Announcing AMD Smart Access Memory, it seems that the company has finally come up with a way to enable Rice CPUs and GPUs to work together to achieve the best performance possible. Our performance tests have proven to work well for Sam, but the future is not limited to AMD-AMD CPU / GPU configurations.
Thanks to PCWorld soon Interview, We have an answer. According to AMD, there are people in the Ryzen team working to get Sam to work on Nvidia GPUs, and in the Radien team there are people working with Intel to work with Intel CPUs and chips. If AMD is comfortable with such an advertisement, it means that there is mutual support in these events, which means that no matter what we hear about the Nvidia / Intel collaboration, we will see the support of the platform, the sellers. It only makes sense for the two companies to work together, as Amhara will benefit AMD’s free operations.
Although AMD does the same thing as anyone else, this proves that Sam is not like AMD-based technology. Adjustable BAR capability (that is for SAM PCA 2.0), but apparently no GPU vendor supports it yet.
If this were AMD-based technology, one might suspect that the company should have designed to use Zen 3 and / or RDNA2. The support was given to Intel and Nvidia hardware because the behavior was not considered appropriate for the problem or because the companies in question did not know that it could actually increase performance until one tried it properly. The latter is rather depressing.
According to AMD, there is some work that needs to be done to properly support the feature, as it has not been activated immediately on Intel and Nvidia platforms. It will be interesting to see what performance we will see when other platforms and hardware take up this capacity – Intel may benefit more than AMD (or vice versa), and AMD GPUs may benefit more from Nvidia cards or vice versa.