AMD Announces 7nm Rome CPUs and MI60 GPUs - Breaking News

By Paul Alcorn November 6, 2018 at 9:00 AM

We're here at AMD's New Horizon Event to bring you up to the minute news on the company's 7nm products. This is breaking news, so check back frequently or refresh the page for updates.

AMD is expected to make major announcements about its new 7nm CPUs and GPUs. Intel continues to struggle with its 10nm manufacturing process, which is delayed until late 2019. If AMD can field 7nm processors early this year, it will mark the first time in the company's history that it has had a process node leadership position over Intel. That should equate to faster, denser, and less power-hungry processors than Intel's 14nm chips.

AMD CEO Lisa Su is delivering the opening statements. She is recounting the path of the EPYC data center processors to market and discussing many of the key roles those processors are used for, such as workloads in HPC, cloud, hyperscale, and virtualization environments. AMD sees the data center as a $29 billion opportunity by 2021, and GPUs are playing a larger role as the industry shifts to artificial intelligence and machine learning workloads.

Amazon Web Services, one of the world's largest cloud service providers, announced that, beginning today, it is offering new EPYC-powered cloud instances. The R5a, M5a and T3a instances purportedly will offer a 10% pricing advantage over AWS's other cloud instances.

Mark Papermaster will cover the new Zen 2 cores and AMD's 7nm process technology. AMD will also revel more about the 7nm Mi60 Instinct GPU for the data center. AMD will also provide early specifications for AMD's Rome, the first x86 7nm processor for the data center. 

AMD is already sampling its 7nm Rome processors, which mark th debut of the Zen 2 microarchitecture, to customers. The firm also has its Zen 3 processors under development. This third-gen microarchitecture will debut on the 7nm+ process, with the "+" indicating this will be a second generation of the 7nm node (a "tock" equivalent). 

The new processors features faster, smaller, and lower-power transistors. As with all new nodes, development requires a significant investment from both AMD and TSMC, which will fab the parts.

Zen 2 will provide up to 2X the compute power per node, improved execution pipeline, doubled core density, and use half the energy per operation. AMD has doubled floating point performance with the Zen 2 microarchitecture.

AMD has improved the branch predictor and pre-fetching. AMD also doubled the vector width to 256-bit. We'll follow up with deeper analysis. This type of technical data isn't well suited for live public disclosure. 

AMD has improved the Infinity Fabric. AMD is now using the second-gen Infinity Fabric to connect a multi-chip design with a 14nm I/O die serving as the linchpin of the design, tying together 7nm CPU chiplets. We'll follow up with deeper analysis of this design.

David Wang displayed the world's first 7nm GPU. The die wields 13.28B transistors and measures just 331mm2. The GPU is based on the advanced Vega architecture and is the first PCIe 4.0 GPU on the market. It also is the first to use the Infinity Fabric over the external PCIe bus.

AMD also presented performance benchmarks highlighting the generational performance gains relative to 12nm GPUs, and also scalability benchmarks to highlight gains due to the increased PCIe 4.0 bandwidth. The company also provided direct comparisons to Nvidia's V100 GPUs.