( ESNUG 426 Item 9 ) -------------------------------------------- [03/31/04]

Subject: Ten Letters On The Free Xilinx XST & Altera Quartus Tools

> First off, overall FPGA synthesis market *shrank* 15% in 2002 ??!!  Huh?
>
> No only that, FPGAs have been getting bigger every year -- thus driving
> a stronger demand for FPGA synthesis tools.  Maybe you can do a 10,000
> gate design using schematic capture, but God help you if you're not using
> simulation and synthesis for a 100,000 gate design.
>
>     - from http://www.deepchip.com/gadfly/gad031204.html


From: Goran Bilski <goran.bilski=user  domain=xilinx spot balm>

Hi John,

Could it be that Altera was starting to ship a lite version of Leonardo 
with their tool set?

    - Goran Bilski
      Xilinx

         ----    ----    ----    ----    ----    ----   ----

From: Arun Chaddha <achaddha=user  domain=tenesix spot balm>            

Hello John,

My 2 cents worth -- no one seems to talk about the low-cost synthesis
tools from Xilinx and Altera.  Being a low-cost company we have been
using Altera tools for about 2 years (cannot afford Synplicity or Mentor
tools).  Designed a STS-48 pointer processor with cross-connect in the
lowest speed EPS125.  FPGA is 90% filled and tool completes in a few
hours (around 3-4 hrs).

When we tested Synplicity in June 02, the support for Stratix was poor
and it seemed that Altera tools were a few months ahead in terms of
optimizing for the then new Stratix part.

    - Arun Chaddha
      Tenesix

         ----    ----    ----    ----    ----    ----   ----

From: [ Kenny from Southpark ]

John,

You missed one point, and two new important players in the market.

As of 2002, Altera Quartus and Xilinx ISE have built in VHDL and Verilog
synthesis tools.  This is what is eating into the profits of Mentor and
Synplicity.

The Xilinx tool, XST (Xilinx Synthesis Tool) only takes very basic VHDL
code, but it has many of the fitting algorithms supplied by Xilinx to
MGC and Synplicity built into it.  As long as you are careful how you write
your code, it works.

The Altera tool is built into Quartus II.  I've played with it enough to
know that it can take a bigger VHDL language set that XST can.  Altera is
really pushing it these days.

Why buy a separate synthesis tool if one comes for free with your FPGA
software?  As long as you are not doing edge of technology designs stick
with what you get for free.

Me?  I'm sticking with Synplicity.  The pipline retiming of Synplify Pro
is something my designers have come to depend on.

    - [ Kenny from Southpark ]

         ----    ----    ----    ----    ----    ----   ----

From: [ An Altera Employee ]

John,

Please let me remain anonymous, but remember that in this timeframe Altera
was OEM-ing Exemplar.  That may well have resulted in increased sales for
Exemplar.

Also note that Altera now provide world-class VHDL and Verilog synthesis
integrated in the Quartus II software - which may well explain why people
aren't spending so much money of FPGA synthesis tools.

    - [ An Altera Employee ]

         ----    ----    ----    ----    ----    ----   ----

From: Kevin Hubbard <kevin.hubbard=user  domain=siemens spot balm>

John, part of this equation is missing. 

Free Synthesis tools are getting better.

I've been using XST (Xilinx Synthesis Techonology) for a year now for
multiple quarter-million gate 160MHz FPGA designs (XC2VP7s).  Have I been
using XST because I'm poor?  No.  I have a Synplify License.  I even have
a DC license I use for my ASIC designs.

I use XST because it works.  Sure its simple, but it does its job well.
FPGA synthesis is not rocket science.  There are no buffers to resize, etc.
Quality Mapping and Place and Route really seem to even out the end
results for performance.

When I compared my designs using Synplify against XST, I saw marginal
performance and area improvement.

Not having to deal with license issues is just a huge bonus. 

Running Linux on my laptop, I can tweak my design on the plane.  Now that's
a nice little niche synthesis market.

