Computex 2022: AMD Ryzen 7000 Processors Offer Faster CPU Cores, RDNA 2 Graphics
Zen 4 Can Hit Very High Clock Speeds, But How Much Faster is It?
AMD has shared some more details about their upcoming Zen 4 desktop processors, and demonstrated impressive clock speeds with an engineering sample of one such CPU hitting 5.5 GHz during their Computex 2022 showcase.
Quoting AMD’s press release (link):
Continuing to expand on the innovation and leadership performance of the Ryzen 5000 Series desktop processors, the new Ryzen 7000 Series desktop processors are based on the extremely efficient 5nm “Zen 4” architecture. The new processors will double the amount of L2 cache per core, feature higher clock speeds, and are projected to provide greater than 15% uplift in single-thread performance versus the prior generation, for an unmatched desktop PC experience. During the keynote, a pre-production Ryzen 7000 Series desktop processor was demonstrated running at 5.5 GHz clock speed throughout AAA game play. The same processor was also demonstrated performing more than 30% faster than an Intel Core i9 12900K in a Blender multi-threaded rendering workload.
In addition to new “Zen 4” compute dies, the Ryzen 7000 series features an all-new 6nm I/O die. The new I/O die includes AMD RDNA 2-based graphics engine, a new low-power architecture adopted from AMD Ryzen mobile processors, support for the latest memory and connectivity technologies like DDR5 and PCI Express 5.0, and support for up to 4 displays.
In addition to faster performance over the previous generation (more on this in a moment), AMD has outlined the platform that the new CPUs will reside in:
The new AMD Socket AM5 platform provides advanced connectivity for our most demanding enthusiasts. This new socket features a 1718-pin LGA design with support for up to 170W TDP processors, dual-channel DDR5 memory, and new SVI3 power infrastructure for leading all-core performance with our Ryzen 7000 Series processors. AMD Socket AM5 features the most PCIe 5.0 lanes in the industry with up to 24 lanes, making it our fastest, largest, and most expansive desktop platform with support for the next-generation and beyond class of storage and graphics cards.
The AM5 family features three levels of motherboards:
- X670 Extreme: Bringing the most connectivity and extreme overclocking capabilities with PCIe 5.0 support for two graphics slots and one storage slot
- X670: Supporting enthusiast overclocking with PCIe 5.0 support on one storage slot with graphics support optional
- B650: Designed for performance users with PCIe 5.0 storage support
Single Thread Performance Uplift Examined
While some have taken the “greater than 15% uplift in single-thread performance” as an indication of Zen 4 IPC improvements, there are a number of factors at play here. AMD does not make any such IPC (instructions per clock) claim, as the uplift is a combination of the architectural improvements (including double the L2 cache per core), increased clock speeds, and the move to higher-bandwidth DDR5 memory.
Let’s look at the relevant footnote from AMD’s press release, explaining under what conditions the new Zen 4 sample was compared to a Ryzen 9 5950X:
“Testing as of May 5, 2022, by AMD Performance Labs. Single-thread performance evaluated with Cinebench R23 1T. AMD Ryzen 9 5950X System: ASUS ROG Crosshair VIII Hero X570, 2×8 DDR4-3600C16. AMD Ryzen 7000 Series: AMD Reference X670 Motherboard, 16-core Ryzen 7000 Series pre-production processor sample, 2x16GB DDR5-6000CL30. All systems configured with Radeon RX 6950XT GPU (driver: 22.10 Prime), Windows 11 Build 22000.593, Samsung 980 Pro 1TB SSD, Asetek 280MM liquid cooler. Results may vary when final products are released in market.”
The 16-core Ryzen 7000 pre-production processor sample was paired with 32GB (2 x 16GB) of DDR5 at 6000 MT/s, CL30. The Ryzen 9 5950X it was compared to had just 16GB (2 x 8GB) of DDR4 at 3600 MT/s, CL16. Why they didn’t use 32 GB with the Ryzen 9 5950X is unknown, though in a single-threaded workload like Cinebench R23 I doubt it made any difference. It still looks a little odd, particularly as AMD used such low latency DDR5 with the Zen 4 system (yes, CL30 at 6000 MT/s is actually very low for DDR5), and CL16 DDR4, rather than CL14, in the Zen 3 system. Sure, the difference would probably be negligible in this test, but I’m sure AMD has access to 32GB of the lower-latency DDR4.
One more thing: I am not seeing 3D V-Cache mentioned anywhere, and AMD did not test the new 16-core Ryzen 7000 pre-production processor sample’s single-thread performance against their recent Ryzen 7 5800X3D – instead focusing on a comparison to the Ryzen 9 5950X. Now, at 5.5 GHz (and with DDR5-6000 CL30) the new CPU was going to win, regardless.
AMD at Computex 2022 Replay
You can peruse AMD’s announcements on their Computex page (linked here), and view the YouTube video of AMD’s full event (embedded below):