( ESNUG 538 Item 10 ) ------------------------------------------- [04/25/14]
Subject: Aart, FD-SOI, Nimbic, Pulsic, SystemC, Jasper, BDA, ARM, not Apple
APRIL SHOWERS: Here's the latest news & rumors I've heard about EDA and the
fabs. Feel free to comment/correct/scold anything!
- SCOOP I: Remember how Aart bailed early on DAC last year instead to
go to the GSA Forum in Munich, Germany? Well, it appears Aart has
"suddenly" found he has to be one of 36 speakers at the June 4-5th
IMEC Forum in Brussels, Belgium -- directly opposite DAC'14!!!
This means Aart will only be at DAC for Monday morning because his
14 hour Lufthansa 9127/1016 flight leaves at 6:55 PM Monday night
so he can speak in Brussels early Wednesday morning.
Why is Aart doing this? As explained in ESNUG 506 #4, Synopsys is
doing its best to tear down DAC because DAC helps those pesky little
EDA start-ups get business. "Why can't users attend SNUG only???"
---- ---- ---- ---- ---- ---- ----
- MORE PROOF: Another datapoint which shows how Synopsys wants to
minimize DAC comes from the upcoming DAC'14 exhibit hall floorplan.
Go to dac.com. You'll see the 2400 sq ft Synopsys DAC booth 1133
is exactly the same size as the 2400 sq ft Atrenta DAC booth 1933!
A whopping $1.96 billion Synopsys is the same as $45 million Atrenta?
"Damn it! Why can't those pesky EDA users just go to SNUG only??"
---- ---- ---- ---- ---- ---- ----
- OOPS!: While you're looking at the DAC'14 exhibit floorplan you'll
see the Atoptech booth is 20 feet diagonally across the aisle from
the Synopsys booth. "Talk about the Hatfields vs. the McCoys!"
---- ---- ---- ---- ---- ---- ----
- SCOOP II: It's an open secret that STmicro has been quietly paying
so-called "independent" external bloggers to chat up all the happy
benefits of going to (ST's) FD-SOI instead of to (TSMC's) FinFET.
Problem is the paid-opinion bloggers failed to sway the engineers,
as the not-paid-opinion-and-honest EE Times reporters saw when they
covered the GSA Silicon Summit:
"FD-SOI won't become a significant technology because FinFETs
have sucked so much talent and resources out of the industry.
Alternatives to FinFETs don't show much of a chance given
the amount of coordinated work among chip designers and
makers needed to deliver process improvements."
- Nick Kepler of VLSI Research (EET 04/10/14)
"I'm not hearing the love on FD-SOI nor any clarity in 450 mm
wafers."
- Rick Merritt, EET Editor (EET 04/10/14)
Damn! Doesn't it suck to pay for all that blogger influence and
get nothing back to show for it? (And thank God the EE Times guys
were around to snag the real story!)
---- ---- ---- ---- ---- ---- ----
- REALLY?: There are reports that TSMC, AMD, and GlobalFoundries got
a combined $200 M boost in 2013 sales selling GPU's and specialized
ASIC's dedicated to bitcoin mining.
"The mining machines can fetch in excess of $20,000 each,
placing chip manufacturers such as TSMC, AMD, and
GlobalFoundries in prime position to take advantage."
- Gil Luria of Wedbush Securities (01/04/14)
Even wilder is a GlobalFoundries spokesman, Jason Gorss, confirmed
in a Businessweek article that GF was manufacturing bitcoin mining
chips for "several companies". (A $200 M boost? Really?)
---- ---- ---- ---- ---- ---- ----
- SCOOP III: My spies report that Nimbic, Raul Camposano's cloud-based
chip-package-board 3D field solver is being shopped around to the
"usual EDA acquiring companies". Nimbic tools sound an awful lot
like what Apache does/did -- signal integrity, power integrity, EMI
for chip-package-I/O -- but with one big twist -- Nimbic does it all
inside their proprietary cloud.
---- ---- ---- ---- ---- ---- ----
- SCOOP IV: Remember Pulsic? That small full analog tool maker in the
UK? Word is they've working on a fully automated analog layout tool.
