ZYCAD ANNOUNCES INDUSTRY'S FIRST SIMULATION SUPERCOMPUTER; VERIFIES HIGH-END DESIGNS IN HOURS, NOT DAYS Breakthrough Architecture Pushes Verification Performance Envelope FREMONT, Calif., March 11 /PRNewswire-AsiaNet/ -- Targeting the high-end of the electronic design market, Zycad Corporation (OTC: ZCAD) today announced the industry's first massively parallel platform that delivers supercomputer-level performance for simulation of microprocessor, ASIC and system designs with up to 16 million gates. Zycad claims that the architecture of its new LightSpeed family of simulation servers allows a 400 times performance improvement over the simulation products on the market today. According to Zycad, the LightSpeed simulation simulator attains an industry-leading performance in the range of 2,000 instruction cycles per second (cps) in benchmark simulations of a 486 processor. By comparison, simulation of a 486 using a fast implementation of Verilog on a workstation can be done at about 5 cps. "LightSpeed is the only simulation system that delivers the extraordinarily high simulation performance required by semiconductor manufacturers and ASIC design houses building state-of-the-art ASICs and systems," said Martin Baynes, Vice President of Engineering for Zycad's Acceleration Division. "By allowing useful simulations of real-world applications rather than just targeted test cases, LightSpeed enables high-end electronic component and system developers to minimize their chip turns and thereby greatly reduce their design costs and time-to-market for virtually any design." Baynes noted that LightSpeed's high simulation performance will enable many high-end microprocessor and other IC design teams to forego the complex, time-consuming and expensive process of design emulation (also called rapid prototyping). "With LightSpeed, many designs that previously required emulation can now be verified using a familiar, easy- to-implement, and relatively low-cost simulation methodology," he said. "And in cases where emulation is still needed, LightSpeed can also be used synergistically to verify the emulation gate-level netlist before it is translated into the emulator's FPGA netlist, thereby significantly enhancing the overall verification process." SIMULATOR IMPLEMENTS LEADING EDGE TECHNOLOGIES AND DESIGN INNOVATIONS In building LightSpeed, Zycad leveraged the latest component technologies and design innovations to ensure the system's clear performance leadership in simulation of high-end electronic designs. Advanced technologies implemented in LightSpeed include: * Simulation and communications ASICs were implemented in 0.5 micron technology from LSI Logic Corp. The simulation ASIC contains more than 2.2 million transistors (which includes 190 kilobits of SRAM on the ASIC chip). The communications chip contains 400,000 transistors. * "LSI Logic's 500K cell-based ASIC family, with its advanced cell design, is ideal for high performance custom processor applications like Zycad's LightSpeed simulation server," said Steve Henderson, US Business Manager for LSI Logic. * State-of-the-art surface-mount packaging relied on a 388-pin plastic ball grid array (PBGA) technology. "LSI Logic's PBGA family maximizes board packing densities at reasonable cost, as required for LightSpeed," Henderson said. * Very high-speed asynchronous SRAMs. These are the IDT71V016, a one-megabit, 15ns, 3.3 Volt SRAM. "IDT's high performance 16 bit wide SRAM enabled Zycad to achieve their massively parallel simulation supercomputer within a reasonable power budget," said Stewart Sando, Director of SRAM Marketing at Integrated Device Technology, Inc. * Leading-edge board design and assembly, with 68 ASICs on one side of a 20 by 16-inch board, and 192 SRAMs on the other side, packed in an extremely dense design. "With about 25,000 solder joints, the processor card presents a formidable challenge for the testing process and for achieving high manufacturing yields," said Bill Barthel, Process Engineering Manager at Plexus Corporation, the board co-developer and manufacturer. "To achieve the necessary per-joint yield in the single-digit parts per million over such a large area, we are using process control verified with a Four-Pi 5DX system, the industry's most advanced X-ray laminography technology." (more)