Why would the improvement vary significantly from CPU to CPU? The only thing I can think of is if the 7900X3D and 7950X3D both have two 3D cache dies, games that are heavy on 2 threads can put one on each CCD, both using their own separate 96 MB of cache.
But when APU? I would like to see a minipc running 6900HX with 6400Mhz @75W TDP like saw on ETA Prime YT channel.
Can't use unified memory as HBM2e/3 up to 8GB instead RAM and 3D cache like for 8809Gh? Like for servers cpu? Get same speed of traditional cache and produce more perf than DDR5 ram increasing latency and more steps.
4 Vega igpu generation and now directly to RDNA3? This is a joke!
Phoenix APU with RDNA3 12CU @6400Mhz when can reduce costs using Rembrandt APU 6700G up to 16CU with 3D cache or HBM2 (old) up to 8GB moving to trash DDR5.
Ryzen 7000 CPUs with 3D cache are on their way. Rumors have it the processors will launch early next year with gaming performance that eclipses the Ryzen 7000 non-3D chips. So how much difference will the Ryzen 7000X3D SKUs make? Per RedGamingTech, the Zen 4 X3D SKUs can be up to 30% quicker than their non-3D counterparts.