MA17502 Datasheet, PDF (12/30 Pages) Dynex Semiconductor – Radiation Hard MIL-STD-1750A Control Unit

MA17502
5.2.10 Special (S)
This addressing mode is applicable to instructions that do
not follow the above formats.
5.3 BENCHMARKING
Table 6a defines the number and type of machine cycles
associated with each MIL-STD-1750A instruction. This
information may be used when benchmarking MAS281
performance. The Digital Avionics Instruction Set (DAIS) mix,
which defines a typical frequency of occurrence for MIL-STD-
1750A instructions, is used here for this purpose.
One problem with the DAlS mix, however, is that it does not
reflect the impact of data dependencies on system
performance. For example, a multiplication in which one
operand is zero may be performed much faster than one with
two non-zero operands. Also, the DAIS mix does not specify
such time consuming operations as normalization and
alignment.
Realistic benchmarks must therefore take both an
instruction mix and data dependencies into account. To this
end, machine cycle counts in Table 6a which have data
dependencies are annotated with either an “a” suffix to reflect
an average number of machine cycles (where each of several
possibilities is equally likely) or with a “wa” suffix to reflect a
weighted average number of machine cycles (where some data
possibilities are more likely than others). Weighted averages
are only applicable to floating-point operations.
Weighted averages provided in Table 6a, based on the
Sweeney (lBM Systems Journal, Vol. 4, No. 1, 1965)
guidelines, take a wide range of data dependencies into
consideration. Normalization and alignment operations are also
represented. Table 5 defines MAS281 throughput, at various
frequencies and wait states, for the DAIS mix using Sweeney
data dependencies.
It should be noted that using the Sweeney guidelines is a
conservative approach to benchmarking. If best case
assumptions are made and such operations as normalization
and alignment are not considered, MAS281 performance
figures are approximately 50% higher than those indicated in
Table 5.
Table 5: Throughput (KIPS)
12/30