Friday, October 28, 2005

INTEL's rushed introduction of Paxville was a foolish move

Under educated idiot got rich overnite can never be relied on providing good advice. INTEL is a technology company, INTEL is not fundamentally a manufacturing company. You need to get that positioning right. INTEL might have been forced to release 400 watt Paxville under the urging of some idiot who knew zip about CPU design. It was really a silly move, it only showed how far INTEL is really behind. Before Paxville, people say, OK, AMD uses dual core to beat up INTEL single core, that is unfair, INTEL's dual core will frogleap. But with Paxville dual core Xeon, people start saying INTEL is really dumb, FSB is really dead, and AMD is really good, HyperTransport is the future and INTEL doesn't have it.

The Paxville situation is really pathetic. INTEL announced the launch of this chip one day before AMD's blow out Q3 2005 earnings report. But as of today, it's apparent that major hardware reveiew sites (such as Anandtech and Tomshardware) can't get their hands on a couple of these chips to benchmark them. The DELL dude celebrated Paxville, in his usual AMD insulting fashion, but requires people to sign NDAs to buy one. Disgusted by DELL's behaviour, INQ, a site with 30 million page views a day, insulted back. HP is honest as usual, and says Opteron rulez INTEL. IBM, still tweaking their Hurricane chip to work with paxville and is probably busy designing bigger heatsinks and fans. Then, unsurprisingly, DELL dude found that it can't find a good enough Voltage Regulator to handle the Paxville power -- 400 watts that is.

Then INTEL came out with slew of cancellations, and people finally realized that INTEL is indeed 6 years behind -- as I pointed out all along. Experts are now saying INTEL won't be able to compete till 2009.

UBS, which dumped 20% of their INTC holdings in June, upgraded INTC from $23 to a $37 target price on Paxville news, claiming INTEL was probably leapfrogging AMD. Now they know, the frog leaps allright, but it can't catch the antelope.

Tom Yager revealed some secret: AMD is adding some core instructions to Opteron and Athlon64, while INTEL hasn't completed its copying of x86_64 yet. I proposed to AMD long ago that AMD should add instructions that are specific to AMD64 architecture and hard to implement in any efficient way on other architectures, guess AMD folks had this in mind all along.

I said before that DELL is the next Enron. Its $2 bookvalue won't be able to support the $32 stock price no matter how you look at it(It was $38 last time I said it last time). Face it, what DELL has besides being INTEL's favourite darling??? Direct sale model? Everyone is doing it now...what else? a bunch of warehouses, a pile of screw drivers and a brand...? those worth near 0. How about the big revenue? Enron had bigger revenue, but it is the value you added that counts. DELL sells millions of PCes at rock bottom prices, that push the revenue numbers up, but DELL's added value is only the labor to use screw drivers to assemble the third party OEM parts. Now INTEL is tumbling, DELL's feeding tube is in danger of being disconnected, and without a brain (DELL's R&D is so tiny compared to its market cap), it will be declared clinically dead sooner or later.

Now look at SUN, SUN has more net assets than DELL, and they have vast amount of technology and a grand vision. SUN's technology is simply too much to listed here -- what they need to figure out is how to focus and monetize the technology, and the $750 Galaxy server is a good start to make SUN's enterprise expertise available to the masses that can't afford the million dollar machines. The Solaris 10 is the only enterprise class OS and SUN is also making it available to the masses too. But now, Solaris 10 is inapproachable. I am a UNIX person, I used Solaris for years and I use Linux daily, but I am scared by Solaris 10 despite my utmost respect for it. How do I start creating containers, how do I create partitions on ZFS, how do I start sshd? It must be an all new set of command line options again, and I will have to start from zero with "man man"... Once usability is improved, Solaris 10 has the potential to unseat Linux in the data centers. My suggestion to SUN is to make Solaris 10 as easy as Windows Server 2003, with integrated GUI control panel and configure tools to perform most basic system adminitration tasks. This is actually not trivial and requires some deep level of integration. But, if SUN can spend less effort on looking class, but more on Solaris 10 Control Panel, and make every dummy capable of running Solaris 10 server, packed with exotic features such as ZFS and 8000 Containers -- now imagine tha impact of that ---on Windows Server and Linux.

AMD should never deal with DELL. AMD capacity is limited and can only supply HP and SUN. Having DELL using AMD damages the system integrity underlies the AMD brand. AMD selling to DELL would be betrayal to SUN and HP, who are looking to grab DELL's customers. Craig Barret once said if you are one generation behind you are dead. I remember these words well, and I think it should be declared as Craig's Law, no less important than Moore's law. Craig's law is universal and of course applicable to INTEL as well.

The future is indeed bright. Technology can win .


Blogger plantlife said...

Nice job, sharikou.

10:26 AM, October 28, 2005  
Blogger Dr. Dweeb said...


If you actually lived in the VMS, Tru64, NonStop world you would know that the Carley vision was rather different.

HP has reneged on promises (killed the Alpha EV79, killed Tru64) and has been doing its level best to kill OpenVMS and NonStop since the acquisition. Enterprise customers have grown tired of HPs mendacity.

HP has been a disaster for these technologies and Carley's vision has never included them.

Carley has embraced the Itanic as the "one processor" for the HP/UX, OpenVMS and NonStop lines - a chip surely doomed to extinction. Customers are leaving these technologies in droves. The only reason it is not showing up yet is that these systems are so critical for most companies that we need 5+ years to implement alternatives.

I am currently on a project to eliminate OpenVMS from a corporation. It will take years and HP will lose 100% as we will be an IBM AIX site when complete. We are a non-trivial company.

We are one of many. Carley thought we would all embrace her HP/UX Itanic solution which is at best marginal and at worst a dead-end

You should check some facts before you praise that women. There is a good reason that we danced on the desks when she was fired !!!!

9:11 AM, November 28, 2005  

Post a Comment

Links to this post:

Create a Link

<< Home