Benjamin Carr, Ph.D. 👨🏻💻🧬<p>Examining <a href="https://hachyderm.io/tags/RDNA4" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RDNA4</span></a>'s out-of-order memory accesses in detail, and investigating with testing<br><a href="https://hachyderm.io/tags/RDNA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RDNA</span></a> 4's memory subsystem enhancements are exciting and improve performance across a variety of workloads. <a href="https://hachyderm.io/tags/AMD" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AMD</span></a> specifically calls out benefits in <a href="https://hachyderm.io/tags/raytracing" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>raytracing</span></a>. But RDNA 4's scheme for handling memory dependencies isn't fundamentally different from that of <a href="https://hachyderm.io/tags/GCN" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GCN</span></a> many years ago. RDNA 4’s makes the most significant change to AMD’s <a href="https://hachyderm.io/tags/GPU" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GPU</span></a> memory subsystem since RDNA launched in 2019. <br><a href="https://chipsandcheese.com/p/rdna-4s-out-of-order-memory-accesses" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">chipsandcheese.com/p/rdna-4s-o</span><span class="invisible">ut-of-order-memory-accesses</span></a></p>