2 ms·

you're rather missing the point about the exclusion of 3D V-cache on the higher-end processors. If it was a productivity-focused play then you'd see it include

by paulmd 5y ago

you're rather missing the point about the exclusion of 3D V-cache on the higher-end processors. If it was a productivity-focused play then you'd see it included on those as well. The 5800X is a good processor and it's true that people do productivity on it, nobody disputed that and there's no need to rush to its defense. But you've missed the point about the overall way the upgrade is (or rather, isn't) being rolled out.

zozbot234 5y ago
Parent comment pointed out that the whole 3d-cache business might come with challenging thermals on high-core-count desktops. There might be all sorts of technical reasons why V-cache is not being used in this particular segment as of yet.
- paulmd 5y ago
  thermal problems are obvious based on AMD disabling overclocking on the 5800X3D. Even when gains have become minimal, they didn't explicitly disable them, so clearly either potential gains are literally zero, or overclocking might actually damage the cache/lead to conditions where it could be bricked or lead to glitches that might allow control of the PSP (the cache die being separate, and completely exposed, certainly is an interesting attack "surface"). Thermals have always been the giant asterisk on die stacking. Everyone knows it's going to be a problem. But this is the first consumer stacked-die product (that I am aware of?) so we don't really have any sense for overclocking on those processors. The tentative implication here is - bad. Even with a single cache die, it's still got a whole processor underneath it heating it up. There are implications here for GPUs as well, as both AMD and NVIDIA are expected to release stacked-die products next generation. It's looking like they will have to keep clocks under control to make that happen - maybe that is a counter-argument to the "500-600W" rumors (for both brands). But multi-die products have twice the area to dissipate their heat over - just like 5950X doesn't run at higher temperatures than 5800X. Higher power, yes, but twice the area to dissipate it over means thermals are about the same. That's not really the reason. A single-die limitation also wouldn't rule out a 3D version of the 5600X - it's no longer "the bottom of the stack" and finally there are value SKUs underneath it, it would be appropriate to re-release it as a more performance-oriented 5600X3D SKU. Anyway, my personal opinion is this is going to be a very limited product that primarily exists for "homologation". AMD can say that they've released it, it's officially a consumer product, so it can re-establish AMD's place on top of the gaming benchmark charts, but it's going to be a 9900KS-style super-limited SKU that doesn't see any real volume production at least until AMD has sold their fill of Milan-X. It just exists to put AMD back on the top of the benchmark charts now that Intel has retaken it with Alder Lake. The 5800X is the best processor for them to do that. 5950X sets up some small regressions for inter-CCD problems/etc. 5800X is enough for games for now anyway.

by paulmd 5y ago

zozbot234 5y ago

Parent comment pointed out that the whole 3d-cache business might come with challenging thermals on high-core-count desktops. There might be all sorts of technical reasons why V-cache is not being used in this particular segment as of yet.

paulmd 5y ago
thermal problems are obvious based on AMD disabling overclocking on the 5800X3D. Even when gains have become minimal, they didn't explicitly disable them, so clearly either potential gains are literally zero, or overclocking might actually damage the cache/lead to conditions where it could be bricked or lead to glitches that might allow control of the PSP (the cache die being separate, and completely exposed, certainly is an interesting attack "surface"). Thermals have always been the giant asterisk on die stacking. Everyone knows it's going to be a problem. But this is the first consumer stacked-die product (that I am aware of?) so we don't really have any sense for overclocking on those processors. The tentative implication here is - bad. Even with a single cache die, it's still got a whole processor underneath it heating it up. There are implications here for GPUs as well, as both AMD and NVIDIA are expected to release stacked-die products next generation. It's looking like they will have to keep clocks under control to make that happen - maybe that is a counter-argument to the "500-600W" rumors (for both brands). But multi-die products have twice the area to dissipate their heat over - just like 5950X doesn't run at higher temperatures than 5800X. Higher power, yes, but twice the area to dissipate it over means thermals are about the same. That's not really the reason. A single-die limitation also wouldn't rule out a 3D version of the 5600X - it's no longer "the bottom of the stack" and finally there are value SKUs underneath it, it would be appropriate to re-release it as a more performance-oriented 5600X3D SKU. Anyway, my personal opinion is this is going to be a very limited product that primarily exists for "homologation". AMD can say that they've released it, it's officially a consumer product, so it can re-establish AMD's place on top of the gaming benchmark charts, but it's going to be a 9900KS-style super-limited SKU that doesn't see any real volume production at least until AMD has sold their fill of Milan-X. It just exists to put AMD back on the top of the benchmark charts now that Intel has retaken it with Alder Lake. The 5800X is the best processor for them to do that. 5950X sets up some small regressions for inter-CCD problems/etc. 5800X is enough for games for now anyway.