Paul Lussier wrote:
> Synopsys and Cadence are by far the 2 leaders. Specman is an up and comer.
> They're no where as larges as these other 2, but they're getting a lot of
> attention for their product (Verisity) (I can't really give you details, since
> I'm not in EDA, and I don't know much about the tools themselves, just what
> they are and who provides them).
Just to set the record straight; The company is Verisity, and the product is
Specman.
As Verification tools go, they are starting to get the right idea. Specman
(really the 'e' language itself that Specman provides) is a bit more to
my liking than my last project, which was written in Vera (The Synopsys
entry into the field of Verification languages), but they all lack a little
something when you are used to having C++ linked into the simulator PLI.
Let them keep evolving, and someday we'll get what we want.
As for the Linux / EDA connection, we've all been waiting for a while for it
to come together. I work for a startup, so you can imagine that we don't have
bundles of money to spend on our compute farm. My current project has tool
requirements from Cadence (NC Verilog), Denali (memory models), and Verisity
(Specman), to name the major (licensed) tool vendors. All of these tools come
in Linux flavors. I'm in the process of augmenting our compute farm (a handful
of Sparc boxen) with fire breathing Linux boxes. When it comes to bang for the
buck, it's really hard to beat the X86 boxes running Linux. The only holdup
to "doing things right" at the moment is Rational's Atria / ClearCase. They
are promising to support "Tier II" platforms (Linux being one of them) in
the very near future (months). In the mean time, I'll get the farm up and
running with a snapshot and file system export methodology for now.
If anyone is interested, I'll let you know how it goes.
NT based tools? Not for the ASIC I'm responsible for.
Enjoy,
-Carl
**********************************************************
To unsubscribe from this list, send mail to
[EMAIL PROTECTED] with the following text in the
*body* (*not* the subject line) of the letter:
unsubscribe gnhlug
**********************************************************
