IBM RS/6000 - Manual (Page 64)

2.7.1 SPEC95 Software Benchmark
SPEC95 is a forward step in the performance measurement of the core of a
system. It covers the CPU, caches, memory, and compiler. The programs and
datasets which make up the suite cannot "fit" into cache, making the benchmark
more representative of real workloads. SPEC has also standardized the compiler
settings so that the results for "base" measurements are more comparable between
suppliers.
SPEC95 is a software benchmark product produced by the Standard Performance
Evaluation Corp. (SPEC), a non-profit group of computer vendors, systems
integrators, universities, research organizations, publishers and consultants
throughout the world. It was designed to provide measures of performance for
comparing compute-intensive workloads on different computer systems.
SPEC95 contains two suites of benchmarks:
CINT95 for measuring and comparing compute-intensive integer performance
CFP95 for measuring and comparing compute-intensive floating point
performance
The two groups of programs are referred to as component-level benchmark suites
as they test the core of the system, the CPU, caches, memory, and compiler, but
not the I/O sub-system.
One of the goals of SPEC95 is increased portability; the current offering from SPEC
is for UNIX only although the member companies have indicated that the
benchmark programs are portable to various flavors of UNIX, Windows NT and
Open VMS.
SPEC95 introduces a new reference platform against which other systems are
measured, changing from the out-dated VAX 11/780 to a SPARCstation 10/40 with
64MB memory but without Level 2 cache. This is more representative of the types
of systems being sold today, but it is also a machine that will beat few, if any, of
the machines being benchmarked.
The rules have also changed; each benchmark must be run a minimum of three
times to get a valid result, with the median time for all runs being used as the
benchmark time.
The SPEC base metrics (for example, SPECint_base95) are required for all
reported results and have set guidelines for compilation (for example, the same 4
flags must be used in the same order for all benchmarks). The non-base metrics
(for example, SPECint95) are optional and have less strict requirements (for
example, different compiler options may be used on each benchmark).
There are several different ways to measure computer performance. One way is to
measure how fast the computer completes a single task; this is a speed measure.
Another way is to measure how many tasks a computer can accomplish in a certain
amount of time; this is called a throughput, capacity or rate measure. The SPEC
speed metrics (for example, SPECint95) are used for comparing the ability of a
computer to complete single tasks. The
SPEC rate metrics (for example,
SPECint_rate95) measure the throughput or rate of a machine carrying out a
number of tasks.
40 Introduction to PCI-Based RS/6000 Servers