Crusoe vs. Dell And Compaq

Boone^ writes "Yahoo! has an article from ZDNet News that details how Compaq and Dell are shying away from Crusoe notebooks 'for now.' " Basically it says that the performance isn't so hot, and consumers are gonna be burned by the hype of the first generation of Transmeta based laptops. But then again ... the battery life sure ain't a bad thing. Mentions that Hitachi notebooks might be shipping as early as October. Update: 06/28 09:37 by CT : here's some pictures of transmeta laptops.

Comments

News flash: PC Maker dislikes chip they don't use

[Golias]

In other breaking news:
New desktop PC's out-perform laptops
North Dakota expects colder weather than Texas this year
Gerneralissimo Francisco Franco is still dead

Of course Compaq and Dell are going to tell you that this chip is crap. They are not using it. The companies that are using it are just as sure to tell you that the chip is an "Intel Killer" or something.

Several good reasons...

[Christopher B. Brown]

• Linux isn't yet, by itself, a good enough reason for the existence of a laptop product.
• Unless you can get economies of scale from selling lots of 'em, the problem of low volume will result in high prices.
• No WordPerfect, and limited maturity of web browsers.

I'm still watching for the first release of OpenPPC hardware; it too is not expected to be inexpensive. TotalImpact cards sound rather cool, but are apparently expensive enough that the vendor isn't willing indicate any pricing information on their web site; reportedly about $1K per CPU.

The "pricing structure" behind the PPCs just doesn't seem suited to laptop deployment that occurs "because they're low powered."

Laptops are concerned with Performance

[anon1209]

Laptops have always been lower performance than desktops. (IMHO) Why should this be an issue now?

Re:Laptops are concerned with Performance

[FoulBeard]

Its not about performance... they can be no way this is about performance. Compaq, and Dell stick cheap IDE hardrive low memory, crappy sound cards, want to mention performance, what is up with the 2MB vid cards on most laptops? I know laptopss displays are only so big, but what about port replicators, and external monitors. PErformance my ass. These companies are going with Transmeta becuase if they do then they will probably lose their special Intel exclusive partnership. I mean really I have never been sitting at my laptop and had to say, damn P2300 this thing is soooos low, I mean..jezz I have to wait .00004 seconds for Word200 to spellcheck my 200pg document.

On the contrary I am usually sitting their waiting for the underpowered harddrive to slowly bring word, then to wait 15 seconds while I alt-tab to Netscape, and have to yank the entire app out of swap.

Most people dont use laptop for quake3, and Starcaft runs fine on a P200, linux will fly on a P200, (enlightenment wont if you the eye-candy on though, not to mention efm ). Anyway my point is that you could replace all of those P2s. and P3s laptop cpus with a P166 and 95% of people would even notice the tiniest burp. Only power-geeks only *really* notice stuff about that. So Dell, and Compaq are blowing crap out their asses again. Same crap, they want to keep the benefits that Intel provides ala M$ style. If you dont like that then its easy buy a Sony or Toshiba they both make nice laptops for ususally better prices than dells.

Re:Laptops are concerned with Performance

[Spasemunki]

For one thing, the increase in desktop and laptop performance across the board has increased the expectations for what a laptop can do. Consumers are no longer content to use ill-designed, unsupported, and generally flakey laptops just because they don't sit on a desk. The majority of senior management in my office doesn't have desktops- they have a laptop, and a docking station in the office. Laptops are increasingly fulfilling traditional desktop roles, and are a big deal with business users that travel and make off-site presentations. If Crusoe is a real drop in performance compared to a non-Crusoe laptop of comperable cost, the business world ain't gonna bite, and that's where a lot of the new money is.

"Sweet creeping zombie Jesus!"

I keep hearing about...

[B-B]

This supposed GREAT laptop battery life you will get with the Crusoe, but I am not buying it. The processor is ONE component, among many. Even the P3 is not the biggest hog in a laptop.

Open Ports, the LCS, the DVD/CD and the HDD all burn MUCH more power. So BFD, even f the Crusoe gets by on 1/30 the power of a P3, it will not translate into more than an extra 1/2 hour. For chissakes, the G3 gets by on 5 anyway, but the Apple does not have significantly more battery life than a comprable P3.

I am rooting for it. But like with Apple and Linux and other techs, there is more hype than substance.

Tom

Re:I keep hearing about...

[Mr Z]

Is this a Intel PC only concern? The only heatsink in here is on the Proc.

It may be. The PC mobo chipset has to provide all of the legacy support crap in a PC. Apple's been a little more successful, I think, in throwing off cruft in each spin of their machine, due to their tighter control of the platform.

Also, in the PC space, you have more rabid head-to-head competition between vendors, and most are competing directly on performance and cost, so they don't care if they burn a few extra Watts. The Mac can get away from that a bit since the competition within the Mac space is less cutthroat, and the competition between the Mac and PC spaces is an apples-to-oranges comparison (eg. you can't just post a 0.5 FPS difference on Quake 3 timedemo scores and knowingly point at the mobo chipset as the reason).

In the laptop space, this is of course less of an issue, although all of the legacy cruft still gets dragged along. Transmeta's software solution allows them to also virtualize much of the lesser-used legacy crap in software, thereby giving them even better bang/buck and bang/Watt ratios.

