Xilinx, IBM ink cores agreement
Xilinx, IBM ink cores agreement
By Craig Matsumoto, EE Times
July 26, 2000 (11:23 a.m. EST)
URL: http://www.eetimes.com/story/OEG20000726S0007
SAN JOSE, Calif. Xilinx Inc. and IBM Corp. Tuesday (July 25) announced an agreement to meld PowerPC hard cores with Xilinx's FPGAs, allowing an ASIC and an FPGA vendor to help one another in the market. Xilinx will license from IBM the PowerPC as well as the CoreConnect bus. Xilinx will enlist IBM as a foundry to manufacture Xilinx parts both with and without the PowerPC cores. The deal is similar to Altera Corp.'s plans to license MIPS and ARM processors for integration into the Apex line of FPGAs. But Xilinx CEO Wim Roelandts said his company has a more focused goal: using hard cores to get higher performance out of FPGAs, especially for networking applications. "(Altera is) talking about 150 MIPs, but we're talking two or three times that performance," Roelandts said. Meanwhile, Altera (San Jose, Calif.) is negotiating for its own PowerPC l icense, from Motorola Inc. "We would expect to close that discussion in six months," said Cliff Tong, vice president of corporate marketing at Altera. The emphasis on using PowerPC as a high-performance hard core also makes this arrangement complementary to the recent deal to optimize the ARC Cores Ltd.'s 32-bit processor soft core for Xilinx FPGAs, Roelandts said. Xilinx targeted the PowerPC because IBM has been pressing to get the core into embedded communications applications, Roelandts said. "The PowerPC seems to be used more in the communications space than any of the other processors. That's why we chose to do PowerPC first," he said. The foundry portion of the deal augments Xilinx's foundry agreements with United Microelectronics Corp. and Seiko Epson Corp. Xilinx would eventually get access to IBM's manufacturing technology anyway due to an agreement between IBM and UMC, but the ne w arrangement will let Xilinx get first crack at some of IBM's new processes. IBM will be able to use Xilinx's advancements to push improvements in process technologies, much as the Taiwanese foundries have done. "We have different ways of advancing the technology, but this adds to our arsenal," said John Kelly, general manager of IBM's microelectronics division. In addition, Xilinx and IBM make a good fit as business partners, Roelandts said. "We are, for the high-volume customers, mainly used for prototyping. We know these customers will go to an ASIC," Roelandts said. "Therefore, a collaboration with an ASIC company is a benefit for both. We get more prototyping business, and they, for customers will lower volumes, can say, 'Why don't you stick with an FPGA?'" The first products combining Xilinx and IBM technologies will be produced on 0.13-micron design rules, officials of both companies said. There were no details as to when these products would b ecome available. Xilinx does intend to use the PowerPC hard core inside the Virtex-II FPGAs, which are not due to begin shipping until later this year. The deal does not include an IBM license of Xilinx technology. Kelly, of IBM, was willing to entertain the prospect of placing FPGA cores onto ASIC systems-on-chip, a step that LSI Logic Corp. is hoping to pioneer. "It is something that intrigues us a great deal, and it does offer us another degree of freedom if we can do it," Kelly said. He noted that the concept continues to be blocked by technical barriers, but that IBM is "trying to attack that as we speak."
Related News
- Rambus Renews License With IBM
- Xilinx and IBM First to Double Interconnect Performance for Accelerated Cloud Computing with New PCI Express Standard
- Rambus Signs License Agreement With Xilinx
- Xilinx and IBM to Enable FPGA-Based Acceleration within SuperVessel OpenPOWER Development Cloud
- IBM and Xilinx Announce Strategic Collaboration to Accelerate Data Center Applications
Breaking News
- Jury is out in the Arm vs Qualcomm trial
- Ceva Seeks To Exploit Synergies in Portfolio with Nano NPU
- Synopsys Responds to U.K. Competition and Markets Authority's Phase 1 Announcement Regarding Ansys Acquisition
- Alphawave Semi Scales UCIe™ to 64 Gbps Enabling >20 Tbps/mm Bandwidth Density for Die-to-Die Chiplet Connectivity
- RaiderChip Hardware NPU adds Falcon-3 LLM to its supported AI models
Most Popular
E-mail This Article | Printer-Friendly Page |