Your logic is incredibly flawed here. AMD is releasing these not only more then a year after the x3d parts came out, but after their SUCCESOR platofrm came out, AND the successor x3d parts...
The logic isn't flawed, it's obvious.
If products and updates are still being released for a platform, even if the products are just cut-down or modified versions of previous products, the platform is not dead.
It's possible for a CPU manufacturer to support multiple platforms at once (and therefore have multiple platforms that are not dead), and that's exactly what AMD is doing here with AM4 and AM5. Not every new CPU has to be a high-end model released for the newest platform, and a platform doesn't immediately become dead the moment a newer platform comes along.
AMD is treating AM4 as a low-end platform which will exist alongside AM5 for high-end processors, at least until releasing low-end Zen 4 desktop processors becomes more profitable than continuing to sell Zen 3 for AM4, like what they did a few years ago for AM3+ and FM2+, and what Intel did for LGA 1155 alongside LGA 1356. In both cases, 2 platforms were supported alongside each other for desktop processors at different price points.
By
your logic, you may as well say that AM5 is dead too, because sTR5 was recently announced.
Do you even read I said raster. Is Raster the only thing CPUs are responsible for in Games?
No... So why even bother to bring it up?
If a CPU isn't powerful enough to run the rest of the game's calculations that need to be performed every frame (this can include calculating resources, physics, etc) faster than the GPU can generate frames, it will bottleneck the GPU, regardless of what the raster threads are doing.
The point that fevgatos and trsttte are making is that in most modern games at 4K ultra settings, with a Ryzen 7 5800X3D and an RX 7900 XT, performance will be limited by the GPU, so upgrading the CPU to a 7900X3D will not result in a significant increase to fps. This doesn't apply to all games, but it is demonstrably true for most, and benchmarks easily prove this.
What was "feeds my GPU 3-5 more GB/s VRAM" even supposed to mean? The CPU isn't connected to the VRAM, so it doesn't directly affect your VRAM bandwidth in any way. The CPU communicates with the GPU through PCIe. In what situations are you seeing your GPU get more VRAM bandwidth with the 7900X3D than with a 5800X3D, and does it actually affect performance? A more powerful CPU could potentially increase the VRAM usage with reBAR enabled as it would be able to access and use data from the GPU more often, but describing this as "feeding the GPU more VRAM" is completely backwards - it's
using more bandwidth, not
providing more bandwidth.
