Ceremorphic Introduces Custom Silicon Development for Advanced Nodes Using In-House Technology to Speed Customer HPC Chip Development
January 10, 2023 --Hyderabad, India – Ceremorphic, a fabless silicon and system development company, today announced an expansion of its business to include a new business model for the high-performance computing (HPC) industry. Leveraging the company’s proven expertise in developing chips in 5nm and below for AI supercomputing and other HPC applications, Ceremorphic will now also offer its IP and design expertise to develop advanced ICs for select customers looking to differentiate their business with customized silicon. For more information, refer to the new “Custom Engineering” page on the Ceremorphic website.
The development cycle for advanced, lower geometry silicon devices is lengthy and costly due to the complex IP and expertise required, even for well-resourced system companies. To solve this problem, Ceremorphic has been working on infrastructure IP for high performance AI supercomputing for many years and has successfully tested its design methodology with a 5nm device tape out in 2022. The company has also developed new methods for connectivity, advanced analog technology and algorithms and AI applications to address HPC development challenges with cost-effective solutions. This includes critical IPs in CXL/PCIe6, machine learning processors, ASIL-D compliant reliable multi-thread processors along with all analog IP needed for advanced nodes.
“Increasing trends towards developing custom silicon by a system company coupled with the design complexity of silicon in advanced geometries provides an opportunity for capable silicon companies to capture the value for their efforts,” said Subhasish Mitra, Professor of Electrical Engineering and of Computer Science at Stanford University. “Ceremorphic brings unique experience in developing infrastructure technology required for high-performance computing platforms, giving them a significant advantage to capitalize on this market opportunity.”
“The combination of Ceremorphic’s silicon and software technology with its own product development expertise gives us a unique position to offer this model to high-end system developers in the AI, machine learning and HPC space,” said Dr. Venkat Mattela, Founder and CEO of Ceremorphic. “The drive for this business comes more from viability for a system company to own custom IC to differentiate their product than the cost equation. As a result, we will offer this service to very select customers that want to develop an advanced IC with minimal extra effort from their own product roadmap.”
Ceremorphic’s IP is supported by a strong patent portfolio, including over 20 awarded patents, over 100 filed patents, and more than 30 licensed patents. The portfolio covers a range of areas, including system level interconnect, chiplet-to-chiplet interconnect, reliable circuits, multi-thread microprocessor technology, machine learning processors, graph neural processors, low power silicon design, quantum resistant security, programmable logic, low power algorithms, and firmware architectures, all of which are relevant to the development of applications with advanced silicon in 5nm and below.
Ceremorphic to Keynote and Exhibit at the 2023 VLSI Design Conference
As a recognized industry expert, the Ceremorphic CEO Venkat Mattela will be giving a keynote at this week’s VLSI Design Conference on January 10th at 12:45 pm at the Novotel, HICC in Hyderabad, India. Ceremorphic will also be exhibiting at the conference in booth #20.
For more information on the conference or the keynote, please visit the conference website at this link.
About Ceremorphic
Founded in April 2020, Ceremorphic currently has 150 full-time employees dedicated to developing advanced silicon and software products for next-generation computing systems. Leveraging more than 100 patents and proven expertise in creating industry-leading silicon system products, Ceremorphic has built a new architecture that delivers the performance needed for applications such as AI model training, automotive processing, drug discovery, HPC, and metaverse processing with unprecedented energy efficiency and reliability. Designed in advanced silicon technology node (TSMC 5nm), this architecture solves today’s high-performance computing problems in reliability, security and energy consumption across all performance-demanding market segments. For more information, visit https://www.ceremorphic.com.
|
Related News
- Arm and Synopsys Strengthen Partnership to Accelerate Custom Silicon on Advanced Nodes
- IBM introduces powerful software to speed development of custom chips
- eMemory Won TSMC OIP Partner of the Year Award for the Outstanding Development of its NVM IP on Advanced Nodes
- Efabless Announces the Release of the OpenLane 2 Development Platform, Transforming Custom Silicon Design Flows
- In-house software team provides Sondrel with a key advantage for its turnkey ASIC service of turning concept into silicon
Breaking News
- Logic Design Solutions launches Gen4 NVMe host IP
- ULYSS1, Microcontroller (MCU) for Automotive market, designed by Cortus is available
- M31 is partnering with Taiwan Cooperative Bank to launch an Employee Stock Ownership Trust to strengthen talent retention
- Sondrel announces CEO transition to lead next phase of growth
- JEDEC Publishes LPDDR5 CAMM2 Connector Performance Standard
Most Popular
- Arm's power play will backfire
- Alphawave Semi Selected for AI Innovation Research Grant from UK Government's Advanced Research + Invention Agency
- Secure-IC obtains the first worldwide CAVP Certification of Post-Quantum Cryptography algorithms, tested by SERMA Safety & Security
- Weebit Nano continuing to make progress with potential customers and qualifying its technology Moving closer to finalisation of licensing agreements Q1 FY25 Quarterly Activities Report
- PUFsecurity Collaborate with Arm on PSA Certified RoT Component Level 3 Certification for its Crypto Coprocessor to Provide Robust Security Subsystem Essential for the AIoT era
E-mail This Article | Printer-Friendly Page |