Intel recently held its annual innovation event IntelON. IntelON is a developer-oriented conference, and last year’s IntelON event focused on the value of developers as a center of value creation. IntelON 2022 has strengthened its attention to developers and expanded it to open source community . Intel has issued many important announcements in the data center, including a new GPU and thirteenth-generation Core processors.
Moore's Law is still valid
Kissinger (Intel CEO) has always been a strong supporter of Moore's Law, that is, the number of transistors will double with the advancement of each generation of semiconductors. Kissinger predicts that chips will have one trillion transistors in 2030, compared with 100 billion today. Kissinger doubled down on his “Four Year Five Node Plan”.
I believe Intel's differentiated end-to-end foundry technology has incredible capabilities, and I'm glad the company didn't spin out this capability. IFS has four parts—wafers, packaging, software and an open chip ecosystem. Kissinger highlights how novel packaging provides designers with tools to increase the number of transistors per device, thus following Moore's Law.
Intel announced the expansion of the UCIe alliance to create an open chiplet ecosystem. Currently, more than 80 companies are involved in the UCIe alliance, including , Samsung and TSMC. Intel’s vision for UCIe is that components from different companies in the alliance can be packaged together, reducing time to market because it will be easier to package different chiplets than to package and test monolithic designs. I'd love to know how this will affect the software aspect, because software becomes inherently less proprietary as chip designers pursue the open chiplet ecosystem offered by the alliance. I can imagine it either slows down the software development process of chipsets in a short period of time or speeds up software development as more UCIe design progresses.
software is the highest and most time-consuming chip cost. As chip design becomes more complex, chip-small software becomes equally complex. If UCIe can somehow accelerate software development while improving competitiveness, I believe chip design may accelerate in the coming years.
New Data Center and Games GPU
Intel released the data center GPU-Flex series earlier the week before, the first product to support AV1 encoding in data centers and support Tensorflow, OpenVINO and PyTorch. As AI becomes increasingly popular in data centers, support for these AI and deep learning frameworks has become crucial. Intel said it should provide customers with a single GPU solution for a wide range of visual cloud workloads.
Intel also announced that it will launch the "highest performance data center GPU" Ponte Vecchio. Ponte Vecchio uses the Intel Xe HPC architecture for high performance computing (HPC) and artificial intelligence (AI). Xe HPC is a scalable architecture with over 100 billion transistors, and Intel claims its computing power is 45 TFLOPS. Its targets A-H100 HPC GPU
Another GPU announced by Intel is the Intel Arc A770 GPU for gamers. Many people don’t believe Intel will follow its promise to release the Arc A770. Although the A770 is not the fastest consumer-grade GPU on the market, Intel’s goal is to achieve the best cost-effectiveness. Priced at $329, nearly $100 less than the average selling price of the similarly performing NVIDIA GeForce 3060. Intel also claims that the Arc A770's peak ray tracing performance is 65% higher than that of NVIDIA.
Intel did not re-enter the discrete graphics card market until recently, and its mobile version of Arc has achieved some success and won many designs so far. While Arc 5 and Arc 7 are not appearing in laptop , I'm glad to see Intel release it on desktops.
Raptor Lake
Kissinger also announced its 13th generation core processor, codenamed Raptor Lake, has up to 24 performance cores, a single-threaded performance improvement of 15% and a multi-threaded performance improvement of 41%. Intel said it improved the P-core cache architecture and doubled the E-core.
This generation is Intel's second generation multi-provisioned kernel product, taking into account both efficiency and performance core configuration. Intel has increased the number of performance cores from 16 to 24, which is why Intel claims multithreaded performance is 41% higher. It will also have a 6.0Ghz clock speed out of the box. Kissinger claims the 13th-generation Intel Core i9 13900K is the fastest desktop processor in the world.
The 13th generation Core processor also supports DDR5-5600 and DDR5-5200 and DRR4, and supports Thunderbolt 4 and Intel Killer WiFi 6E. I believe Intel's 13th-generation Core processors may have better multithreading performance due to performance core improvements. If this is true, this seems to be a role conversion for AMD and Intel, considering that AMD was once known for its multi-threaded performance, while Intel is known for its single-threaded performance. While clock speed isn't everything, 6.0GHz out of the box shows that there is a lot of performance boost potential.
Summary
Kissinger continues to make Intel a manager of Moore's Law, and Intel's differentiated end-to-end foundry has an incredible future. I'm also interested in learning how UCIe changes the market and how fast chiplets are being developed across a wider range of players.
Intel has also taken huge steps in its standalone GPU, announcing the launch of three new GPUs, two for data centers and one Arc A770, and many even doubt it will be launched this year. Intel also announced its 13th-generation Core processors with 6.0 GHz clock speeds and better single-threaded and multi-threaded performance.
