If AMD’s Big Zen 3 Claim Is True, Intel’s In Trouble

If AMD’s Big Zen 3 Claim Is True, Intel’s In Trouble
Screenshot: AMD

This morning AMD revealed all the juicy details about its upcoming Ryzen 5000-series processors. While I don’t need to upgrade my CPU, or even my motherboard, the leap from AMD’s Zen 2 to Zen 3 architecture appears to be so good that I just might do it anyway. Higher power efficiency, more performance per watt, more instructions per clock, more frames per second, and faster single core performance than Intel?!

OK, let me sit down and take a breath.

Ever since AMD first introduced its Ryzen processors based on its home-grown Zen architecture, it’s consistently made generational improvements year after year. The Ryzen 3000-series was, and still is, fantastic. Ryzen 2000-series, too. AMD sort of muddied the water a bit with slightly juiced-up versions of its Ryzen 3000 processors (and when I say slightly, I mean there’s no point in rushing out to buy one even if they are really good). But now AMD has gone back to keeping it simple. There are four new Ryzen processors coming next month. Just four. That’s it. But that’s what I like about AMD. It doesn’t need a list of CPU SKUs as long as CVS receipt.

Here’s what’s hitting retail shelves on Nov. 5:

  • Ryzen 9 5950X: 16-cores/32-threads, 3.4 GHz base clock (4.9 GHz boost), 105W TDP ($US800 ($1,118))
  • Ryzen 9 5900X: 12-cores/24-threads, 3.7 GHz base clock (4.8 GHz boost) 105W TDP ($US550 ($768))
  • Ryzen 7 5800X: 8-cores/16-threads, 3.8 GHz base clock (4.7 GHz boost) 105W TDP ($US450 ($629))
  • Ryzen 5 5600X: 6-cores/12-threads, 3.7 GHz base clock (4.6 GHz boost) 65W TDP ($US300 ($419))

Editor’s Note: Stay tuned for local Australian pricing and availability.

Still based on 7nm architecture, AMD has completely redesigned the layout of its next-gen Zen 3-based processors, and the company says it’s the most significant architecture update to its Ryzen processors to date. Instead of accessing two separate 16GB 4-core cache complexes, or the part of the processor that helps reduce the time it takes to access data from the SSD, the Zen 3 architecture now has one big 32GB 8-core cache complex. AMD says every core can now communicate directly to the cache for reduced latency, which boosts processing times for every program and game you run on your PC.

According to AMD, the architectural changes also deliver a 2.4x performance per watt increase and 19% higher instructions per clock, or the amount of instructions the CPU can process in a single clock cycle. Ryzen 5000 on Zen 3 also has 24% higher power efficiency, which it claims is 2.8x more efficient that Intel’s Core i9-10900K, and 26% more fps while gaming at 1080p on high. AMD showed a side-by-side comparison of its Ryzen 9 3900XT and Ryzen 9 5900X running Shadow of the Tomb Raider at 1080p, which is one of the games we use in our benchmarking tests, and the Ryzen 9 5900X managed 40 more frames than the 3900XT, 181 fps to 141 fps.

AMD Is Sticking With Its AM4 Socket and That’s Great If You Want to Build a PC

It’s a small thing that only people building and upgrading their own PCs will appreciate but to the delight of AMD enthusiasts, the company recently announced that it would make its future Zen 3 architecture compatible with the older X470 and B450 motherboards, and it would be sticking with its...

Read more

AMD also offered a sneak peak of “Big Navi,” or the company’s next-generation of ray tracing-enabled GPUs. By the looks of things, it will be able to break 60 fps in some games at 4K on ultra, however the games shown, like Borderlands 3, are not ray tracing compatible. The company has a full event for all things RX 6000 later this month, so hopefully we’ll get a better idea of how the RX 6000 series compares to Nvidia’s latest generation of GPUs. But in the meantime, I’m remaining sceptical performance-wise until I have one of those graphics cards in my hands to test.

But perhaps the most exciting part of this morning’s event was AMD’s side-by-side comparison of its Ryzen 9 5900X and Intel’s Core i9-10900K running on Cinebench; AMD appears to have overtaken Intel in single core processing speed by a whopping 87 points. Now, we don’t know how AMD configured its test bench, so it’s unclear how these results will scale in comparison across the other Ryzen 5000 chips and Intel’s 10th-gen, especially at different spec configurations. But at a glance, it’s absurdly impressive.

The Core i9-10900K is supposed to get up to a 5.3 GHz frequency, which should still make it the faster CPU in single threaded workloads because its frequency is higher than the Ryzen 9 5900X. However, there are a few scenarios where AMD’s chips could be faster. In my testing, the i9-10900K never went above 5.0 GHz. Intel’s frequency boosting tech never kicked in because the CPU temperature exceeded the maximum threshold, so that extra 300 MHz was lost. AMD’s higher-end processors are 100-200 MHz slower than the i9-10900K, but it seems like the Zen 3 architectural design changes make up for that. I’m looking forward to getting the Ryzen 5000-series in my hands to see it for myself.