AMD explains Smart Access Memory and radiator design of the Radeon RX 6000

Yesterday, AMD unveiled the long-awaited new generation of graphics cards and architecture in the form of the Radeon RX 6000 and RDNA 2. In addition to the promised competitive performance and ample memory in all models, the company also introduced new features, such as Smart Access Memory. In a new video, AMD’s Frank Azor explains how the technology works, and gives an insight into the cards’ cooling solution.

Smart Access Memory is part of what the company calls “the AMD Advantage”, which also includes Smartshift technology. Smartshift was unveiled during CES this year and allows laptops to dynamically switch power budgets between processor and graphics part depending on where performance is most needed. The technology is also used in the game console Playstation 5, where clock frequencies for processor and graphics part can vary with the load.

With Smart Access Memory, AMD develops the benefits of using the company’s components, as Ryzen 5000 processors get the full utilization of memory on Radeon RX 6000 graphics cards. It has previously been possible for the processor to access the graphics memory in 256 MB sections. Frank Azor points out that this was not a problem a few years ago when the usual amount of memory measured 2 to 4 GB, but that it limits game developers today.

With the Radeon RX 6000, developers can instead access the graphics card’s full memory. The fact that the processor does not have to reach and leave smaller sections of the graphics memory also means that it does not have to reach the primary memory as often, which reduces latency and improves overall performance. The company’s performance example combines a Ryzen 5000 processor with the Radeon RX 6900 XT, where the gain with Smart Access Memory goes from 5 to 11 percent.

AMD announces that Smart Access Memory works in the background and requires a Ryzen 5000 processor, a Radeon RX 6000 series graphics card and a motherboard with 500 series control circuits. The review also highlights the company’s existing technologies Radeon Anti-Lag and Radeon Boost, which lower latencies and increase performance without requiring specific interfaces or hand-laying from game developers.

Next, the coolers used in the Radeon RX 6000 series are discussed. Here, the top model RX 6900 XT is compared with Nvidia’s top card Geforce RTX 3090, where the latter has a larger power level of 350 watts against the RX 6900 XT at 300 watts. AMD also points out that the RX 6800 XT is less energy-intensive at 300 watts versus 320 watts with rival Geforce RTX 3080.

The comparison also extends to the physical size, where the RTX 3090 occupies three slots while the RX 6900 XT stays within 2.5 slots. AMD also emphasizes that both the RX 6900 XT and the RX 6800 XT stay within the 10.5-inch length (27 cm) that has long been a yardstick in component terms. This means that the cards fit in more existing chassis of a more compact model compared to Nvidia’s larger counterparts.

In terms of the coolers’ physical construction, a design with three axial fans is applied as previously uncovered. The cards support a silent mode where the fans stop completely at low load, and at load the sound level is stated to stay at 37 dBA. The company also highlights that the power connectors are located at the lower end of the cards with the connectors facing the chassis side panel, which should facilitate wiring.

Read more about AMD Radeon RX 6000: