By Nitin Dahad, EE Times
AMD certainly got an edge over Intel at Computex 2019 in Taipei last week, announcing pricing and availability of the third-generation Ryzen processors based on its 7-nm–based Zen 2 core. Meanwhile, Intel announced that it is shipping its 10-nm Intel Core Ice Lake processor and is bringing artificial intelligence (AI) to the PC. NVIDIA also revealed its EGX edge AI platform.
AMD president and CEO Lisa Su gave her first-ever Computex opening keynote on Monday. She said, “We made significant strategic investments in next-generation cores, a breakthrough chiplet design approach, and advanced process technologies to deliver leadership 7-nm products to our high-performance computing ecosystem. We are extremely excited to kick off Computex 2019 together with our industry partners as we prepare to bring our next generation of Ryzen desktop and EPYC server processors and Radeon RX gaming graphics cards to market.
AMD claims that the new Zen 2 core performance features up to 15% estimated instructions per clock (IPC) uplift over its predecessor Zen architecture. Powering both the AMD Ryzen and EPYC processors, it also includes significant design improvements, including larger cache sizes and a redesigned floating-point engine. With the new AMD Ryzen desktop processor family, AMD has introduced a new category of Ryzen 9 desktop processor with its flagship 12-core/24-thread Ryzen 9 3900X; the family is rounded out with eight-core Ryzen 7 models and six-core Ryzen 5 models. The new Ryzen desktop processor family will be available beginning July 7, with prices ranging from $199 to $499.
AMD also introduced a new X570 chipset for socket AM4, claiming to be first PCIe 4.0 supported chipset, with 42% faster storage performance than PCIe 3.09, enabling high-performance graphics card, networking devices, and NVMe drives. More than 50 new motherboard models are anticipated from ASRock, Asus, Colorful, Gigabyte, and MSI, as well as PCIe 4.0 storage solutions from partners including Galaxy, Gigabyte, and Phison. Major OEMs and system integrators like Acer, Asus, CyberPowerPC, HP, Lenovo, and MAINGEAR said that they’d be supporting the new platforms with new Ryzen processor-based gaming desktop systems in the next few months.
For its data center business, Su carried out a competitive demonstration of AMD’s second-generation EPYC server platform with Intel’s server. This showed a 2P 2nd Gen AMD EPYC based server versus a 2P Intel Xeon 8280-based server running an NAMD Apo1 v2.12 benchmark test. The pre-production AMD processor-powered server outperformed the Intel Xeon-powered servers by more than 2× on the NAMD benchmark. The second-generation EPYC server processor family is expected to launch in Q3 2019.
Microsoft Azure also announced that it achieved previously unobtainable levels of performance for computational fluid dynamics (CFD) using an Azure HB cloud instance running on a first-generation AMD EPYC processor-based system. The EPYC’s memory bandwidth enabled Azure HB to scale the Siemens Star -CCM+ application across more than 11,500 cores using a Le Mans 100 Million Cell simulation, well beyond the company’s 10,000-core goal, which had never been achieved before.
Intel brings AI to PC, reveals Project Athena innovation program
On Tuesday, May 28, Intel’s senior vice president and general manager for its client computing group, Gregory Bryant, used his industry keynote to announce that its 10th generation 10-nm Intel Core Ice Lake processors are now shipping, bringing AI to the PC for the first time, and revealed first specifications of its Project Athena innovation program.
The 10th Gen Intel Core processor wafer. The processors feature up to four cores and eight threads, up to 4.1-GHz max turbo frequency, and up to 1.1-GHz graphics frequency . (Image: Intel Corporation)
Bryant said, “No one wants to compromise. People want it all: battery life, performance, responsiveness, connectivity, and slick form factors. Our job is to come together as an industry and deliver incredible and differentiated PCs, purpose-built to what real people want. [The] 10th Gen Intel Core processors — our most integrated CPU — and Project Athena are great examples of how our deep investments at a platform level will help fuel innovation across the industry.”
Intel said that its 10th generation Intel Core processors — built on the company’s 10-nm process technology, new Sunny Cove core architecture, and new Gen11 Iris Plus graphics engine — bring high-performance AI to the PC at scale with Intel Deep Learning Boost (Intel DL Boost). The processors will range from Intel Core i3 to Intel Core i7, with up to four cores and eight threads, up to 4.1-GHz max turbo frequency, and up to 1.1-GHz graphics frequency.
