IP promise still can be kept
IP promise still can be kept
By Brian Fuller, EE Times
June 5, 2003 (10:09 a.m. EST)
URL: http://www.eetimes.com/story/OEG20030605S0022
ANAHEIM, Calif. The original promise of the IP methodology was broken sometime ago, but the industry still needs to keep the promise, a panel agreed here Wednesday (June 4) at the Design Automation Conference. "Thet original promise of IP was that you could use it as a virtual component. Yes, it's been a largely broken promise. However that is still what the IP industry should aspire to do," said Craig Lytle vice president of the IP business unit at Altera Corp. (San Jose, Calif). The panel, sponsored by VSI Alliance and moderated by EE Times senior editor Ron Wilson, shot holes through the notion that the IP business itself is dead. Bart De Loore, general manager of reuse technology at Philips, said his company has 450 cores from 20 internal and external providers. "Most are assumed for design re-use. Design for reuse has been elevated to a company policy and it's being used all over the place," he said. But where perhaps th e under-appreciated model starts to have difficulty is in the area of optimization. Some panelists insisted that no matter how rock solid a core may be, customer usually wants to optimize it for their designs. Among other things this can lead to costly continued maintenance with the customer. Bernie Rosenthal, senior vice president of marketing and sales for Tensilica Inc., noted that his company has automated the tool chain "to, in effect, constrain the user to valid states of IP." Lytle said Altera has sold more 13,000, and of those 250-300 have been customization projects. "We'll do them ourselves and optimize the RTL for them," he said. Ultimately you have use the IP in some sort of differentiated type of way," said Ronnie Vasishta, vice president of technology marketing and Coreware at LSI Logic. "That's where a lot of initial IP companies failed to realize. They said 'well, here's the RTL. Now we're done." Kurt Wolf, director of library management at TSMC, cautioned about an oversimp listic and over-demanding view of IP providers. Too many people still think acquiring an IP core is like shopping for food at a grocery store "and it's magically going to work. It doesn't work that way. It won't and it never has. Yeah, there's science to what we're doing but there's also art." Wolf said everybody in the industry needs to be "healthy" for the whole SOC industry to thrive. "You have to be careful what you wish for in terms of beating up your IP providers for more and more at less and less cost," he said.
Related Semiconductor IP
- HBM4 PHY IP
- Ultra-Low-Power LPDDR3/LPDDR2/DDR3L Combo Subsystem
- MIPI D-PHY and FPD-Link (LVDS) Combinational Transmitter for TSMC 22nm ULP
- HBM4 Controller IP
- IPSEC AES-256-GCM (Standalone IPsec)
Related Articles
- Creating IP level test cases which can be reused at SoC level
- SOC: Submicron Issues -> Deep signal integrity can be assured
- RF ICs can be quite complex
- Can the chip design - verification divide be plugged in with some knee jerk band aid type strategies?
Latest Articles
- ElfCore: A 28nm Neural Processor Enabling Dynamic Structured Sparse Training and Online Self-Supervised Learning with Activity-Dependent Weight Update
- A 14ns-Latency 9Gb/s 0.44mm² 62pJ/b Short-Blocklength LDPC Decoder ASIC in 22FDX
- Pipeline Stage Resolved Timing Characterization of FPGA and ASIC Implementations of a RISC V Processor
- Lyra: A Hardware-Accelerated RISC-V Verification Framework with Generative Model-Based Processor Fuzzing
- Leveraging FPGAs for Homomorphic Matrix-Vector Multiplication in Oblivious Message Retrieval