an editorial on companies in the business

[Sazar]

please do not link this to the front page newsposters... I do not think it is appropriate for that

I'll give a brief overview so as to explain something before you embark on the read...

uttar has for the past year or so been posting various tidbits of info regarding gpu's and the like... he knows a fair deal about gpu's and marchitecture and was generally well regarded wrt his info... the reason I say WAS is because of something he released that IMO was inappropriately handled... but thats another story...

the following article is an opinion piece... please read it as such... it has various bits of information in it that some may take as fact...

my suggestion is not to take the article's comments as fact at all nor believe all the numbers and other print you see... rather see it for what it really is... an opinion piece backed by some numbers from sources who may or may not be right...

I normally don't preface a link or a post with information such as what I am doing now and the reason I have decided to add a preface will become clear when you read the article...

after reading... post away... its always fun to have a discussion on topics like this...

happy reading

uttar's editorial

 

[Geffy]

got to page 2 got bored and went somewhere else

 

[LeeJend]

Uttar hasn't been around much has he?

Corporate Facts - based on 30 years in engineering in 10 different companies ranging from 10 employess to 200k employees:

Every new company goes through a life cycle. They are started by entreprenuers with a good idea (or more likely some friends with a good idea). The entreprenuers become corporate executives who think one hot idea translates into eternal clairvoyance. The neo corporate executives drive away the remaining talent by becoming the oppressive *******s they used to hate. (Famous quote "Slaves don't want to be free, they want to change places with their oppressors".)

BTW the same thing I said above has been published by a top manager in his book "Life Cycles of a Corporation".

The real question to ask about Nvidia is what 5-10 employees left taking 95% of the corporate talent with them?

The other, slimmer possibility is linear thinking. CPU speed grows linearly for a time then flat tops because the straight line growth in clock speed can't be supported by the materials technology anymore. We're at a brick wall now. Look at the tom's hardware table of speed tests for processors over the past 10 years. Straight line to a wall then a quantum jump.

8086 to 286
286 to 386
Pentium I/II/III to P4
Pentium I/II/III to Duron/Athlon
P4/ Athlon to ????? we're due but the 64 bit chip isn't it.

Voodoo ran the straight line into oblivion.
Nvidia almost (and still may) do the same.
AMD did the same and still may not survive it.

ATI has been through it's first corporate cycle and proved smart enough to survive it. Cheers to their management.

Intels been around for a long time and proven it's staying power. Note they have ~70% of the video chip market and if ATI or Nvidia piss them off watch for the fastest video cards in the world to all say Intel Inside.

Most likely Nvidia squandered their corporate talent producing the Nforce chipsets hoping to steal Intels 70% "integrated solution" market share. Result the hot shot video designers not interested in (eww, yuk) integrated chipsets leave and the "B" team takes over. End of story, end of company.

The scary part is:
$500 video cards that are only 25% faster, need multiple cooling solutions and extra power connectors.

100W CPU's when we can't adequately (quietly) cool a 75W CPU.

Liquid or even freon based CPU cooling solutions on your desk or worse yet, in your living room.

Time to get off this straight line and start a new curve.

 

[Sazar]

uttar knows a fair deal about gpu's... he may assume that he knows more about business models than he really does but tbh he is basing a lot of his stuff on sources who many not be entirely accurate..

nvidia's main problems are rumoured to be overstretching engineers on the xbox and not dedicating enough to the nv3x lineup... the r3xx being better than anticipated causing them to do the tweaks et all that they had to...

ati had a terrible structure and only with the acquisition of ArtX and management changes have they managed to change things around and become both profitable and start to lead the industry in tech...

I dont really go along with the nforce chipset design being a waste leejend... in fact that is one of the more succesful recent ventures nvidia has had...

one thing to keep a track of... gpu's are becoming more complex than cpu's... they have more transistors by a fair degree and the next iterations will only widen this gap...

the products will likely be running cooler with new developments in manufacturing.. I am personally not too worried about that..

 

[LeeJend]

You misunderstood. I love what they did with the nforce chipset. It was a quantum leap.

The problem is companies usually dilute their existing engineering team to break into new areas. That leaves the B team running the bread and butter product line and trying to make enhancements. It's never worked at any of the companies I've worked for. Usually ends up with a crisis for the bread and butter product line while the new stuff isn't generating any significant profits yet. I could name companies and products but then I'd be unemployed.

Wait I can name one. Apple. When they had the whole Apple II / LISA / Macintosh debacle that nearly put them out of business.
Very similar to nvidia right now.

Read "West of Eden" story of the early days at Apple. Good reading, and very sad.

 

[Electronic Punk]

Still makes for interesting reading

 

[LeeJend]

So does West of Eden. Grab a copy. You won't know whether to laugh or cry.

 

[Sazar]

thread resurrection of the day?

 

[LeeJend]

Don't look at me, Punk started it!