Friday, February 03, 2006

Fan boys ask world to wait for INTEL

Hans Mosesmann at Moors & Cabot are asking people to wait for INTEL's Woodcrest CPU, which is expected to be out in 3Q06. He claimed that Bensley (Dempsey CPU + Blackford chipset) is being shipped for testing in 1Q06 (now) and will be out in volume in 2Q06, but then the socket compatible Woodcrest CPU can replace the Dempsey CPU in 3Q06.

Does Hans Mosesmann think people are stupid? First, in Q1, they buy INTC Paxville, which is too dumb, too slow and too hot; then in Q2 they buy INTC Dempsey the Athlon MP imitator, which is slower than old Opteron 246 in some tests, and then 3 months later, they find Dempsey too hot and "pull in" INTC Woodcrest, which is supposed to be cooler than Dempsey. If people were indeed stupid as Moors & Cabot liked to be, then INTC would make the same money three times in a row, leading to some $1.7 EPS.

The good news for the shrewd server buyer (such as Marc Andreessen) is, AMD will release quad-core Opteron in mid-2006. The fact these quads can plug into existing Socket 940 boards means their power consumption is no more than 95 watts. There are no surprises here. The Turion MT-40 is a 2.2GHZ CPU with max power of 25 watts, 4 of these make 100 watts. AMD should have no problem making a quad-core 2.2GHZ Opteron at 95 watts.

The choices for Opteron buyers are clear. They can buy single or dual core Opterons now with power of 30 watts (Opteron EE), 50 watts (Opteron HE) or 95 watts. Then they can upgrade to quad-core Opterons later with a simple BIOS flash.

No wonder that 90% of Rackable's revenue comes from Opteron, and it's no surprise that SUN could snatch a $400 million workstation deal with Opteron at the last minute of the bidding process.

Hans Mosesmann's FUD on Socket F Opteron and HyperTransport is pitifully laughable. Let me assure you this, the Socket F Opterons will be so advanced and strike so hard, INTEL will be beaten into guacamole.

BTW: Look at this score card, zero stars on seven of of the eleven companies covered. A monkey could have done a better job by throwing darts.

Analyst Hans Mosesmann, with analytical accuracy seemingly less than that of a dart throwing monkey, wrote with a sense of desperation,

On 2006 outlook: "AMD Valuation. Our 12-month price target of $10 for AMD shares is based on a 20x multiple to our 2006 EPS estimate of $0.50. The 20x multiple is an 11% premium to the company's historical 18x P/E ratio and is actually the same multiple we are applying to INTC shares, which is justified by our view given that AMD is a significantly better managed company today, under CEO and Chairman Hector Ruiz."

On AMD lawsuit: "[T]he timing of the complaint and the company's tone of indignation (perhaps exasperation) give us a sense that AMD's traction in processors that they enjoyed last year is not meeting expectations"

"It strikes us that AMD's approach is based on throwing everything but the kitchen sink at Intel in the hopes of getting something to stick."

Angry man, how is your $12 AMD short doing?


Anonymous aaron said...

sharikou, i have thoroughly enjoyed reading your insightful comments since i first found your site. it was with pure amazement that i read the "analysis" put forth by hans moeesmann. it must be a type-o, because i can not understand the price target of 10 or the 2006 eps estimate of .50. i've got degrees in both finance and economics, i think it is time for me to get my resume over to moors and cabot. i am trying to finish my graduate work, but i can't see how i could not land a job with them right now. i certainly could do better than him, maybe not the dart throwing monkey, but certainly him.

12:07 PM, February 04, 2006  
Anonymous Anonymous said...

Where's the beef...

Nice hatchet job on Linux/AMD by Sun. Heaven forbid they open up the hood and let people look at how they came by their numbers. $288/month support fee for Linux? WTH???

Btw, Mark Andreesen is an idiot one-hit wonder.

What did Mark Say?

* We originally planned to use Linux on Intel commodity hardware (because we were too stupid to live until Sun showed up with their marketing money)
* Solaris is a better Linux than Linux (Based on what? Oh right, the Sun marketing money)
* Solaris has similar economic and OpenSource advantages as Linux (Since when does "vastly inferior" translate into "similar"? Oh yeah, when you start sucking on the teat of Sun marketing money)
* It's very easy to move from Linux to Solaris (It's even easier to move from Solaris to Linux - just ask all the people who've done it over the last 5-10 years)
* Solaris 10 reliability is a very big deal for us (The reliability of Solaris 10, especially on x64, has yet to be proven since it is BRAND SPANKING NEW)
* Sun is now riding aggresive side of the price/performance curve (Based on what?)
* With Sun hardware, we know everything has been tested and intergrated together (We also know that our marketing material that we publish for all to see has been through a "speel chekker")
* With Solaris, we know if we need vertical scalability, we'll actually be able to get it (... when we move to Solaris on SPARC and lose all the advantages of being on Opteron to begin with)

Sun is not going to ride the crest of the next tech wave because they assume their customer base is as dumb as they are. This may have worked in the 90's, but not now.

12:57 PM, February 04, 2006  
Blogger Sharikou, Ph. D said...

