( ESNUG 563 Item 8 ) -------------------------------------------- [03/09/17]
Subject: Dean and Sawicki on Big Data tapeout predictors, John Lee's Gear
DAC'16 Troublemakers Panel in Austin, TX
Cooley: Dean. You did Big Data last year. You did some little tool.
What the hell was it? Envision? (See ESNUG 550 #5)
Dean: That's good.
Cooley: Your Envision tapeout predictor tool got number one in my
DAC Cheesy Must-See List for that year -- and so did
John Lee's Apache Gear.
Dean: We've continued to get tremendous interest in Envision
throughout the year, and it's been a huge attraction even at
this DAC. The Envision tool was designed to basically provide
analytics on your design process. We all know as engineers, to
improve something, you have to measure the results and then you
can actually figure out how to improve it.
Cooley: Right, that was the Big Data concept.
Dean: Well, it's not always "Big Data", but it is in this particular
case. So if you measure the results of your design process,
where you're spending your time, who's working on what, what's
slow, what's behind, what's ahead, what's coming and is
expected -- you can then actually use Big Data and make your
design process go a lot better and improve it.
That continues to get interest. In fact, we've had a handful
of IC Manage GDP, our Global Design Platform, deals that have
closed basically because the customers wanted to get Envision
and make it work well. We can make Envision work with any
design management system, but obviously it's going to work a
little easier and a little better with ours.
So the high-level big design customers really want some tools,
so they can do "kaisen".
Cooley: Ok.
Sawicki: You know one thing I would say? You know if a buzzword has
gone out of control where it puts that tool (Envision) with
the Gear Apache tool. Gear Apache is a power analysis tool.
This Envision is a design management tool to see how you're
whole thing's coming together. The "Big Data" buzzword is
out of control if you're doing comparing and contrasting
with those two.
Dean: John, you are just out of control. [audience laughter]
Sawicki: "Quelle Surprise!" [audience laughter]
Cooley: Joe, you have Big Data experience. Mentor has a lot of
experience doing big data stuff because you were doing it
before it was "Big Data". You were doing it in litho, fab,
things like that.
Sawicki: Ours is mostly around diagnostics, and we can argue whether
that's big data.
Cooley: You're data mining, you're going through...
Sawicki: We are. But if you look at, I think John Lee did a great
job on that tool, and Gear did a great job on finding an
exit out with Ansys. And Ansys needed it, because RedHawk
was a tool that was getting very, very stale.
But why they call it a "Big Data" tool is God's own mystery
to me.
Amit: Yeah, I think big data, or deep learning, or machine learning,
is more of a methodology within the algorithms to produce the
results, right?
It's tough to say it's a category by itself of tools.
Cooley: OK.
---- ---- ---- ---- ---- ---- ----
Related Articles
Anirudh and Sawicki on IC Compiler II, Innovus, Nitro-SoC, Antun
Amit and Sawicki on Cadence Spectre, Synopsys HSPICE, Mentor BDA
Dean and Sawicki on Big Data tapeout predictors, John Lee's Gear
Raik and Hogan on formal apps, formal engines, and SNPS VC Formal
Anirudh and Sawicki on iffy Apache IR-drop #'s vs. Voltus/Innovus
Join
Index
Next->Item
|
|