Remember how C compiler companies (Watcom, Borland, Micro$oft) used to
boast about performance optimization?  Remember when compilers actually
cost money?  Synplicity needs to watch out and expand outside of FPGA
synthesis (i.e. own the Structured-ASIC arena), or go the way of Borland,
Watcom, etc.

    - Kevin Hubbard
      Siemens
  
         ----    ----    ----    ----    ----    ----   ----

From: Gary Seely <gary=user  domain=aldec spot balm>

John,

I'm just a sales guy at Aldec, but the word from our customers and prospects
in the FPGA world is that the free FPGA vendor tools are making a huge dent
in the synthesis market (as well as other EDA areas).  Many designers are
getting by using Xilinx XST where they would have bought Synplicity or
Exemplar in the past.  There was also a very aggressive marketing campaign
for Precision as a replacement for Leonardo during this time frame. 

    - Gary Seely
      Aldec

         ----    ----    ----    ----    ----    ----   ----

From: [ An Anon Engineer ]

One explanation to the declining sales could be that FPGA users are
migrating towards the zero-cost tools offered by FPGA vendors.  Could there
be a place for Scenario 3?

The next logical step would be for an SCO-like scenario where some of the
"Synthesis Gurus" sue FPGA **vendors** for violating FPGA Synthesis
Intellectual property...  Now wouldn't that be a hoot...

    - [ An Anon Engineer ]

         ----    ----    ----    ----    ----    ----   ----

From: Ramesh Narayanaswamy <ramesh=user  domain=tharas spot balm>
             
Hi John,

While there might be more FPGA starts, it may not reflect in revenue
numbers because customers with tight budgets might use the nearly free
Xilinx/Altera tools.  Did this increase in 2002?

    - Ramesh Narayanaswamy
      Tharas

         ----    ----    ----    ----    ----    ----   ----

From: Herman Schmit <herman=user  domain=everychip spot balm>

John,

Aren't these numbers an illustration that more people are using the 
integrated synthesis with Xilinx and Altera tools?  There is no reasonable
way to measure the revenue of XST and Quartus, because they are largely
subsidized by chip sales.

    - Herman Schmit
      Everychip, Inc.

         ----    ----    ----    ----    ----    ----   ----

From: Michael Whalen <mwhalen=user  domain=intimesw spot balm>

John,

Here's an auditable approach to the numbers derived from Synplicity's 2002
Annual Report:

    Total License revenue 2002: $25,830,000
    Total License revenue 2001: $32,126,000

And from the same report:

  "ASIC Solutions

   Our software products for use with ASICs represented approximately
   13%, 6% and 7% of our total bookings in 2002, 2001 and 2000,
   respectively."

That would put FPGA license revenue at 87% of $25.830 M for 2002 and 94% of
$32.126 M or

    FPGA License revenue 2002: $22.472 M
    FPGA License revenue 2001: $30.198 M

These appear to be within 10% of the Gartner numbers's you reported, so
it's pretty close:

    2002 Synplicity  ############################ $28.2 (54%)
    2001 Synplicity  ################### $19.4 (44%)

Black and white, from the SEC docs, a decline of 26% from 2001 to 2002.

As for Mentor - not so easy.  They don't break things out in their 2002
Annual Report, so it's hard to find an auditable source to their $19.0 M
for 2002.

    - Mike Whalen
      InTime Software, Inc.                      Sunnyvale, CA


 Sign up for the DeepChip newsletter.
Email
 Read what EDA tool users really think.


Feedback About Wiretaps ESNUGs SIGN UP! Downloads Trip Reports Advertise

"Relax. This is a discussion. Anything said here is just one engineer's opinion. Email in your dissenting letter and it'll be published, too."
This Web Site Is Modified Every 2-3 Days
Copyright 1991-2024 John Cooley.  All Rights Reserved.
| Contact John Cooley | Webmaster | Legal | Feedback Form |

   !!!     "It's not a BUG,
  /o o\  /  it's a FEATURE!"
 (  >  )
  \ - / 
  _] [_     (jcooley 1991)