In a surprising bit of news, it turns out AMD may have postponed or straight-up canceled the Zen 3+ lineup, codenamed Warhol in favor of a slightly earlier Zen 4 launch (sometime in October or November 2022). Keep in mind that this, like every other piece of news about Warhol, is still a rumor, and the existence of the Zen 3+ processors hasn’t been confirmed by AMD at any point.
The other day, a certain YouTuber shared in a video that Warhol/Zen 3+ would feature a higher IPC gain than originally anticipated, despite featuring the same core architecture as Vermeer. Even if the Zen 3 refresh isn’t canceled, it’s highly unlikely that we’ll see an IPC increase of more than 2-3%. The overall performance may, however, increase due to the use of a more mature derivative of the N7 (7nm) process, but that’s about it.
As it stands right now, this means that Intel’s 10nm Alder Lake-S processors will go up against the Ryzen 5000 chips, and almost certainly should beat them by a notable margin. We are, after all, looking at a much denser node (10nm SFE), a new core architecture (Golden Cove) with an IPC improvement of 25-25% as well as a new memory interface (DDR5), and a brand new platform (LGA1700).
AMD might be able to hold the lead in content creation workloads thanks to the Ryzen 9 processors which feature up to 16 cores, but the gaming crown will certainly fall back to Intel, even if for a brief interval. The hybrid core architecture raises a few concerns, primarily with respect to core prioritization, especially in mobile processors. (gabapentin) It’ll fall to the operating system to distinguish between low-power and high-performance cores and use the right ones for the appropriate workloads.
Either way, as far as AMD goes, we’ll likely see Warhol in early 2022 (if not late 2021) in the form of the Ryzen 5000 XT lineup, or perhaps, not at all. The company has yet to release its hex and octa-core Zen 3 parts, as well the quad-core offerings, so we could see those in Warhol’s stead. Again, all this is pure speculation and should be taken as such.