You would threaten Hans Mosesmann's SVP position if you apply a job at Moors&Cabot. In my opinion, the smartest (IQ-wise) analyst covering AMD is Joe Osha, and the dumbest is Hans. You can tell that by just reading their reports, you see Joe Osha has a coherent thinking. Even if Joe Osha was FUDing, he made it far less obvious than Mosesmann's. Hans Mosesmann sounded more desperate and irrational than the shorts on Yahoo msg board, I can safely bet he and his moors cabot customers have a large AMD short position just above $10. The shorts on yahoo board would say INTEL will launch a price war, so AMD will not see profit even though it gains market share. Mosesmann's argument that Woodcrest will pull in Dempsey is plainly idiotic. The fact INTEL is advertising a platform that will be incompatible with today's Paxville will definitely hurt the sales of 1Q06. Even if you are die hard INTEL fan, you don't want your purchase be obsolete in just 6 months. If customers know in 6 months the 2.33GHZ Woodcrest will be cooler and faster than today's 3.2GHZ Paxville, they will think more about alternatives now. The same applies to desktop and notebooks. The best thing for INTEL right now is most people are unaware of INTEL's plans and unaware of AMD's huge lead, this is especially true in the 3rd world.

12:58 PM, February 04, 2006  
Blogger Sharikou, Ph. D said...

To be fair, Linux is definitely not the best breed. Linux was based on the old book on AT&T UNIX, the design was orthodox, but also quite ancient. Solaris 10 is based on BSD, the Berkeley Unix written by SUN co-founder Bill Joy. There were some performance benchmarks on FreeBSD and Linux, and FreeBSD won most of the benchmarks, especially networking performance. Also, the Linux filesystem EXT3 is definitely not the best. The appeal of Linux to many was that it was a 100% rewrite from scratch. People stick to EXT3 not because it's the best, but because of pride. Linux is also not as stable as people wanted, I personally experienced kernel problems under extreme load, in another occassion, my linux box was rendered broken after the root filesystem became full, these indicated that there were problems in Linux scheduler, kernel reliability and system integrity.

Now look at Solaris 10, it's simply amazing. The only problem SUN needs to solve is to make Solaris 10 easy to use by any Linux or Windows user.

1:15 PM, February 04, 2006  
Blogger Sharikou, Ph. D said...

Let me add a further comment on SUN's software drive. The plan of bundling enterprise software stack and make money failed. The software industry is a winner take all game. The only best thing SUN has are Solaris 10 and Identity(maybe a few other minor things), most others are 2nd class. I am not sure open source SUN J2EE server will attract more developers. But it might succeed, as we see from the Apache Tomcat project, which is derived from SUN's reference JSP engine.

1:32 PM, February 04, 2006  
Anonymous Anonymous said...

Wow. Just wow. Where do I begin on this steaming pile of FUD that you've dished up for Linux?

I've run production FreeBSD and Linux mail chains, DB systems, J2EE app tiers, you name it side by side for several years now. In almost every application (except the mail chain architecture), Linux has beaten FreeBSD. The gap widened considerably once the 2.6 kernel came on the scene. FreeBSD had nothing to compete with that, and to this day is struggling to catch up while mired in compressed release cycles and the no-man's-land of their M:N threading implementation.

I've found uptime and stability to be quite equal under load (daily concurrent session loads of 600 to 700), so no complaints there. It's just that, for ever application other than e-mail, the FreeBSD machines just start to get "tired", for lack of a better way to describe it. I can keep a Linux node up against that inrush for months at a time, with the only "restart" coming from application upgrade cycles (the node itself stays up). Not so with the FreeBSD machines - they live in the mail chain only now.

Bottom line as a generic observation... Google on Linux is whipping Yahoo! on FreeBSD like a red-headed step child. Some portions of Yahoo! are either moving or have been moved to Linux ( last time I checked) because of the -limitations- of FreeBSD. Both Linux and FreeBSD have their limitations, so you have to *test and benchmark under load* to flush them out for your given application.

At least it's not Windows Server *shudder*.

Back on topic re: Linux/AMD vs. Solaris-x64/AMD. Show me the apps and show me that they run better/faster on one vs. the other. I have a sales compensation package running side by side on SLES 9 for AMD64 vs. Solaris 10 x64 as we speak. This is not a trivial app - companies use our product to cut checks and perform commission resolution. What we're finding is that the Sun 1.4/1.5 JREs are faster on Linux than they are on Solaris 10 x64. You would not -believe- the level of regression we went through to confirm this, since it was such a political hot potato. I'm sure the situation will improve over time (especially given our feedback LOL), but for now Solaris 10 x64 should be considered "new" and not necessarily ready for prime time on some applications.

Oh, and Oracle 9i EE 64-bit AMD64 would be nice to have for Solaris 10 x64. I can get that for Linux, btw - it runs like a scalded ape.

5:19 AM, February 05, 2006  
Blogger Sharikou, Ph. D said...

I think the Linux 2.6 kernel has improved a lot. I use it everyday. So far, I had few problems with it . The udev stuff is also good. FreeBSD has few developers working on it, and can't match Linux in general. Last time I checked, it didn't have a AMD64 port. However, in terms of technicals, Solaris 10 is far more advanced at the kernel level, we should not deny that.

2:28 PM, February 05, 2006  

Post a Comment

Links to this post:

Create a Link

<< Home