I'm not sure if it's like the old Ciranova Helix or not. It takes in
OpenAccess schematic and PDK, and creates multiple extraction-ready
layouts. Supposedly your constraints are automatically recognized
from the schematic and can be user-edited. No programming needed to
generate the layouts. I'm guessing it'll be at DAC'14 in 5 weeks.
---- ---- ---- ---- ---- ---- ----
- CHANGING OF THE GUARD: DeepChip covered the first chip designer
engineering inklings on the joking possibility of doing actual
chip design from C-source waaaay back in 2001 in SNUG'01 #14.
"C sucks for hardware design. It's not concise nor does
it offer parallelism like Verilog without major revisions.
The overhead provided by SystemC or Cynlib is excruciating
to design in."
- Gregg Lahti of Intel (03/28/01)
That started 13 years ago. Over time the C-to-chip tech discussion
morphed into a very vocal online battle between Brett Cline of Forte
vs. Shawn McCloud of Mentor-then-Calypto, with some George Harper
of Bluespec thrown in.
I'm sad to report that Shawn has left Calypto. He's been replaced
by Mark Milligan. Mark's background is SpringSoft Laker full custom
layout, then some OSCII/SystemC stuff, and then Coware ESL. With
Cadence-Forte joining up against Mentor-Calypto it will be fun to
see how Mark squares up against Brett on the C-design battlefield.
---- ---- ---- ---- ---- ---- ----
- MISSED SCOOP: Last week the white-hot rumor my spies were reporting
was Jasper was about to be acquired. I cried because I had it a
full 6 days before it was officially announced! :(
What held me back was some personal family news distracting me plus
the added rumor that one of Jasper's VC's was shopping the deal
around to SNPS/MENT/ANSYS -- which explain why Jasper sold for such
a hefty price of $170 M minus $24 M Jasper had in cash == $146 M.
(Different spies tell me Jasper made from $28 M to $50 M per year,
so the multiple is from 5X down to 3X. If 5X, wow! If 3X, on par.)
Now the big question is what'll CDNS do with Kathryn Kranen? She's
a force of nature inside EDA and the big reason why Jasper did so
well in formal sales. If she leaves, poor Lip-Bu will have had
paid some serious coin just for some empty tool source code.
The other question is how long will it take for Lip-Bu to see a big
enough return to justify that $146 M purchase price?
---- ---- ---- ---- ---- ---- ----
- FUN FACT:
2013 $M YTY
IC sales change
-------- ------
Intel 48,321 -2%
Samsung 32,520 7%
TSMC 19,859 17%
Micron 14,255 88%
SK Hynix 12,970 43%
Texas Instruments 10,794 -5%
Toshiba 9,868 9%
Renesas 6,405 -14%
STmicroelectronics 5,847 -6%
GlobalFoundries 4,261 6%
UMC 3,959 6%
NXP 3,534 14%
Infineon 3,402 11%
Freescale 3,371 6%
Fujitsu 3,210 -16%
Sharp 1,811 -4%
Sony 1,271 -34%
source: IC Insights McClean Report
That +88% jump in 2013 sales by Micron includes Elpida's $2.5 B in
sales. Bill McClean's report contains 400+ tables and graphs.
---- ---- ---- ---- ---- ---- ----
- SPEAKING OF BDA: The BDA lawyers who successfully got CDNS to settle
in the infamous CDNS vs. BDA SKILL OASIS "contract dispute" fiasco,
O'Melveny & Myers (OMM), is defending Mentor Graphics from Synopsys
in the Synopsys/EVE vs. Mentor patents lawsuit. (That is, I heard
MENT chose OMM because how well OMM did in the CDNS/BDA lawsuit.)
---- ---- ---- ---- ---- ---- ----
- SPEAKING OF BDA (Part II): My spies report as the new blood of the
re-energized MENT custom IC tool group, Ravi Subramanian plans to go
after the SNPS HSPICE/FineSim/CustomSim memory franchise. Why?
