Leslie D. Kohn

Introducing the Intel i860 64-bit microprocessor

The authors describe the single-chip i860 CPU, a 64-bit, RISC (reduced-instruction-set-computer)-based microprocessor that executes parallel instructions using mainframe and supercomputer architectural concepts. They designed the 1,000,000-transistor, 10-mm*15-mm processor for balanced integer, floating-point, and graphics performance. They discuss the RISC core, memory management, floating-point unit, graphics, bus interface, software support, and interfacing to a DRAM system.<<ETX>>

The Intel i860 64-bit processor: a general-purpose CPU with 3D graphics capabilities

A description is given of the Intel i860 processor, which was designed for use in low-cost 3D graphics workstations. These products need high performance to support the operating system, simulations, transforms, and shading calculations. The integer unit provides high performance for Unix and other operating systems, the floating-point unit provides high throughput for computationally intensive applications and the processor includes 3D graphics hardware. The sustained high computation rates are supported by on-chip caches. The processor is supported by compilers and other development tools. The i860 processor will enable cost-effective interactive scientific visualization. This includes not only the numeric computation and 3D matrix transforms but also the pixel-intensive shading calculations that normally are either very costly or very slow. Integrating all these capabilities into one component allows design of a desktop unit or a single add-in card for existing personal computers.<<ETX>>

The i860 TM 64-bit supercomputing microprocessor

The Intel i860TM processor is a RISC-based microprocessor incorporating a RISC core with memory management, a floating point unit, and caches on a single chip. The 1,000,000 transistors allow a single chip implementation with highly optimized interunit communication and wide internal data buses. The parallelism and pipelining between the execution units, and the innovative cache management techniques are under explicit control of software. Vectorizable applications can use the pipelined adder and multiplier units to achieve up to 80 Mflops at 40 Mhz for the inner loops of common calculations. Special instructions allow using the data cache as a flexible vector register and support high data bandwidth from main memory. Finally, to support visualization of data, special hardware for 3D graphics is included.

Distributed processing with the NS16000 family

The paper discusses the benefits of microprocessor-based distributed processing systems, which are greater than those of conventional timeshared mini or main-frame systems. It highlights the advantages of the NS16000 microprocessor family for this application, and it explains how the NS16000 operating system supports distributed processing.