This is interesting, I recently worked on a hardware team using Forte to
synthesize a processor from SystemC, and helped debug the flow at times.
It seemed like promising technology. - Connie
From: [ John Cooley of DeepChip.com ]
Subject: Cadence to acquire Forte Cynthesizer at a rumored fire sale price
Cadence has announced it is buying Forte Design Systems, the maker of the
Cynthesizer SystemC-to-Verilog-RTL synthesis tool, for unknown terms.
Forte is a 13 year old privately held EDA company with 35 employees. It was
formed when David Evans' Chronology and John Sanguinetti's CynApps merged in
2001. Products are Cynthesizer SystemC-to-Verilog-RTL synthesis, CellMath
Designer datapath synthesis and CellMath datapath IP. (The CellMath line
came from its acquisition of Arithmatica, Ltd. back in 2009.)
In 2012 revenues for Forte were $15.5 M -- with $12.3 M for Cynthizer and
$3.2 M for the CellMath tools/IP. Forte's VC funders are Lucio Lanza,
US Venture Partners, Infinity Capital, and 3i US.
Forte's more notable customers are: Sony, Samsung, Toshiba, Ricoh, Realtek,
LG, Hitachi, Fujitsu, Sanyo -- who are mostly in Japan, with some in Europe.
---- ---- ---- ---- ---- ---- ----
On the rumors side, I've had multiple spies tell me Forte sold to Cadence
for $10 to $12 million -- and one lone spy claims it went for $20 million.
Either way, if any of these spies are right, it's bad news because it means
Forte went to CDNS for around 0.6X to 1.3X annual revenue -- that is, it was
at a "fire sale" price. The rule of thumb is a good acquisition price is
around 3.0X annual revenue or higher, i.e. $45 million or more.
WARNING: I can't tell if these lowball "fire sale" rumors are just rivals
trying to make either one or both of the two companies look bad or not.
(Plus the few people who know the real acquisition price won't talk!)
---- ---- ---- ---- ---- ---- ----
On the staff side, word is almost all of the Forte R&D and support people
will be offered a job at Cadence and only some admin people might be cut.
And, of course, I had to ask if Cadence was going to keep Brett Cline of the
infamous "SystemC chicken suit" bet of DAC 2005 (see ESNUG 441 #12),
around. My spies reported "Brett already has been offered the job.".
Beyond wearing chicken suits, Brett's been a fanatically vocal advocate for
SystemC-based chip design for close to a decade:
Brett gloats over Gary Smith's C-synthesis market data reversal
Brett Cline of Forte warns of CatapultC/Synfora ANSI-C dead ends
Brett defends the honor of SystemC, but no Moshe for Specman "e"
Brett on 5th Grade Math, Architectural Weenies, and SystemC
Brett wrong! User eval confirms Mentor CatapultC does SystemC
Carl Icahn accountability forced MENT to trade away Catapult C
Brett Cline on why "SystemC Won't Go Quietly Into The Night"
Plus given the fact that he speaks fluent Japanese, it would be extremely
surprising if Cadence didn't also pick up Brett in this acquisition.
---- ---- ---- ---- ---- ---- ----
On the tech side, even though Forte Cynthesizer and Cadence C-to-Silicon
both do the same thing -- namely SystemC-to-Verilog-RTL synthesis -- meaning
that Cadence just bought a duplicate tool, EDA analyst Gary Smith said:
"They can target to different markets. The Cadence C-to-Silicon
tool is faster to RTL, whereas the Forte Cynthesizer is a more
serious ESL synthesizer -- it can do better optimization and
WHAT IF analysises than C-to-Silcon can."
- Gary Smith of GSEDA (02/05/14)
The most recent Market Trends Report for C-to-RTL synthesis is:
2012 ESL Synthesis Market: $39.8 million total
Forte Cynthesizer ######################### $12.3 (31%)
Mentor Catapult C ##################### $10.7 (27%)
BlueSpec ############### $7.6 (19%)
Cadence C-to-Silicon ############## $6.8 (17.0%)
SNPS Synphony C (Synfora) ##### $2.4 (6%)
Which means with this Forte acquisition, Cadence owns 31% + 17% == 48% of
the High Level Synthesis (HLS) market. (Not bad for a 20% CAGR niche.)
---- ---- ---- ---- ---- ---- ----
What's colorful here is that Gary Smith has been predicting ESL/HLS since
1997 and has received relentless grief for it for that entire 17 years.
To be fair Gary didn't fully predict ESL/HLS design was going to take off
until 2005. His recent Market Trends data shows that ESL/HLS design has
started to take off in 2011 -- OK, so he was off by 6 years -- with the
big adopters being mostly in Japan and some in Europe.
"Cadence has done a good job with acquisitions lately. They've
picked up a string of IP companies. Lip-Bu is buying into his
company's strengths the same way Wally does.
This Forte acquisition puts Cadence and Mentor Calypto CatapultC
way ahead of Aart's Synfora Synphony C Compiler. The Synopsys
HLS tool pretty much only does algorthmic mapping to RTL, whereas
Forte and CatapultC can do algorthmic mapping, control logic,
datapaths, registers, memory interfaces, etc. -- a whole design.
I don't hear of chips being taped out with Synopsys HLS, but I
know plenty that have been taped out with Mentor or Forte HLS."
- Gary Smith of GSEDA (02/05/14)
On the user side the one big question that customers will have with a tool
"merger" like this is: "which tool dies?" No matter how CNDS answers this,
there will be users who get burned because the tool *they* personally
invested their man-hours and their $$$ in will be gone. (Fun!)
According to the joint Cadence/Forte press release: "The acquisition is
expected to close within 30 days."
- John Cooley
DeepChip.com Holliston, MACadence to acquire Forte Cynthesizer at a rumored fire sale price
'via Blog this'
No comments:
Post a Comment