"With the MENT's sales reach, AE bandwidth and big company
stability, if Ravi can win 20% of SNPS' $250 M+/year memory
SPICE monopoly, Aart stands to lose $50 M/year while Ravi
4X's his revenue -- it goes from $15 M to 65 M! This would
give MENT a 1 year ROI on its $50 M purchase of BDA."
What this spy reports seems plausible considering that BDA AFS Mega
already has TSMC and Samsung as early adopters -- plus this tool
was announced 12 months ago -- i.e. it's had a chance to harden.
---- ---- ---- ---- ---- ---- ----
- ARM Holdings PLC: WHAT GOES UP
- 04/21/2010 Apple is rumored to acquire ARM - AppleInsider
- 09/24/2010 Oracle is rumored to acquire ARM - Bloomberg
- 10/11/2011 Citigroup upgrades ARM shares - Motley Fool
- 11/28/2012 Intel is rumored to acquire ARM - SeekingAlpa
- 01/09/2013 Deutsche Bank upgrades ARM shares - WSJ
- 04/17/2013 Goldman Sacks upgrades ARM shares - WSJ
- 12/12/2013 Google rumored making ARM servers - Bloomberg
EVENTUALLY COMES DOWN
- 12/12/2013 Don't believe the Google/ARM rumor - SeekingAlpa
- 01/09/2014 Goldman Sacks downgrades ARM shares - WSJ
- 01/09/2014 Deutsche Bank downgrades AMM shares - WSJ
- 01/24/2014 Can ARMH sustain growth in 2014? - SeekingAlpa
- 02/04/2014 Slower smartphone growth hits ARMH - CNN.com
- 02/04/2014 ARM drops 9%, triggers short-sales - MarketWatch
- 04/07/2014 ARM Holdings stock is overvalued - SeekingAlpa
- 04/11/2014 ARM retreats 4.5% to 958.5 pence - Bloomberg
- 04/23/2014 ARM royalty growth slows to a crawl - Fool.com
- 04/23/2014 ARM Holdings hits a speed bump - SeekingAlpha
But don't be surprised when ARMH goes back up again. Simon Segars
isn't the type who'd tolerate a downturn like this for long.
---- ---- ---- ---- ---- ---- ----
- FUN PAST QUOTE:
"As far as R&D teams are concerned; Magma brags they can achieve
better results with a 5 man team than Synopsys can with a 50 man
team, and recently they have been proving their point."
- Gary Smith, EDA analyst (GSE 12/31/11)
---- ---- ---- ---- ---- ---- ----
- LIP-BU'S PUZZLE: Earlier this week during the CDNS 1st quarter 2014
earnings call, Lip-Bu Tan boasted (on record):
"We won a major competitive displacement at a large consumer
electronics company, which decided to adopt the full Cadence
flow, RTL through signoff, for multiple FinFET production
designs that should be in silicon later this year."
- Lip-Bu Tan, CDNS CEO (1Q14 Earnings 04/21/14)
While it's clear that this means CDNS stole away one of Aart's top
tier customers, it's unclear exactly who CDNS stole.
I know it's not Samsung nor Apple. They'd be too big not to have
everyone blabbing. Also, both Apple and Samsung use a mix of tools
from SNPS/CDNS/MENT/ANSYS. Lip-Bu is talking about a displacement.
This means SNPS got booted out, to be replaced by an all-CDNS flow.
I'd love it if anon sources told me who CDNS stole. (Hint! Hint!)
---- ---- ---- ---- ---- ---- ----
Two weeks ago my 81 year old mother was unexpectedly rushed to the hospital
in serious pain. It turns out she has cancer that has metastasized all
over her body. She will be having home-hospice care and I'm flying down to
Florida today to spend some time with her and to help out my dad who's
been taking care of her. My brothers will also be flying in. We simply
don't want her to be alone for even one instant during this time.
We don't know if my mom is going to die in 5 days or in 5 months.
For those of you who've gone through this, I understand now.
- John Cooley
DeepChip.com Holliston, MA
Join
Index
Next->Item
|
|