RISC computers

Table 3. SPECmark Ratings for Popular RISC Computers...

PCI has many benefits over other bus types. First, it supports both 64-bit and 32-bit data paths, so it can be used in both 486 and Pentium-based systems. In addition, it is processor independent. The bus communicates with a special bridge circuit that communicates with both the CPU and the bus. This has the benefit of making the bus an almost universal one. PCI buses can be found in PCs, Mac OS-based computers, and RISC computers. The same expansion card will work for all of them you just need a different configuration program for each. 

Knuth, D. E. 2005. The Art of Computer Programming Vol. 1, Fascicle 1. MMIX A RISC Computer for the New Millennium. Reading, MA Addison-Wesley Professional. 

Vendors such as SUN and MIPS introduced lines of computers based on RISC (reduced iastmction set computer) chips. These computers offered significant performance advantages over the CISC (complex iastmction set computer) minicomputers, at least for CPU-bound work. Although there are stiU active debates about what RISC and what CISC are, the essence of RISC is simplicity. 

The philosophy of RISC is that the CPU performs a very small number of very simple operations. Whereas a CISC-based computer might have an iastmction that fetches a number from memory and updates a counter, a RISC system implements such an operation with multiple, but simple, iastmctions. By keeping the CPU simple, it can be more readily scaled up to ever greater speeds. The idea is that, although it might execute many more iastmctions than a CISC machine, it can perform its simple iastmctions so much faster that it gets more work done ia a given time period. 

The RISC versus CISC conundmm has led to the much abused and ultimately extremely confusiag term MIPS (millions of iastmctions per second). Measures of performance that can be more directiy related to a computer s abiUty to perform usehil work should always be preferred over machine MIPS. The throughput of a computer is a function of the number of iastmctions to be executed, the average number of iastmctions that can be executed per clock cycle, and the time per clock cycle. 

Technical RISC work stations have revolutionized computational science in five years. 

Eor all the excitement and enthusiasm of the computer architects, these computers did not meet with great success in the marketplace, and few companies remain as viable entities. One of the primary reasons for their demise seems to have been the simultaneous rise of the RISC work stations, which killed off numerous other architectural initiatives, hence the term killer micros (11). 

A recent victim of the killer micros was Evans and Sutherland s parallel computer development effort, halted ia 1990. Their architecture combiaed a small number of approximately 1-MFLOPS processors iato semi-iadependent functional units. Several of these units could, ia turn, be combiaed to form a processor hierarchy, building up to systems that were expected to cost between 1 and 8 million dollars. With the advent of lO-MFLOPS uniprocessor killer micros, such an architecture became irrelevant and the project was halted. The RISC killer micro could deUver the same level of performance as could the combiaed efforts of 10 of the 1-MFLOPS processors, evea with the unlikely assumptioa that the problem could be perfectiy parallelized across 10 processors. 

Computational Details. Restricted Hartree-Fock (RHF) calculations were carried out using Gaussian 94 (45) and ACES II (46) on an IBM RISC/6000 computer. The gauge independent atomic orbitals (GIAO) method was used for the shielding calculations (47). All second-order many-body perturbation theory (MBPT2, also referred to as MP2) calculations were performed with ACES II (46). 

Tel. 412-621-2050, fax 412-621-3563, e-mail info gaussian.com Gaussian 92. Ab initio molecular orbital calculations (Hartree-Fock, Direct HF, Moller-Plesset, Cl, Reaction Field Theory, electrostatic potential-derived charges, vibrational frequencies, etc.). Input and output of molecular structures in formats of many other molecular modeling systems. Browse for archival storage of computed results. VAX, Cray, DEC-RISC (Ultrix), Fujitsu (UXP/M), Kubota, IBM RS/6000, Multiflow, Silicon Graphics, Sun, and other versions. Gaussian 90 for Convex, FPS-500, Fujitsu (MSP), IBM (VM, MVS), HP-700, and NEC SX/3 systems. 

K. Dowd, High Performance Computing RISC Architectures, Optimization, and Benchmarks, O Reilly and Associates, Sebastopol, CA, 1993. 

---