The top semiconductor and hardware companies have arrived in Taipei, Taiwan for Computex 2019. The world’s biggest computer show is a preferred stage for the likes of AMD, Intel, NVIDIA, and Arm to showcase their latest chip designs. The tech giants are also vying for the vanguard in the 2020 micro processor market, and hoping to boost confidence in the semiconductor market, which has been hit hard in the intensifying US-China trade war.
Synced’s picks of the biggest rollouts at Computex 2019.
AMD: New Ryzen 3000 series challenges Intel
In her keynote speech AMD CEO Lisa Su announced the new 7nm processors for PCs, Ryzen 3000 series. The US chip manufacturer promoted the compelling performance and cost-effectiveness of the Ryzen 3000 chip family compared to the offerings from rival Intel. The market reacted positively, with AMD shares rising 10 percent today.
All Ryzen 3000 processors are based on the new Zen 2 core architecture and are supported byPCIe 4.0 PC readiness. The high-end flagship processor Ryzen 9 3900X features a 12-core and 24-thread design, a base frequency of 3.8 GHz, a turbo frequency of 4.6 GHz, a TDP of 105W, 40 PCIe 4.0 lanes, and costs only US$499. AMD claims the Ryzen 9 3900X outperforms the Intel i9 9920X by more than 16 percent.
Below is the full Ryzen 3000X product stack. Amandtech joked that products will ship on July 7 because they are built on TSMC’s 7nm process.
Also announced in the AMD keynote were the new 7nm graphics cards AMD Radeon RX 5700-series — which feature high-speed GDDR6 memory and support PCIe 4.0 — and the second-generation AMD EPYC server processor — which will double the performance-per-socket and quadruple the floating performance-per-socket over the previous generation. The products will be shipped in July and Quarter 3 respectively.
Intel: 10th Gen Intel Core CPUs bring AI capability to laptops
Intel’s big announcement today was its long-awaited 10nm Ice Lake processors — officially named “10th Generation Core” processors. The 10th Gen Core processors so far only range from the Intel Core i3 to Intel Core i7, with up to 4 cores and 8 threads, up to 4.1 max turbo frequency and up to 1.1 GHz graphics frequency. They are based on the new Sunny Cove core architecture and the new 11th Generation graphics engine Iris Plus to double the visual performance for gaming.
What’s new about 10th Gen Core processors is their support of AI workloads on laptops, delivering up to 2.5 times the AI performance over the previous generation thanks to the new Deep Learning Boost (DL Boost) technology designed to accelerate the inference performance of deep learning workloads. DL Boost includes AVX-512 Vector Neural Network Instructions to accelerate CNN-based models, and a brain floating-point format for use hardware accelerating machine learning algorithms.
Arm: New Cortex-A77 CPU and Mali-G77 GPU are coming to smartphones in 2020
UK-based, SoftBank-owned intellectual property (IP) provider Arm provides chip solutions for most of the smart electronics we use daily — from IoT devices to smartphones to assisted driving systems. The company iterates the microarchitecture design of its CPUs (Cortex), GPUs (Mali) and other chip series annually, and did so this year at Computex.
Arm announced its new Cortex-A77 CPU, Mali-G77 GPU and Arm ML (machine learning) processor, which are designed to empower OEMs’ 2020 premium smartphones. Improvements over previous generations:
- The 7nm Cortex-A77 CPU brings 20 percent more instructions per cycle (IPC) performance improvement, 35 percent more floating-point (FP) performance, and 15 percent more memory bandwidth over the previous Cortex-A76.
- The Mali-77 GPU is Arm’s graphics unit based on the new Valhall architecture, which boosts its performance by 40 percent over the previous Mali-G76. The Mali-G77 also achieves a 30 percent increase in performance density, 30 percent energy efficiency improvement, and 60 percent improvement in machine learning.
- Since Arm announced its ML processor last year, the latest IP has doubled its energy efficiency to 5 TOPs/W, tripled its memory compression, and boosted peak performance to 32 TOP/s.
One more thing: The recent US sanctions on Chinese telecommunication giant Huawei led Arm to suspend its business with Huawei last week. Because Huawei will not be allowed to use Arm’s latest microarchitecture IPs in its next-generation flagship phones, the world’s second largest smartphone maker is pinning its future on its semiconductor subsidiary, Shenzhen-based HiSilicon, a chip veteran founded in 1991.
NVIDIA: New EGX drives AI at the edge
While US GPU giant NVIDIA didn’t unveil any new chips in Taipei, the company signaled an ambitious move by announcing EGX, a new edge computing platform which enables low-latency AI acceleration at the edge — 5G base stations, warehouses, retail stores, factories and beyond.
It is estimated that by 2025 over 150 billion machine sensors and IoT devices will be streaming continuous data. All that will need to be processed. NVIDIA Vice President and General Manager of Enterprise and Edge Computing Bob Pette told the Computex 2019 audience that “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. 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 product stack ranges from NVIDIA Jetson Nano, a cost-effective robotic processor for beginners; to NVIDIA T4 servers, which deliver more than 10,000 TOPS for real-time AI tasks. NVIDIA also leverages technologies from Red Hat, Mellanox (which it acquired months ago), and Cisco to enhance its enterprise-grade operation, network, and storage. Major cloud service providers including AWS and Microsoft Azure, have supported the EGX platform.
Computex 2019 runs to June 1 at the Nangang Exhibition Center in Taipei, Taiwan.
Journalist: Tony Peng | Editor: Michael Sarazen