first 3D V Cache benchmarks appear.
And interestingly, the benchmarks show that for applications that don't fill up the L3 cache, actually perform Worse on the new models. This makes sense, we know the new v cache chips run slower than the old ones, so if a piece of software is not utilizing the extra L3 cache available then all you have is a CPU with useless cache and slower clock speeds for your efforts.
When all the L3 cache is filled, overall memory latency is reduced below the old models and this benefits gaming - this is why some games will see a nice performance gain on the 5800x3d and some will see no difference or even a reduction in performance. To get any of the benefit of the new v cache requires the game/application to fill the L3 v cache, otherwise bandwidth and latency is no different to the old models and because the boost clocks are lower the game/application performs the same or slower.
This means reviews for the 5800x3d are going to be very interesting, because you will able to cherry picks benchmarks to show the 5800x3d running slower than the 5800x in some games and cherry pick benchmarks to show the 5800x3d running significantly faster than the 5800x in other games.
https://www.tomshardware.com/news/amd-3d-v-cache-benchmarks-mixed-results-milan-x-cpus
). Beyond rumours we don't know precisely when.