IBM Crams Power2 onto Single Chip; 8/26/96
ثبت نشده
چکیده
While putting most of its emphasis on PowerPC products from its partnership with Motorola, IBM continues to maintain its own line of POWER processors. A relatively small team, in less than a year, took the eight-chip Power2 design (see 071301.PDF) and collapsed it into a single die, delivering a powerful new product. The phenomenal integration provides manufacturing cost savings and, more important, improved performance by eliminating the overhead of chipto-chip communication. The new device, dubbed P2SC, is the most complex single-chip processor yet announced. It can issue and execute six instructions per cycle, two of which can be floating-point multiply-add instructions. It supports register renaming and out-of-order execution, although only for floating-point code; also, the P2SC’s out-of-order design uses a primitive technique compared with the reorder buffers of more modern processors. The chip includes 160K of on-chip cache, more than any other microprocessor, for a total transistor count of 15 million. Putting so much onto one chip requires an advanced IC process: the P2SC is the first processor to take advantage of IBM’s 0.29-micron CMOS-6S process (see 090905.PDF). IBM Crams Power2 o P2SC Offers Incredible Memory Bandw
منابع مشابه
POWER2 instruction cache unit
Introduction IBM introduced the POWER-based RISC System/6000@ (RS/6000) workstation in February of 1990. This system was well received in the industry and helped IBM capture a sizable share of the workstation market. The POWER2TM processor goals were to build on the strengths of the original POWER design and to overcome its shortcomings. The POWER and POWER2 systems partition instruction proces...
متن کاملPerformance Measurements of Parallel and Distributed Systems Using On-Chip Counters
Present generation of high performance computing (HPC) systems are often built around commodity processors due to their high performance at a low cost. These processors provide a number of new architectural features, including support for performance measurement through on-chip counters. On-chip performance counters can provide reliable measurement of processor-level activities that are too com...
متن کاملApplication Mapping onto Network-on-Chip using Bypass Channel
Increasing the number of cores integrated on a chip and the problems of system on chips caused to emerge networks on chips. NoCs have features such as scalability and high performance. NoCs architecture provides communication infrastructure and in this way, the blocks were produced that their communication with each other made NoC. Due to increasing number of cores, the placement of the cores i...
متن کاملThe IBM ASIC/SoC
This paper describes the methodology employed by the IBM Microelectronics Division for the design of its Blue Logic applicationspecific integrated circuits (ASICs) and system-on-a-chip (SoC) designs. This methodology is used by both IBM ASIC and SoC designers, as well as OEM customers. A key focus of the IBM ASIC/SoC methodology, outlined in the first section of this paper, is the first-time-ri...
متن کاملApplying recursion to serial and parallel QR factorization leads to better performance
We present new recursive serial and parallel algorithms for QR factorization of an m by n matrix. They improve performance. The recursion leads to an automatic variable blocking, and it also replaces a Level 2 part in a standard block algorithm with Level 3 operations. However, there are significant additional costs for creating and performing the updates, which prohibit the efficient use of th...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1996