Recent chip design leaks and industry sources indicate that AMD's "Strix Halo" series of mobile APUs, including the Ryzen AI Max+ Pro 395, will incorporate a TSV (Through-Silicon Via) structure to support 3D V-Cache technology.
This design choice not only paves the way for future X3D variants of these processors but also suggests a significant potential for CPU performance gains through the addition of extra L3 cache. Such an enhancement is expected to be particularly beneficial in gaming and professional workloads, where larger caches can dramatically improve processing efficiency.
This move aligns with AMD's strategy of utilizing 3D V-Cache to enhance performance in cache-sensitive applications. By stacking additional L3 cache directly on top of the CPU cores, AMD can provide a substantial amount of super-fast cache memory. This approach has proven effective in other AMD processors, leading to faster game loading times, smoother frame rates, and improved overall system responsiveness.
The presence of TSV connections in the "Strix Halo" die, as confirmed by industry insiders, strongly suggests that AMD is indeed preparing to bring the benefits of 3D V-Cache to its mobile APU lineup. This development is generating excitement among enthusiasts who anticipate further performance leaps in upcoming AMD mobile processors, especially in demanding tasks like gaming and content creation.
Detailed analysis of the Strix Halo chip layout reveals that AMD has strategically reserved space above the L3 cache layer between the Zen 5 cores specifically for 3D V-Cache stacking. This vertical connection is facilitated by TSV (Through-Silicon Via) technology.
Mirroring the design approach of desktop Ryzen X3D processors, this allows AMD to significantly expand the cache capacity without increasing the overall silicon footprint of the chip.
Furthermore, the Strix Halo APU incorporates a brand-new interconnect solution. Compared to traditional SERDES (Serializer/Deserializer) designs, this innovative interconnect achieves a remarkable 42.3% reduction in physical size. This contributes to an overall chip size reduction of 0.34mm while simultaneously lowering latency and power consumption. These advancements not only benefit the Strix Halo generation but also pave the way for the future Zen 6 architecture and its continued pursuit of efficiency and performance.
0 Comments