HPFC-6600/6640 Tachyon QE4
Quad-Channel 4 Gb Enhanced Fibre Channel PCI-Express Controller With T10 DIF

The Tachyon QE4 device (HPFC-66xx Series) is a high-performance 4- port 4 Gbit/s Fibre Channel controller. It features an 8-lane, native PCI Express bus, enabling full-duplex operation. QE4 is an integrated single chip solution ideal for a variety of high-performance I/O applications.

The Tachyon proven State-Machine architecture scales directly with system CPU performance and is not limited by the constraints of an embedded microprocessor. QE4 supports up to 4 processors per FC link using 4 independent register and API (Application Programming Interface) queue structures. This feature allows systems to scale to required performance as necessary while maintaining compatibility with the Tachyon programming model at the register level, and with the Tachyon Software Development Kit (TSDK) API tools.

The QE4 supports the T10 Standard Protection Information Model as described in the SCSI Block Commands - 2 (SBC-2) Standard, which incorporates a Data Integrity Field (DIF) in data records. Additionally, the QE4 has masking capabilities for enhanced functionality. T10 DIF provides a robust CRC-based data protection scheme with mechanisms to initiate, extend or terminate a protection domain, which can span both the PCI-Express interface and the Fibre Channel interface. When combined with Fibre Channel's standard CRC, T10 DIF offers complete data protection in the I/O data path. T10 DIF is performed in hardware to maximize performance.

QE4 offers additional performance by providing advanced framehandling functions that are normally handled by the host processor. These additional enhancements make it possible for the system to utilize intelligent frame handling while reducing CPU utilization.

The Tachyon architecture of the QE4 enables a complete hardwarebased design in a single-chip FC solution. Numerous independent functional blocks concurrently process inbound data, outbound data, control and commands in hardware. This results in the lowest latency and the highest level of performance available today.