data on this topic has emerged from Igor’s Lab reviews, providing a precise overview of the situation. The core idea behind the DDR5-4400 1:1 configuration was that it would maintain a cleaner synchronization between the memory and the memory controller, thus avoiding latency penalties. On paper, this sounds appealing, especially for the AM5 platform, where the interplay between memory, controller, and Infinity Fabric has always been significant. However, theory often diverges from real-world performance.
In Cyberpunk 2077 (version 2.3), running at 1440p with RT Overdrive and an RTX 4090, the Ryzen 7 9800X3D with DDR5-4400 1:1 achieved an average of 105.8 FPS and 82.1 FPS in 1% lows. In comparison, DDR5-5600 JEDEC yielded 107.1 FPS and 83.3 FPS (1% low). DDR5-6200 EXPO reached 108.3 FPS and 86.6 FPS (1% low). DDR5-6400 1:1 scored 109.1 FPS and 86.8 FPS (1% low), while DDR5-8200 2:1 delivered 107.4 FPS and 84.9 FPS (1% low).
This clearly indicates that the DDR5-4400 1:1 configuration is the slowest among the tested options, both in terms of average FPS and 1% lows. A similar trend was observed in another well-optimized game, Shadow of the Tomb Raider.
Here, DDR5-4400 1:1 managed 332.8 FPS average and 235.2 FPS in 1% lows. DDR5-5600 JEDEC achieved 331.5 FPS average but dropped to 224.7 FPS in 1% lows. However, faster configurations once again took the lead: DDR5-6200 EXPO recorded 333.7 FPS and 242.1 FPS (1% low), DDR5-6400 1:1 reached 334.0 FPS and 241.1 FPS, and DDR5-8200 2:1 offered 334.0 FPS average with 240.5 FPS (1% low). While DDR5-4400 wasn’t drastically underperforming, it was not the superior choice.
Theory Dismissed: Practice Shows Faster is Still Better, Though 3D V-Cache Mitigates Differences
In Assetto Corsa Competizione, the advantage of faster memory and more aggressive configurations becomes apparent. The DDR5-4400 1:1 configuration delivered 271.0 FPS average and 191.6 FPS in 1% lows. DDR5-5600 JEDEC saw a slight dip in 1% lows (189.9 FPS) but achieved 268.6 FPS average. DDR5-6200 EXPO reached 271.7 FPS and 191.1 FPS (1% low), while DDR5-6400 1:1 led with 273.9 FPS and 192.7 FPS (1% low). DDR5-8200 2:1 followed closely with 272.4 FPS and 192.3 FPS (1% low).
Therefore, China’s assertion that DDR5-4400 1:1 is the ideal gaming configuration for Ryzen 9000 CPUs is incorrect. While it may offer technical advantages in synchronization and provide decent frame times without being a complete performance disaster, the benchmarks do not support its claim as the best option. In these tests, the most balanced configurations appear to be around DDR5-6200 and DDR5-6400, as they provide better average and minimum FPS without relying on extreme DDR5 speeds.
The 3D V-Cache technology in the Ryzen 7 9800X3D makes it less sensitive to RAM scaling. However, even a small performance increase from faster RAM means the slowest option is not automatically the most recommended. These comparisons would likely show even more significant differences on CPUs without 3D V-Cache, which Igor’s Lab may explore next. It has been demonstrated that DDR5-4400 1:1 is far from the optimal choice for Ryzen 9000. Nevertheless, it remains competitive and, considering the current RAM market prices, it’s not an entirely dismissible option for budget-conscious users.
