Originally Posted by epic1337
what are you guys implying? my point is its time to change architectures, hint *netburst repeat*.
It doesn't matter what or how good the architecture is. There will always be a market for the best of the best. So whatever is available, the HFTs will go after whatever offers the best performance for them - which happens to be the highest-clocked of something
- even if it is so far into the realm of diminishing returns that it sounds stupid to those unfamiliar with the industry.
I can't imagine Intel putting any real effort into "making" these 9990XE chips. They're simply taking the best of what the binning process yields and selling them for a premium.
So it's not taking any resources away from their "next architecture".
Originally Posted by AlphaC
I'm guessing the AVX 512 performance isn't there and AVX2 requires a bigger clock drop than on Intel.
It's already public knowledge that Zen 2 won't have AVX512. (AMD pretends it doesn't exist, and it isn't in their GCC patches.)
On the AVX2 throttling, Anandtech did an interview with AMD. They said it won't throttle, but gave a very unsatisfying reason.
IC: With the FP units now capable of doing 256-bit on their own, is there a frequency drop when 256-bit code is run, similar to when Intel runs AVX2?
MP: No, we don’t anticipate any frequency decrease. We leveraged 7nm. One of the things that 7nm enables us is scale in terms of cores and FP execution. It is a true doubling because we didn’t only double the pipeline with, but we also doubled the load-store and the data pipe into it.