Flood of serial entrants confounds designers
Flood of serial entrants confounds designers
By Rick Merritt, EE Times
October 28, 2002 (10:15 a.m. EST)
URL: http://www.eetimes.com/story/OEG20021024S0010
About the time microprocessors were approaching gigahertz speeds, I/O was the systems performance bottleneck. Not so today. Fast serial I/O technologies, prepared for anything muscular multiprocessor systems can throw at them, are at the ready-a plethora of them, in fact.
Today's systems engineer faces technically complex and sometimes politically difficult decisions about which interconnect to use. To help, we asked a group of experts to weigh in on which interconnects they are using where, when and why in this week's In Focus section. It may not make a designer's decision any easier, but it will at least remind you that others feel your pain-and perhaps provide some insights as other engineers trace their thinking processes.
While the main focus in this section is the embedded world, we also check in with the giants of computing. After all, new technologies are often tested out in the mainstream computer industry first.
Thus, t op engineers at IBM and Hewlett-Packard are among our authors, as is a representative of server chip set maker ServerWorks. Interestingly, Mellanox, which makes Infiniband chips, decided it could not determine which local interconnects would dominate in the future.
On the embedded side, Bustronic details its picks from the growing world of fabrics for backplane-based systems. PLX Technologies discusses the roles of HyperTransport and PCI Express, while the RapidIO backers articulate their views of what the embedded world needs. And Intel provides an update on its Advanced Switching spec, designed to bring PCI Express to communications systems.
On the chip-to-chip front, follow the microprocessor. Intel will push PCI Express through its Pentium desktop and notebook chips sets in 2004. AMD is likely to do the same with HyperTransport in its high-end Opteron desktops and servers. Ditto ServerWorks with PCI-X 2.0 in server chip sets. Motorola and IBM are gearing up their embedded PowerPCs with R apidIO. All will find enough of a foothold to hang on, creating a polyglot world.
Things get even more complex in the system-to-system and backplane arenas. Fibre Channel has finally caught on in storage, just as Infiniband is trying to gain traction in computing. Ethernet casts a long shadow over all, but at the gigabit-and-above level it will require significant silicon work.
Related Semiconductor IP
- Sine Wave Frequency Generator
- CAN XL Verification IP
- Rad-Hard GPIO, ODIO & LVDS in SkyWater 90nm
- 1.22V/1uA Reference voltage and current source
- 1.2V SLVS Transceiver in UMC 110nm
Related White Papers
- Altera courts ASIC designers with block-based Stratix PLD
- Reconfiguring Design -> Reconfigurability: Designer's key strategy
- High-Performance DSPs -> Serial interconnects back high-performance computing
- MPEG Standards -> MPEG-4: A system designer's view
Latest White Papers
- OmniSim: Simulating Hardware with C Speed and RTL Accuracy for High-Level Synthesis Designs
- Balancing Power and Performance With Task Dependencies in Multi-Core Systems
- LLM Inference with Codebook-based Q4X Quantization using the Llama.cpp Framework on RISC-V Vector CPUs
- PCIe 5.0: The universal high-speed interconnect for High Bandwidth and Low Latency Applications Design Challenges & Solutions
- Basilisk: A 34 mm2 End-to-End Open-Source 64-bit Linux-Capable RISC-V SoC in 130nm BiCMOS