Read-modify-write operations are typically not atomic at an instruction
level on a RISC machine.
byte granularity: A property of memory systems in
which adjacent bytes can be written concurrently and independently by
different processes or processors.
CISC: See complex instruction set computer.
compatibility: The ability of programs written for one
type of computer system (such as OpenVMS VAX) to execute on another
type of system (such as OpenVMS Alpha).
complex instruction set computer (CISC): A computer
that has individual instructions that perform complex operations,
including complex operations performed directly on locations in memory.
Examples of such operations include instructions that do multibyte data
moves or substring searches. CISC computers are typically contrasted
with RISC (reduced instruction set computer)
computers.
concurrency: Simultaneous operations by multiple
agents on a shared object.
cross development: The process of creating software
using tools running on one system, but targeted for another type of
system; for example, creating code for Alpha systems using tools
running on a VAX system.
granularity: A characteristic of storage systems that
defines the amount of data that can be read or written with a single
instruction, or read or written independently. VAX systems have byte or
multibyte granularities while disk systems typically have 512-byte or
greater granularities.
image information file (IIF): An ASCII file that
contains information about the interface between VAX images. VEST uses
IIFs to resolve references to other images and to generate the
appropriate linkages.
image section: A group of program sections with the
same attributes (such as read-only access, read/write access, absolute,
relocatable, and so on) that is the unit of virtual memory allocation
for an image.
interlocked instruction: An instruction that performs
some action in a way that guarantees the complete result as a single,
uninterruptible operation in a multiprocessing environment. Since other
potentially conflicting operations can be blocked while the interlocked
instruction completes, interlocked instructions can have a negative
performance impact.
jacket routine: A procedure that converts procedure
calls from one calling standard to another; for example, calls between
translated VAX images, which use the VAX calling standard, and native
Alpha images, which use the Alpha calling standard.
load/store architecture: A machine architecture in
which data items are first loaded into a processor register, then
operated on, and finally stored back to memory. No operations on memory
other than load and store are provided by the instruction set.
longword: Four contiguous bytes (32 bits) starting on
any addressable byte boundary. Bits are numbered from right to left, 0
to 31. The address of the longword is the address of the byte
containing the low-order bit (bit 0). A longword is naturally
aligned if its address is evenly divisible by 4.
multiple instruction issue: Issuing more than one
instruction during a single clock cycle.
natural alignment: Data storage in memory such that
the address of the data is evenly divisible by the size of the data in
bytes. For example, a naturally aligned longword has an address that is
evenly divisible by 4, and a naturally aligned quadword has an address
that is evenly divisible by 8. A structure is naturally aligned when
all its members are naturally aligned.
page size: The number of bytes that a system's
hardware treats as a unit for address mapping, sharing, protection, and
movement to and from secondary storage.
pagelet: A 512-byte unit of memory in an Alpha
environment. On Alpha systems, certain DCL and utility commands, system
services, and system routines accept as input or provide as output
memory requirements and quotas in terms of pagelets. Although this
allows the external interfaces of these components to be compatible
with those of VAX systems, OpenVMS Alpha internally manages memory only
in even multiples of the CPU memory page size.
PALcode: See privileged architecture library.
privileged architecture library (PAL): A library of
callable routines for performing instructions unique to a particular
operating system. Special instructions call the routines, which must
run without interruption.
processor status (PS): On Alpha systems, a privileged
processor register consisting of a quadword of information
including the current access mode, the current interrupt priority level
(IPL), the stack alignment, and several reserved fields.
processor status longword (PSL): On VAX systems, a
privileged processor register consisting of a word of privileged
processor status and the processor status word itself. The
privileged processor status information includes the current interrupt
priority level (IPL), the previous access mode, the current access
mode, the interrupt stack bit, the trace trap pending bit, and the
compatibility mode bit.
processor status word (PSW): On VAX systems, the
low-order word of the processor status longword. Processor
status information includes the condition codes (carry, overflow, 0,
negative), the arithmetic trap enable bits (integer overflow, decimal
overflow, floating underflow), and the trace enable bit.
program counter (PC): That portion of the CPU that
contains the virtual address of the next instruction to be executed.
Most current CPUs implement the program counter as a register. This
register is visible to the programmer through the instruction set.
quadword: Four contiguous words (64 bits) starting on
any addressable byte boundary. Bits are numbered from right to left, 0
to 63. The address of a quadword is the address of the word containing
the low-order bit (bit 0). A quadword is naturally aligned if
its address is evenly divisible by 8.
quadword granularity: A property of memory systems in
which adjacent quadwords can be written concurrently and
independently by different processes or processors.
read-modify-write operation: A hardware operation that
involves the reading, modifying, and writing of a piece of data in main
memory as a single, uninterruptible operation.
read-write ordering: The order in which memory on one
CPU becomes visible to an execution agent (a different CPU or device
within a tightly coupled system).
reduced instruction set computer (RISC): A computer
that has an instruction set reduced in complexity, but not necessarily
in the number of instructions. RISC architectures typically require
more instructions than CISC architectures to perform a given
operation, because an individual instruction performs less work than a
CISC instruction.
RISC: See reduced instruction set computer.
synchronization: A method of controlling access to
some shared resource so that predictable, well-defined results are
obtained when operating in a multiprocessing environment or in a
uniprocessing environment using shared data.
translated code: The native Alpha object code in a translated image. Translated code includes:
translated image: An Alpha executable or shareable
image created by translation of the object code of a VAX
image. The translated image, which is functionally equivalent to the
VAX image from which it was translated, includes both translated
code and the original image. See VAX Environment Software
Translator.
Translated Image Environment (TIE): A native Alpha
shareable image that supports the execution of translated
images. The TIE processes all interactions with the native Alpha
system and provides an environment similar to OpenVMS VAX for the
translated image by managing VAX state; by emulating VAX features such
as exception processing, AST delivery, and complex VAX instructions;
and by interpreting untranslated VAX instructions.
translation: The process of converting a VAX binary
image to an Alpha image that runs with the assistance of the TIE on an
Alpha system. Translation is a static process that converts as much VAX
code as possible to native Alpha instructions. The TIE interprets any
untranslated VAX code at run time.
VEST: See VAX Environment Software Translator.
VAX Environment Software Translator (VEST): A software
migration tool that performs the translation of VAX executable
and shareable images into translated images that run on Alpha systems.
See translated image.
word granularity: A property of memory systems in
which adjacent words can be written concurrently and independently by
different processes or processors.
writable global section: A data structure (for example, FORTRAN global common) or shareable image section potentially available to all processes in the system for use in communicating between processes.
6459P016.HTM OSSG Documentation 22-NOV-1996 13:07:29.38
Copyright © Digital Equipment Corporation 1996. All Rights Reserved.