backtop


Print 48 comment(s) - last by blwest.. on Sep 16 at 3:06 AM


AMD's Rick Hegberg  (Source: AMD)
AMD loses another executive

In late August, DailyTech reported that Henri Richard would leave his position as executive vice president and chief sales and marketing officer for AMD on September 8. Richard had been with the company since 2002 and saw the Sunnyvale, CA-based company go through good times (Opteron successes) and bad times (post ATI acquisition difficulties).

Today, The Wall Street Journal is reporting that another AMD executive is jumping ship. Rick Hegberg, AMD's VP of worldwide sales, is leaving the company although a company representative declined to indicate when the resignation would take place or his prospects afterwards.

Hegberg was a former ATI employee and came to AMD after the 2006 acquisition. Hegberg is also the second former ATI executive to leave AMD -- former ATI CEO Dave Orton left in July.

It's interesting that three key executives are abandoning the company prior to the release of Barcelona. AMD's native quad-core processor family is just around the corner and the company has already announced pricing for the new chips.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: Merger was a Mistake
By DFranch on 9/5/2007 12:50:11 PM , Rating: 2
New CPU's are in development for years, the ATI merger is a recent development, and had little to do with the fact that AMD had no answer for the core 2 duo. AMD got caught with it's pants down with the core 2 duo plain and simple. ATI has also been having trouble since before the merger keeping up with Nvidia. The 7 series video cards were better than anything ATI had at the time, and the 8 series just blew them out of the water. Only time will tell if the merger is a success. The problems you are seeing now would have probably happened to the two companies independently anyway. Hopefully the merger will pay off eventually.


RE: Merger was a Mistake
By DrKlahn on 9/5/2007 1:17:29 PM , Rating: 2
Incorrect on the 7 series. The X1800 was a match for the 7800GTX with AA/AF enabled. The X19xx series was better all around then it's 7xxx counterpart. Only with the R600 series are we seeing ATI falter.


RE: Merger was a Mistake
By maroon1 on 9/5/2007 2:00:31 PM , Rating: 4
First of all, don't forget that 7800GTX came 6 months earlier than X1800

The midrange ATI X1xxx series were worse that than Nvidia midrange 7 series. 7600GT was much faster than X1600XT, and 7600GS was faster than X1600Pro. Thats why ATI replaced X1600XT and X1600Pro with X1650XT and X1650pro


RE: Merger was a Mistake
By Jedi2155 on 9/5/2007 2:49:21 PM , Rating: 4
The X1800 series was developed and design in time to be competitive with the 7800GTX, but it was due to a bug in a third party design library (they are using a hardware description language to develop these chips), that caused a leakage problem in the chips that caused them months to locate.

This is also why the x1800 and x1900 series were only 3 months apart in their respective launches. The different design teams had their designs already done, so when they found the bug it was a simple matter to fix.


RE: Merger was a Mistake
By jarman on 9/5/2007 4:53:48 PM , Rating: 2
Um, no. The X1950XTX was an improvement on the debacle that was the X1800 series, but the crossfire implementation was a poor alternative (dongle anyone?) to nVidia's superior SLI technology (e.g. 2 x 7950 GTX) in the market at the time.


RE: Merger was a Mistake
By Darkskypoet on 9/5/2007 5:43:43 PM , Rating: 4
AMD didn't get caught with it's pants down...

What AMD did do, was gamble.. and they lost.

First off, Barcelona is late, very late. Does anyone remember the Athlon XP stepping that was Thoroughbred A? Remember how bad it scaled? Take that and multiply it by a lot. They had to take an architecture that was of common place complexity, and rework and spin it quick.. and voila, Thoroughbred B. What did it take? An extra metal layer, and that baby scaled up incredibly well.

Fast forward to the first iterations of the Barcelona core... The gamble was in going monolithic vs MCM. When the biggest baddest Fab guys on the block (intel) don't go monolithic.. there might just be a technical reason. AMD's error was in the hubris of believing they could pull it off in time. (very few if any fab based companies can pull of Intel's manufacturing prowess. This is simply a fact)

They didn't and couldn't, and now we see the fall out. Barcelona, and a competitive IPC core to core 2 duo was supposed to arrive well before now. Had it, we would still be in the throes of an actual CPU performance war instead of a price one. Sadly, we are not. (except in some of the most obscure yet profit rich environments)

2nd idiotic gamble: Phasing out 939 far to early. The only way you hold the legions of enthusiast converts from going C2D is to give them a drop in replacement for their aging single core chip. Force them to buy a platform, and they will not usually buy yours if you are not competitive.

Dropping 939 sent the market share back Intel's way in a decently (not wholly responsible) large way. Especially when combined with the channel screwing that occurred.

Miss a deadline, and kill you're only solution that would delay platform upgrade for your installed base is ludicrously stupid.

However, Thats what happens when you have very few fabs. Opportunity cost baby!

Had AMD inked a deal with chartered (tsmc, etc)to make the less complex 65nm core for 939 and AM2, moving themselves wholly to AM2, it might have been a smarter move. Thereby gradually phasing out 939. Obviously they learned nothing from ATI's major mfg partner Sapphire, who still releases AGP cards to an existing un-catered-to market.

R600...

When strapped for resources to allocate to all of your stream processing needs (GPU, CTM Server boards, Pro Grphx solutions), you go for the architecture that is applicable to as many market segments as possible.

Enter R600. Yes it hasn't shown up as a dominating top end Single Gaming card solution. However, look at the professional level benchmarks for the card, and look at the power unleashed from such a "GPU" in CTM custom coded apps. (for example folding)

When Pro level Fire GL cards sell for up to 5 times that of the 2900XT for the same cost, do you allocate resources to make the gaming card better, or to make a kick ass general purpose vector processing monster?

The Fire GL will eat the Quadros for breakfast.

There are very few, if any (that I am aware of) methods to use 8800 series parts for folding, or other massive vector processing ala the CTM push from ATI/AMD. This is yet again another sector of the market that the R600 is king in.

Many of us on this site are very self centered to believe that a large company is going to always focus very scarce resources (especially now) on winning the uber gaming card contest fanboys love and live for. Is it preferable? Yes. Is it necessary on a strategic level? no. Not when you already have multiple massive OEM design wins for the midrange (yuck this time around) and low end discrete market.

In sum:

K10

Pants down with Core 2 Duo? Hardly. MCM Barcelonas could have been released (if they went that route) a long time ago. They gambled, and lost. Those are balls of brass!

R600

Not everyone's priority is to win the Gaming card competition to the exclusion of all else. And if you haven't noticed Crossfire is currently superior to SLI, and will only continue that superiority as it seems scalable up to and past 3 GPUs. There is a reason Dual 2900XTs hold the current performance crown. Also the new Fire GLs are coming... I doubt we'll laugh at the 2600xt core again.


"This week I got an iPhone. This weekend I got four chargers so I can keep it charged everywhere I go and a land line so I can actually make phone calls." -- Facebook CEO Mark Zuckerberg

Related Articles
AMD Prices "Barcelona"
September 4, 2007, 5:06 PM
ATI's Dave Orton Resigns
July 10, 2007, 2:31 PM













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki