Verilog vs VHDL

Jim,

Any snapshot of just the last year, at the calendar year mark, can be 
safely ignored if you try to get too much from it.

If you take the last five years, then maybe you get to see the average 
trend (which is evident when you look at the overall market share -- 
hard to get to 50+% without actually showing some pretty impressive 
growth...).  That Altera is (still) ~33% is also not lost on anyone. 
The whole market has more $$$, so we all show growth in $$$, even if 
some of us are not really gaining market share.

A quick review of "How to Lie with Numbers" would prove instructive to 
anyone who wanted to spin the data to their advantage.  That said, I 
have my own bias, and the way I chose to post the numbers could be 
considered highly biased (not intentionally, really).

So, the whole reason for the post was to show that 2004 vs 2005 was a 
generally good year for PLD suppliers compared with digital ASIC suppliers.

It was also to show that PLDs are about 1/5th the total logic market, so 
VHDL and verilog are pretty useful to know.

Austin

Austin Lesea wrote:

> Jim,
> 
> Any snapshot of just the last year, at the calendar year mark, can be 
> safely ignored if you try to get too much from it.
> 
> If you take the last five years, then maybe you get to see the average 
> trend (which is evident when you look at the overall market share -- 
> hard to get to 50+% without actually showing some pretty impressive 
> growth...).  That Altera is (still) ~33% is also not lost on anyone. The 
> whole market has more $$$, so we all show growth in $$$, even if some of 
> us are not really gaining market share.

I was wondering why the Altera results differed in two areas, but 
Xilinx's did not. Did the report not clarify that, or was it a typo ?

> A quick review of "How to Lie with Numbers" would prove instructive to 
> anyone who wanted to spin the data to their advantage.  That said, I 
> have my own bias, and the way I chose to post the numbers could be 
> considered highly biased (not intentionally, really).
> 
> So, the whole reason for the post was to show that 2004 vs 2005 was a 
> generally good year for PLD suppliers compared with digital ASIC suppliers.

Yes, and an even better year for Foundries, who some could consider the
ultimate 'ASIC suppliers'  :)

> 
> It was also to show that PLDs are about 1/5th the total logic market, so 
> VHDL and verilog are pretty useful to know.

Did those numbers include OnSemi/Philips/TI etc standard logic devices ?

-jg

Jim,

Answers, below,

Austin

-snip-

> I was wondering why the Altera results differed in two areas, but
> Xilinx's did not. Did the report not clarify that, or was it a typo ?

Not sure what you are referring to.  One thing is that Hardcopy is
including in those Altera numbers, and we don't see Hardcopy as anything
but an ASIC, so it makes it hard to compare how the programmable logic
sector is doing (is Altera falling more rapidly behind in FPGAs?).

> Yes, and an even better year for Foundries, who some could consider the
> ultimate 'ASIC suppliers'  :)

Foundries are doing well, right now.  Tough business.  Low margins.
Cycles up and down.  Huge capital costs.  Glad I am not in it.

> Did those numbers include OnSemi/Philips/TI etc standard logic devices ?

No, they did not.  All standard logic products were not included in any
of the numbers.  Neither were memories, microprocessors, nor DSP.

Standard logic in 2005 was $1,579 million total.
DSP was $7,569 million total.
uP+Memories was $104 billion!

As I said, you can make the numbers tell any story you want.

For example, comparing 2005 with 2002: (using company reports)
2002:
Altera, $712 million revenue for the year (No Hardcopy back then)
Xilinx, $1,016 million revenue
2005:
Altera, $1,124 million (4% or $45 million is Hardcopy, really $1,079
million, as noted in their public filing)
Xilinx, $1,573 million

So we grew 1573/1016= 55% and Altera grew 1079/712= 52%.

Could have fun all day with numbers, but I need to get some work done now.

Austin Lesea wrote:
>>I was wondering why the Altera results differed in two areas, but
>>Xilinx's did not. Did the report not clarify that, or was it a typo ?
> 
> 
> Not sure what you are referring to.  One thing is that Hardcopy is
> including in those Altera numbers, and we don't see Hardcopy as anything
> but an ASIC, so it makes it hard to compare how the programmable logic
> sector is doing (is Altera falling more rapidly behind in FPGAs?).

#Looking at PLDs alone (no surprises here):
#Xilinx up 3.7% with 50.3%
#Altera up 7.6% with 33.1%
#Which makes these 5 with 96.8% of a $3.17 billion market
#So, what happens when you combine the Logic ASIC with the PLD markets,
#and look at them together?
#Xilinx up 3.7% with 10.5%
#Altera up 10.6% with 7.2%
#A combined 45.7% of the total market, which accounts for $7.13 billion.

In both blocks, Xilinx is +3.7%, but in one Altera is +7.6%, and the 
other +10.6%, so that was what caught my eye.

The 'second Altera' seems to be slightly larger, which may be tbe $45M
hardcopy you mentioned, included in that block, but not in the first ?

-jg

Jim,

Could be (where is Hardcopy included/not?).  I used company reports for
the second posting, and an independent market firm's results for the
first posting.

I am pretty sure the marketing firm has all of Altera's revenues under
PLD (including Hardcopy).  The yearly report clearly states Hardcopy as
4% in 2005.

It is pretty picky of us to not include it (exclude from all
comparisons), but we do so to call attention to the volume of business
that they do, that detracts from their mindshare in FPGAs.  It is as if
they also make buggy whips.  Who cares how much money they make on buggy
whips.  They state in their yearly report, that Hardcopy could/might
become 10% or more of their business.  Given that LSI bailed, and had
$145 million, that might be too low an estimate.  If Altera gained all
LSI business, that would be almost $200 million for structured ASICs in
their basket.

No small number, but from our experience with Hardwire 1,2, 3 etc. that
business is a huge distraction, with poor to non-existent margins.

It does offer customers a path for cost reduction, and its primary value
is in the fact that most customers can't stop changing things, so they
must stick with FPGAs far longer than they wanted to.  Or, they can't
make the Hardcopy work, so they have to use the FPGA for much longer
than they wanted to.  Timing closure is still the number one problem
with any "hardened" approach.

However, when the hardened version doesn't work, the customer usually
demands a price concession on the FPGA until it does work.  Such
'difficult conversations' are conducted on an adversarial basis with
customers and generate negative "goodwill" and are never good for a
company.  I'd rather not be in a business where I have to get tough with
a customer to stay in business!  Once burned, they tend to go away, forever.

I'd rather not deal with a manufacturer that has tough guys used to very
angry yelling and screaming customers that I must negotiate with...

I am surprised that no one pointed out how good 2001 was and how bad
2002 was.  The dot-com bust really burst many bubbles.  Again, for
really good comparisons, you might want to go all the way back to 1985.

Austin

I am in no way surprised that this thread turned into a FPGA vs ASIC
and then a Xilinx vs. Altera (or Xilinx vs. the world) war. What I
don't understand is how come Jim hasn't recommenced any Lattice devices
yet ;)


Jokes aside, please stop the madness

burn.sir@gmail.com wrote:

> I am in no way surprised that this thread turned into a FPGA vs ASIC
> and then a Xilinx vs. Altera (or Xilinx vs. the world) war. 

Hmm - I don't see much of a war here  ? - have I missed something ?

Austin put some numbers out, and I queried some details of that,
but I had no real issue with the numbers themselves.

It all seemed quite civil to me ?
None of Austin's replies offended me, and I don't think I offended him ?

> What I  don't understand is how come Jim hasn't recommenced any Lattice devices
> yet ;)

and how would that be relevent, to this thread, or why should I ?

> 
> Jokes aside, please stop the madness

madness ?  - I'm sure I missed something ...

-jg

burn,

OK, I will no longer post to his topic, since you object.

Austin

burn.sir@gmail.com wrote:
> I am in no way surprised that this thread turned into a FPGA vs ASIC
> and then a Xilinx vs. Altera (or Xilinx vs. the world) war. What I
> don't understand is how come Jim hasn't recommenced any Lattice devices
> yet ;)
> 
> 
> Jokes aside, please stop the madness
> 

The most offensive posting here was by burn... ,
dressing down Ron as if he were a dumb schoolboy.
Otherwise it was the usual banter...
Peter Alfke