( ESNUG 426 Item 1 ) -------------------------------------------- [03/31/04]
Subject: Bernie Also Questions Dataquest FPGA Synthesis Marketshare Data
> What's unbelievable here is that, apparently while nobody was watching,
> Mentor/Exemplar somehow managed to catch up with Synplicity??!! That
> is, Mentor sales remained steady while Synplicity sales dropped 31% ??!!
> And Synopsys FPGA synthesis market share, although small, somehow manged
> to almost double at the same time, too????
>
> None of these "facts" pass the "smell test"!!!
>
> - from http://www.deepchip.com/gadfly/gad031204.html
From: [ Bernie Aronson, Synplicity CEO ]
Hi, John,
Thanks for helping us clear this one up. We were also puzzled by the 2002
FPGA synthesis market share numbers reported by Gartner and as a result
looked into them a bit further. Here are some facts followed by some
conclusions based upon more recent data from 2003.
It turns out, in addition to other items, Synplicity's FPGA synthesis OEM
revenue was not included in Gartner's report for the years 2001 & 2002.
This revenue comes from our FPGA partners (Actel, Lattice & QuickLogic) for
rights to provide Synplify to their customers. Gartner now knows this and
said they will publish corrected numbers in their next (2003) report. If
we properly include these numbers, the market share picture looks like this
(below) and may be confirmed with Nancy Wu of Gartner.
The Corrected Dataquest/Gartner FPGA Synthesis Market Share Numbers
Synplicity 2001 ############################# 57%
Synplicity 2002 ######################## 47%
Mentor/Exemplar 2001 ################## 35%
Mentor/Exemplar 2002 ##################### 41%
Synopsys 2001 #### 7%
Synopsys 2002 ###### 11%
While improved, we still do not feel these numbers fairly represent actual
customer adoption and usage of FPGA synthesis tools. Other important
factors that may be influencing the market share numbers are as follows:
a) Synplicity does not have FAM (Flexible Access Model) or similar license
bundling models so our reported numbers fairly represent actual
customer need and use. Mentor & Synopsys can include licenses (like
FPGA synthesis) as part of bigger FAM/bundle type deals and can
therefore allocate revenue to products whether they are actually
used or not.
b) Mentor released Precision RTL in 2002 which we expect was included in
many FAM/bundle deals.
c) Synplicity separates maintenance and product revenue for time-based
licenses. We are not sure if Mentor and/or Synopsys separate them.
If they don't, it would show higher reported license revenue, which
would not be indicative of higher market share.
d) 2002 was the last year for revenue for Mentor from the Altera OEM of
Leonardo Spectrum, which we believe was in the $4 - 5M per year range.
e) Based upon excerpts taken directly from Mentor's Analyst calls in 2003
(see below) we can make some additional observations regarding Mentor's
FPGA business for 2003.
- Q1 - No mention was made regarding FPGA bookings, let alone FPGA
synthesis bookings.
- Q2 - Overall FPGA (ModelSim, HDL Designer + synthesis) was flat
and FPGA synthesis was down due to termination of the Altera OEM
arrangement.
- Q3 - FPGA synthesis bookings were down 50% compared to the prior
year, all due to termination of the Altera OEM. Assuming the OEM
generated $5M / year, Mentor's non-OEM FPGA synthesis bookings
in Q3 was ~$1.25M. Synplicity's quarterly FPGA license revenue
for the same Q3 was nearly 5X that number.
- Q4 - Mentor's overall FPGA bookings were down 35%, with FPGA
synthesis down 40% to 85%. Synplicity's FPGA synthesis license
revenue remained steady year-over-year for that same period.
In summary, it is clear that Mentor's FPGA synthesis business was down
sharply in 2003, and we would expect our market share to increase sharply
for that year once reported by Gartner.
Another journalist perspective (FPGA Journal) is at:
http://www.fpgajournal.com/articles/windmills.htm
and it provides additional information on this topic.
Synplicity continues to invest heavily in the FPGA segment and all of the
data tells us we are the market leader. We strongly believe the products
we have developed provide excellent value to FPGA designers both at the
high-performance end and high-volume end of the market.
- Bernie Aronson, CEO
Synplicity Sunnyvale, CA
P.S. Excerpts taken directly from Mentor's 2003 Analyst Call Transcripts
Q2 2003:
No specific mention of the FPGA business
Q2 2003:
Product booking climbed 30%, while consulting dropped 15. System-on-chip
design and verification, which includes our emulation product line or
hardware-software co-verification product line designed for test and
embedded software was the most impressive, up 70% Deep Submicron
including the Calibre and the advanced MS Simulation family was up 35%
System Design climbed 30%, while FPGA was flat as a synthesis portion
of our ALTERA OEM contract lapsed.
Q3 2003:
This quarter strength was seen in board products up in aggregate 35%,
cabling, as I said before, up 50%, and DFT, more than doubling, Calibre
RET up 35%, and embedded software up 20%. Particularly weak were such
systems related products, as FPGA synthesis, emulation and hardware
software applications all down 50%, but in the case of FPGA synthesis,
I remind you that this was the decrease was all due to the lapse earlier
this year of the ALTERA OEM contract.
Q4 2003:
System-related products on the other hand were generally weak, emulation
and FPGA synthesis all declined ranging from 40-85%. Board design
compared to exceptional Q4 '02 was also relatively weak. By product
line, systems designs measured against a very strong Q4 02 that was then
up 85%, was down about 40. SOC design and verification was down 45%,
FPGA down 35%. The deep submicron, which includes our Calibre as well
as analog/mixed signal families was up 170%. I don't have the details
at my finger tips right now, Rich, but we would have had ModelSim, we
traditionally classify within our FPGA product line. That would have
been up substantially I think somewhere around 25% and we would have
had very strong bookings in North America and PacRim would have been
quite strong as well.
---- ---- ---- ---- ---- ---- ----
From: Kevin Morris <kevin.morris=user domain=techfocusmedia spot balm>
Hi John,
I really enjoyed your article on Dataquest numbers for FPGA synthesis!
So much, in fact, that I wrote my own "Tilting at Tech Market Windmills"
for FPGA Journal. If you're interested, the article is here:
http://www.fpgajournal.com/articles/windmills.htm
You're mentioned. Not in vain, I hope.
- Kevin Morris
Editor - FPGA Journal
---- ---- ---- ---- ---- ---- ----
From: Jeff Garrison <jeff=user domain=synplicity spot balm>
Hi John,
Bernie Aronson already responded to you with some information regarding
FPGA synthesis market share and Synplicity's position on it, but I want to
point out a recent survey prepared by FPGA Journal which contains much more
current data than Gartner's 2002 numbers. A summary of this survey is on
the web at: http://www.fpgajournal.com/articles/windmills.htm
In this survey of FPGA design projects performed between August 2003 and
December 2003, 310 users responded to the question of what FPGA synthesis
tool they use. The results were as follows:
################## 35% use Xilinx XST
################ 32% use Synplicity Synplify / Synplify Pro
######### 17% use Mentor Leonardo Spectrum
####### 14% use Altera's Quartus synthesis
# 2% use Synopsys FPGA Compiler II
If we remove the free tools provided by FPGA Vendors from the equation and
compare EDA vendor's share (same as Gartner) we get 158 (310 - 108 Xilinx -
44 Altera) survey respondents using EDA synthesis tools. The market share
then looks like:
Synplicity ############################### 62.7%
Mentor/Exemplar ################# 33.3%
Synopsys ## 3.9%
I suspect this is a closer representation of current FPGA synthesis market
share.
- Jeff Garrison
Synplicity Sunnyvale, CA
|
|
