( ESNUG 588 Item 5 ) ---------------------------------------------- [09/19/19]
Subject: Badru wants PnR users to know Nitro-SoC is alive and well and #3!
CHOICE IS GOOD: A whole bunch of digital PnR users were quite happy to
see that AtopTech ... err... correction ... make that "Avatar" was back
in the EDA game. There were two groups of users commenting on this:
1. the older earlier ATOP users who were generally happy that their
PnR tool of choice could now compete on a technology level instead
of on a "who has the sleaziest lawyers level".
2. and a number of potentional *new* Avatar users who love the idea
of having a viable 3rd choice in PnR over the CDNS vs. SNPS
duopoly -- especially now that MENT Olympus and Nitro-SoC are gone.
(01/30/2019 Edit: Mentor says Olympus and Nitro-SoC are NOT gone.)
- Avatar/Atop is back; but MENT Nitro-SoC is dead (01/30/19)
From: [ Badru Agarwala of Mentor Calypto ]
John, John, John...
Oh, the chaos you cause is so fun, my friend. Our sales department has a
different view of you, however (but we won't go into that).
You sent an email to your many readers that in passing errantly suggested
that both Mentor's Olympus and Nitro PnR tools are "no more." That was a
shocker to a lot of our customers, especially our newest ones, and of
course prompted many sales calls to calm nerves.
But after we called you to set the record straight and you corrected it.
Thank you for doing that! Of course, many people likely read your initial
email and didn't return to it to see this correction, which is why I'm now
writing to you.
I must admit too that we're partly at fault for going a bit radio silent
on the implementation tools' outbound marketing. Here's why:
- As you often point out, PnR is a high cost, very complex and
complicated part of the design process. PnR requires highly
skilled engineers who need to understand all engines'
capabilities, with endless customization possibilities.
For many customers this translates to a small army of mixed
talent from AEs, R&D and customers to get chips implemented.
- This, often brute force approach is unsustainable, and just
doesn't scale where time to market defines success, or for
markets where an army of PnR engineers isn't available.
- Over the past 2 years Siemens has invested heavily in our
group to fully automate the PnR process for the "power
sensitive system market" -- what you'd call IoT, AI, mobile,
vision, space, and automotive markets -- to have killer PPA
right out of the box with our Nitro-SoC.
- The idea is Nitro-SoC makes all the right decisions for the
PnR engineer without needing custom scripts. It automatically
handles all the dependencies between all stages -- design import,
placement, clock tree synthesis, routing, fixing set-up & hold,
DRC, and export -- for successful final design closure.
- Of course, Nitro-SoC isn't 100% totally automated, but with machine
learning (like what Sawicki outlined on the DAC'19 panel) it's 90%
of the way there. Nitro-SoC carefully manages the dependencies
between the different stages, but the PnR user can adjust each
phase for the perfect recipe to close each unique design with
minimal human intervention in the loop.
Since I know you don't believe me until I "name names" and "give hard data",
here's the list of Nitro-SoC 19.1 customers whom I'm allowed to share:
ST, GlobalFoundries, Ostendo, Mythic, Dolphin Integration, Stargazer
Systems, Vishay Intertechnology, and Kalray. (There are more, but I can't
share them.) They're using Nitro-SoC 19.1 to scale their existing PnR
talent pool while lowering the cost of ownership for technology nodes to
16nm for power sensitive IoT, AI, vision, space, automotive chips.
I'm proud to say with machine learning, we've been able to do a lot more
with far fewer PnR engineers. Automating tasks that otherwise would
require significant human expertise (and time) to complete like fixing
routing congesting and reducing DRC violations all while lowering total
power.
In short, Nitro-SoC is alive and well -- and the 3rd choice in PnR is NOT
Avatar, it's Nitro-SoC -- and our goal is to make it so that through our
better PnR-specific machine learning having Nitro-SoC reduces the overall
cost of ownership of PnR.
- Badru Agarwala
Mentor Calypto Fremont, CA
---- ---- ---- ---- ---- ---- ----
Related Articles:
Joe Sawicki on ML, Calibre, Solido, VC funding, and heuristics
ST used Nitro-SoC for full PnR on ultra-low power IoT chips
Avatar/AtopTech's comeback in digital PnR is #4a "Best of 2018"
How Samsung uses Oasys-RTL inside their IC Compiler II flow
Nvidia uses MENT Nitro physical floorplanning for ICC/ICC2
Join
Index
Next->Item
|
|
