| APM
and SMM Advanced
Power Management |
is
an API from Intel and
Microsoft that allows programs and operating systems to
slow down various hardware components, thereby saving power.
System
Management Mode
is the
onchip hardware
and eight additional instructions that CPUs need in order to respond to
APM instructions. |
| CISC
Complex
Instruction Set |
CISC's multiple addressing
modes and many variablelength instructions require additional onchip
microcode to decode and execute.
|
| ClockDoubling |
| This means that the CPU
internally runs at twice the clock speed of the memory
system/motherboard |
| RlSC
Reduced
Instruction Set |
| RISC
stresses simplicity of instructions, addressing, and construction to
gain maximum hardware performance. RISC processors typically require
more sophisticated and optimized compilers than do CISC designs. |
| Integer
Unit |
The
heart
and soul of your CPU.
Intel's Pentium includes two
parallel integer execution units to pump up performance. |
| Pipeline |
| Pipelines
inside CPUs allow several instructions both integer and floating to
progress in
orderly stages through the execution process. Once the pipeline is
full, performance is greatly enhanced unless branches or jumps occur. |
| Process
Technology |
| The
actual
construction used to fabricate chips, including CPUs. Advanced process
technology uses multiple layers
with thinner line sizes measured in
fractions of a micron to pack more circuitry into smaller spaces. |
| Superpipelining |
| Superpipelining
increases the number of pipeline Stages to avoid execution
stalls and keep information flowing
faster in superscalar designs. |
| Superscalar
Architecture |
| A
superscalar architecture provides two pipelines to execute multiple
instructions in parallel, for faster processing and
higher
performance. |
| FPU
floating-point unit
|
A
formal term for the math coprocessors
FPUs perform certain calculations faster than CPUs
because they specialize in floating-point math, whereas CPUs are geared
for integer math. Today, most FPUs are integrated with the CPU
rather
than separately |
| Multi-branch
prediction |
| Boosts
processor performance by predicting with high accuracy the next
instructions needed. |
| Speculative
execution |
| Allows
the
pipelines to continuously execute instructions
following a branch without stalling the pipelines |
| WriteBack
Cache |
An
advanced
caching technique, found in Cyrix's FasCache and Motorola's 68040 CPUs,
among others.
Writeback
caches only write data back to main memory if they have been modified.
They offer increased performance compared with writethrough designs,
such as the one in Intel's 486 chips. However, writethrough caches are
somewhat safer, since all data is immediately written back to memory. |
| Data
dependency Removal |
Provides
instruction results to both pipelines simultaneously so that
neither pipeline is stalled
|
MFLOPS
Million
Floating Point
Operations Per Second
|
| The
raw speed of a computer's floatingpoint processor |
| Math Co-Processor |
The
math coprocessor is a special purpose processor, that works together
with a general purpose or main processor (CPU).
While general purpose processors require software routines to
execute floating point and large integer calculations, math
coprocessors execute these functions in hardware |
| MHZ
MegaHertz |
It is a unit
of measurement named for
Heinrich R. Hertz,
a German
physicist, who first detected electromagnetic waves in 1883.
1 Hertz (Hz) is equal to 1
cycle
per second. Consequently MegaHertz is a unit of measurement
for
indicating the frequency of one million electrical vibration cycles per
second. The speed of processors is measured in MHz
|
