RUSSIAN came
Summarized by Gris
We recently came across a very
interesting information, which we would
like to share with you, dear readers. It is a pity that we can
not you provide the information fully (yes you
it is not necessary) due to its large ob'ema
(Approximately 300Kbayt). For this reason,
give you a distillation of this document.
Just want to say that the text set out below provides a
technical description of the new processor and its
modifications developed by Russian engineers, who bears the
name "Elbrus". To forget that the whole text is
on behalf of developers.
Project Description
The aim of this project is to develop new
postRISC-native-architecture Computer, the corresponding
crystal microprocessor, workstation-based this crystal,
optimization compiler, as well as the binary system
compilation, which provides a highly efficient and 100%
reliable performance binary code for the x86 processor
architecture firm Intel. Maintained a high level of security
programming.
PRODUCTIVITY
Proposed in this design method allows the use of a broad
command-level parallelism of operations for improve
performance. Achieved theoretical limit of the speed of the
logical for a given amount of equipment. With the introduction
of explicit parallelism in the machine language of most of the
hard work Planning computing transferred from
Hardware optimizing compiler.
This allows a high logic
speed with the ability to implement a more sophisticated
optimization and at the same time, simplifies hardware and
allows to achieve high clock frequencies. Speed
microprocessor is well known testing programs SPEC
performance is 135 SPECint95 and 350 SPECfp95 at a clock
frequency equal to 1200 MHz on the technology
0.18 microns. Thus, we achieve an almost two-fold advantage in
speed over the fastest microprocessor Western next decade.
This microprocessor is actually The second implementation of
this new approach, developed by a team since 1986
First implementation was "Elbrus-3 -
extremely fast computer,
made in 1991 It was 2 times faster than the best performance
while super computer Cray Y-MP, despite the extremely imperfect
domestic manufacturing technology crystals.
COMPATIBILITY
The project achieves the full
compatibility with the architectural platform
x86 on a binary level. The project is fully
legally independent from the intellectual
property due to Intel's new methodology binary compilation
developed by the team.
SECURITY SOFTWARE
Secure programming, ease of debugging, ease
build very large and reliable software systems from individual
parts developed by many different programmers, with the
development of computers are becoming
greater significance. One reason for this
The situation is in wide use
Internet and the growing size of a typical
software applications. Secure programming to software
developers in order to be able to successfully debug
large applications. It is necessary, and users of computers
that are not afraid to use the Internet and free software to
read from her other authors. Today, the Java language from Sun
Microsystems suggested that high levels of security,
with the result that specialists are well
adopted this language to adopt.
However, the developers have chosen another
approach. This microprocessor supports data types in hardware
and verify their correct processing during implementation. This
is only possible resolution of conflicts between security on
the one hand, and the versatility and efficiency, on the other.
We have 20 years of experience developing and supporting three
generations of computers based on this philosophy. As a result
of this technology developed and tested
In practice, our staff, the proposed microprocessor: improves
the safety of existing languages such as Java, Fortran, etc.
and, more importantly, adds a high level of security
very widely spoken languages with
and C + +. Due to hardware support
This microprocessor provides a high
level of security.
INTERNET SECURITY
Based on our 20 years of experience in
development of several generations of computing systems
"Elbrus" with highly reliable and provides the security context
protection we can offer a new improvement of this architecture
to support secure programming and complement existing operating
systems and file systems means working in Internet. This
project will improve the organization of the Internet and will
enable secure sharing of remote files in a secure and safe mode.
Small crystals
With sufficient funding,
addition to the large crystal can be
create one less option for embedded and network computing.
Benefits our development here is the most
high performance for a given
crystal size and ease of debugging
complex codes, programs running in
real time. This is achieved through a high level
safety and strict control of hardware supported types of teams
in run time.
Multimedia applications
Our architecture is a broad team is well-suited for multimedia
extensions set of commands. Currently, several
firms, such as Chromatic (Mpact),
Philips (TriMedia), Microunity, TI,
announced a powerful specialized
crystal for multimedia applications. Along with this, all the
major firms developing a universal
microprocessors, have expanded their collections
teams with special teams for multimedia applications. Our
purpose microprocessor and a small crystal
have a highly efficient hardware support for multimedia and 3D
graphics.
CURRENT STATUS
Developed:
- Verilog description at RTL
equipment with a preliminary estimate
possible clock frequency (it is now
is in the process of detailed logical verification);
- Optimizing compiler, performing all the SPEC packages and
other great application (not finished all the optimization);
- Exact (performed at the clock level) device simulation
apparatus Perform more applications;
- Compile binary program, performing large applications that
are compiled from a binary package x86 (not all of full
optimization). In the binary compilation performed by all
versions of Windows operating systems company Microsoft.
TIME FRAME
POSSIBLE END
DRAFT
After the start of funding necessary
have 3 years to complete hardware
and software and an additional one year to complete verification
system.
MARKET CONSIDERATIONS
In the near future the company will release Intel
microprocessor Mesrced, based on the new-postRISC-native
computer architecture (IA-64) c use large team with explicit
parallelism EPIC. "If Intel can install IA64 as a software
standard She set a new level
performance, which is unlikely to be comparable with the
performance of pure IA-32 processor, released its
competitors, or even most
processors based on RISC, which would make
Intel the leader in performance ...
Intel will have a strong portfolio of patents that protect the
IA-64. Since most if not all, current
licensing agreements
by Intel, will expire by 2000., other
producers would be nearly impossible
create a consistent crystals without a license from Intel. In
this situation, Intel prefers to maintain its position
monopolist. "(MicroDesign Resources
( 1996, p.12.)). In this situation,
Our project is very significant
advantages. According to the indicators
Performance, published in the
same report (p. 6), our processor will be
3-fold faster than Merced. On binary codes Intel it will be
more rapid through a binary compilation. Our
project is not based on IA-64 architecture.
It is protected by our own patents, so there is no legal
problems. Our microprocessor allows
security when using the widely
common languages such as C and
C + +, and conventional operating systems such as Linux.
That's how things stand right now with our
processors. Now to you,
dear reader, feel the full
issues of the day, we offer you a slight
Processor Merced, which was
taken from the official website of the company Intel.
IA-64 architecture:
new features
Intel Architecture
Architecture (IA)
Market for computer systems actively
diversifying and becoming more
segmented in the whole range of
mobile PC to the server. In an effort to
create the best solutions for each
market segment, Intel is expanding its series of products.
Processors IA-32 architecture will continue to provide the
opportunity to boost performance in the mobile sector
and desktop PCs, servers and desktops
stations. At the same time, processors with
IA-64 will extend the scope of architecture
Intel's high-performance sector
servers and workstations. The first processor with IA-64 will
Processor Merced. Like other processors with IA-64 processor
Merced will have a full hardware
compatibility with IA-32 processors on
level binary code, which guarantees
Protect your investment in end-user
software.
IA-64 architecture combines
a unique range of innovative technologies, including explicit
parallelism, branch prediction, speculative execution, etc. A
high degree of scalability to satisfy the
requirements of different market segments
servers and workstations. The most important
principle in developing the IA-64 was a complete binary
compatibility with IA-32 architecture at the hardware level,
enabling end users to freely use software tools
designed to IA-32. Thus,
IA-64 architecture is different scalability, the highest level
performance and full compatibility with existing software for
IA-32 architecture.
