|
What are the new design challenges?
The new design challenges are led by the transition from single application
systems to multi application systems. This transition has a significant
impact on the hardware architecture and software development. In particular
local memory sub-systems are replaced with complex memory hierarchy,
local shared bus are replaced with intelligent interconnects (NoC),
single processor IP are replaced by larger multi processors/cores
and interconnect IP. Finally, "single' stand alone software stacks
are being transitioned to multiple and dependent software stacks.
What are the results of the old design methods breaking down?
IP, semiconductor and electronics companies have taken a platform-based
approach to product development. They achieve competitive differentiation
through architecture and software. Hardware architecture design has
traditionally been addressed at the RT level and software development
through the use of development boards. These approaches are breaking
down. Venture Development Corporation (VDC) reported in June 2007
that the top engineering task believed to be the cause of a schedule
delay on current projects is the system architecture design and specification. At
the same time, Gary Smith EDA reported that "this year the cost of
software development for a SoC is passing the cost of the actual IC
design."
What do you mean by the use of virtual hardware platforms by the
larger enterprise and its ecosystem through marketing and business
development functions?
IP and semiconductor companies have traditionally gone to market
using paper datasheets and the availability date of silicon. Unfortunately,
this approach has created a gap between specification and the ability
to validate and use the design by their customers. Using virtual hardware
platforms, marketing and business development functions at IP and
semiconductor companies can provide an executable specification to
their customers as to validate the requirements of their customer,
enable earlier development and establish higher customer satisfaction.
ESL 2.0 is characterized by the end of the traditional go-to-market
approach and the use of executable specifications to establish continuous
and iterative go-to-market strategies.
Is ESL 2.0 a CoWare product?
No, ESL 2.0 is not a CoWare product. ESL 2.0 refers to a second generation
of ESL solutions, which aim to facilitate the design and development
of processor-centric, software-intensive products with complex interconnect
and memory architectures, in a production environment. ESL technologies
and methodologies and their use have reached an inflection point.
Prior to this infection point, the ESL community of suppliers and
users was focused on new methods and new technologies applied to a
limited set of users in a proof of concept or experimental environment.
Traditional development approaches were still being used in production.
The inflection point is caused by the emergence of new design challenges
and the breaking down of old design methods. The requirements beyond
the inflection point are linked with the considerations that need
to be given to the use of ESL by a larger community, in production
and to establish new go-to-market strategies.
Does ESL 2.0 refer to a standard?
No, ESL 2.0 does not refer to a standard, although we should point
out that in the ESL 2.0 framework the use of standards is an important
factor as it provides an underlying infrastructure that helps in addressing
the larger community, facilitating the deployment in production and
the establishment of more effective go-to-market strategies.
What does a "CoWare ESL 2.0 product" mean?
A CoWare ESL 2.0 product is a product whose direction and features
are being implemented to specifically address the objectives of ESL
2.0. CoWare's new release fits this definition by delivering solutions
for the larger community, features that accelerate the roll out and
use in a production environment and the enabling of more effective
go-to-market strategies.
What are some of the technologies and standards relevant to ESL
2.0?
Although the term may suggest a new definition of ESL, it does not
refer to an update of the underlying technology concepts and standards
in use or evolution but rather a maturing of these technologies and
methodologies such that they can be applied by the mainstream professional
engineer in a production ready and productive manner.
The core of ESL 2.0 is virtual hardware platforms, which can be used
by the wider community, exchanged within or across company boundaries
and applied to production designs. Standards relevant to the development
and use of virtual hardware platforms include SystemC, TLM, SPIRIT,
OCP/IP and the standardization of APIs around integration of debugging
and analysis tools for software development.
ESL 2.0 is also marked by the move from proprietary technology to
standards-based technologies and the move from proprietary internally
developed technologies to commercial technologies. These two trends
are reflecting traditional technology adoption stages.
Is there any relation between Web 2.0 and ESL 2.0?
Yes and No. No because obviously Web 2.0 and ESL 2.0 are addressing
different problems in different areas with different technologies.
However it is worth noticing that some of the conceptual ideas behind
ESL 2.0 can be related to web 2.0.
As examples
- Web 2.0 aims to facilitate collaboration and sharing
between users. This concept can be found in ESL 2.0 where
the use of virtual hardware platforms aim at communication
between different engineering design teams within or between
enterprises.
- Web 2.0 is based on the principle of the web as a
platform. In ESL 2.0, the principle is based on the use of
virtual hardware platform and their modeling at a higher level
of abstraction.
- Web 2.0 saw the value moved "up the stack" to services
delivered over the web platform. In ESL 2.0, this move up
is reflected by the use of a virtual hardware platform, not
the simple ability to model at a higher level of abstraction.
- Web 2.0 does not refer to an update to the web technical
specification. In ESL 2.0, there is no attempt to create
new standards but rather build upon the existing mature standards.
What were the characteristics of the pre ESL 2.0 era?
The pre-ESL 2.0 era is often referred to as a proof of concept or
experimental era. Some of the characteristics and challenges of this
era were:
- Its EDA origin had placed ESL often as an abstraction
level above RTL in the mind of the hardware developers.
This view left out the context of the system (hardware, software,
the environment it evolves in).
- Early ESL technologies were either internal to a
company trying to find an answer to a specific design problem
or based on proprietary languages and technologies. It made
the deployment of such technology design task and company
specific, and as a result difficult to deploy on a wider scale.
- ESL brought up a lot of innovation around new enabling
methods and technologies. The emergence of SystemC, TLM and
many other technologies marked a significant step in the initial
evolution of ESL and, as any standard, faced distraction and
competition from proprietary technologies. As a result, the
standards were not initially perceived as standard and confusion
around the objectives of the standards were created.
- In the early years (1997-2004), designs did not require
a significant change in the way they were developed, single
core application, simple interconnect did not break down the
existing design methods and as a result a status quo came
in place that the current way was the right way. This perception
has radically changed over the past two years.
- A significant part of the design is reused from project
to project. Earlier ESL technologies have often assumed
that a new design would start with ESL technologies and have
not incorporated the previous design reuse component.
|