The processors enable high-performance AI on the laptop, delivering approximately 2.5× AI performance with Intel DL Boost for low-latency workloads. The new graphics architecture delivers up to 1 teraflop of vector compute for heavy-duty inference workloads on highly mobile, thin, and light laptops. For low-power AI usages on the PC, Intel Gaussian Network Accelerator (GNA) is built into the SoC. The new processors also deliver both integrated Thunderbolt 3 and Intel Wi-Fi 6 (Gig+) to enable nearly 3× faster wireless speeds. Intel’s Gig+ implementation of Wi-Fi 6 connectivity delivers greater than 1-Gbps wireless speeds.
Intel also shared more details on its innovation program code-named Project Athena, including the 1.0 target specification expected to result in the first wave of laptops in the second half of this year. It previewed some of the first designs to come from partners for consumer and commercial, including Acer, Dell, HP, and Lenovo.
Project Athena is a program prioritizing enabling experiences reflective of real-world conditions as measured by “key experience indicators” (KEIs) for laptops. Intel’s goal for the new KEI metrics is to work with the ecosystem to drive innovation that noticeably impacts people’s experiences on the laptop and evolve those capabilities year over year. The first wave of KEI targets includes consistent responsiveness on battery; 16 or more hours of battery life in local video playback mode and nine or more hours of battery life under real-world performance conditions; and system wake from sleep in less than one second. The specification includes platform-level requirements across six areas: instant action, performance and responsiveness, intelligence, battery life, connectivity, and form factor.
As part of Project Athena, Intel is providing co-engineering support across the ecosystem — with more than 100 companies signed on — and new tools and open lab facilities to support verification and testing of laptops.
Nvidia targets low-latency edge AI
Continuing the AI theme, NVIDIA also used Computex to announce its EGX accelerated computing platform enabling low-latency AI at the edge to perceive, understand, and act in real time on continuous streaming data between 5G base stations, warehouses, retail stores, and factories. The company said that NVIDIA EGX was created to meet the growing demand to perform instantaneous, high-throughput AI at the edge with guaranteed response times while reducing the amount of data that must be sent to the cloud.
Targeting 150 billion machine sensors and IoT devices by 2025 that will stream continuous data needing to be processed, edge servers using its platform will more likely be distributed around the world to process data in real time from these sensors. “Enterprises demand more powerful computing at the edge to process their oceans of raw data — streaming in from countless interactions with customers and facilities — to make rapid, AI-enhanced decisions that can drive their business,” said Bob Pette, vice president and general manager of enterprise and edge computing at NVIDIA. “A scalable platform like NVIDIA EGX allows them to easily deploy systems to meet their needs on premises, in the cloud, or both.”
The EGX platform ranges from the small NVIDIA Jetson Nano, which, in a few watts, can provide 0.5 trillion operations per second (TOPS) of processing for tasks such as image recognition all the way to a full rack of NVIDIA T4 servers, delivering more than 10,000 TOPS for real-time speech recognition and other real-time AI tasks.
The company has partnered with Red Hat to integrate and optimize NVIDIA Edge Stack with OpenShift, the enterprise-grade Kubernetes container orchestration platform. NVIDIA Edge Stack is optimized software that includes NVIDIA drivers, a CUDA Kubernetes plugin, a CUDA container runtime, CUDA-X libraries and containerized AI frameworks and applications, including TensorRT, TensorRT Inference Server, and DeepStream. NVIDIA Edge Stack is optimized for certified servers and downloadable from the NVIDIA NGC registry.
Developer support for NVIDIA EGX includes video-analytics applications, which are ideal for large retail chains and smart cities, from software vendors such as AnyVision, DeepVision, IronYun, and Malong Technologies, as well as health-care–specific software offerings from 12 Sigma, Infervision, Quantib, and Subtle Medical.
EGX servers are available from enterprise computing providers ATOS, Cisco, Dell EMC, Fujitsu, Hewlett Packard Enterprise, Inspur, and Lenovo. They are also available from major server and IoT system makers including Abaco, Acer, ADLINK, Advantech, ASUS, Curtiss-Wright, GIGABYTE, and Wiwynn.
Early adopters include more than 40 companies and organizations, including BMW Group Logistics. Drawing from NVIDIA’s EGX edge computing and Isaac robotic platforms, they are able to bring AI directly to the edge of its logistics processes and handle increasingly complex logistics with real-time efficiency.
This article was originally published at sister publication EE Times .
Advertisement
Learn more about Electronic Products Magazine