How much power could this draw, though, vis-a-vis the LCD or the DVD drive[?]

Not really sure. My wild-ass-guess is that the PCI stuff could burn as much as 50% as much power the CPU does in a laptop (much lower ratio in a desktop), but it's really just a WAG. How that factors into the overall system's equation is even more of a mystery.

You're right about the display being the really "sucky" element. I can't wait until low-power LCD goggles are available. (I know of someone who's working on such a toy...)

--Joe
--

Re:I keep hearing about...

[Mark F. Komarinski]

According to Intel, a 400Mhz Celeron PPGA uses 24.6W. According to IBM, their Travelstar 14GS (14GB 2.5" drive) takes a maximum of 5W, and that's for powerup. Normal is on the order of 2-3W. Don't know about the LCD, but a Teac notebook CD-ROM is about 3W (600mA * 5V) while active. Sharp 12.1" LCD screen takes 7.5W.(see intel.com, ibm.com, and teac.com, sharp-world.com for the specs).

You're now looking at instead of a draw of about 40.1W with an Intel Celeron versus a draw of 16.5, or a drop in almost 2/3. This would imply that all things being equal, you'd triple your battery life by going from a Celeron 400 to an Crusoe. And the Celeron 400 is pretty low power compared to the PIII chips.

-Mark

Dell & Compaq are missing the point

[ender-]

It seems to me that Dell and Compaq are completely missing the point of the Transmeta chips. All the portable chips lose performance by being made for portable use. But the point of the Crusoe is to lose a *little* bit more performance and in exchange get a huge increase in battery life.

Personally I'd kill to have a laptop that performs like say, a PII300 but has 7+ hrs of battery life [that isn't an apple :) ] as opposed to a PIII 700 that only has 2.5hrs of life. I don't use laptops for that much CPU intensive work. They aren't designed for it. They're designed to get some work done while travelling [Ok, some games too] A 2hr battery is useless on a 6hr cross-country trip.

I personally think Transmeta has a huge winner with this chip. Sure future versions will probably increase performance but since what they provide NOW is totally adequate, anything more is just icing on the cake.
And of course that's not even considering the cool web-pad like devices that can be made with them... :)

Hitachi, IBM, NEC or Fujitsu will almost definitely get my $ when these laptops are available.
Ender

Performance outstrips productivity in the office

[Outland Traveller]

Honestly, aside from the engineers no one in our company needs the performance of even the lowest end of Dell's CPU offerings (500mhz).

We habitually buy the cheapest CPU's we can get, a little extra RAM than usual, a moderately sized HD, and the vast majority of our userbase is happy. No one's office software comes close to maxing out their CPU, and even the lowest end systems now have respectable 2D video performance well capable of marketing presentations. I wish we had the option of buying even slower-CPU systems for a reasonable discount but alas, chips only get so cheap...

Unless you're a designer/engineer/scientist/gamer, you're not scratching the capability of the CPU, and if you are you need to kill winamp, your software dvd player, SETI, and l0pthcrack and get back to work :P

Before you flame me keep in mind I'm talking about normal corporate users, not enthusiasts.

-OT

What's wrong with gcc?

[slothbait]

First, I'm going to agree that writing "native" code for Crusoe doesn't really make sense. Vital pieces of architecture are implemented in low level software, so if you bypass that layer, you are left with an incomplete microprocessor. Plus, your compiler backend would only be useful for that *particular* rev of the Crusoe. That layer of abstraction allows Transmeta to reorganize their architecture far more easily than other chip makers who have to worry about compatibility at the silicon level. I have no doubt that they will be using that advantage.

But I want to ask about another part of your post.

I keep hearing people snicker at gcc, but I haven't heard any solid arguments as to why it's "not that great". How is it inferior to other compilers? Does it not optimize as well? Is it less efficient generating code on certain architectures? Is it not as ANSI compliant as we might like?

My experiences with gcc have always been pleasant, and I use it instead of a vendor's system compiler whenever I can. Perhaps I'm biased since I've encountered way too many Sun's that still use ancient, pre-ANSI C compilers. No doubt, well maintained system are in better shape, but I'm wondering how much better than gcc they really are.

Please fill me in.

--Lenny

Re:What's wrong with gcc?

[Mr Z]

I personally find that GCC tends to be the same or better than the SparcWorks compiler on "integer" control codes, but it sucks wind on floating point codes. (GCC is also simply a much more forgiving compiler.) In most of the other anecdotal evidence I've heard about other platforms (Alpha, mainly), it's a similar story -- integer is ok, floating point sucks. Here on Slashdot, we've seen multiple stories about Compaq's Alpha compiler and math libraries and how they outperform GCC. GCC 2.95 and newer might be better given several recent developments, but I'm not in a position to test it presently.

Until recently, GCC lacked several optimizations, such as software pipelining and even accurate pipelining modeling for straight-line scheduling, so it did a poor job of keeping highly parallel pipes full. (Note that on x86, this isn't a problem since those CPUs have relative narrow pipes these days.) For instance, GCC's architecture description primitives weren't expressive enough to describe how to order Sparc instructions so as to generate a schedule that would issue four instructions per cycle. Many fixes have occurred as part of the EGCS / GCC 3.00 project that's now in progress, but that's not the mainstream GCC currently.

These days, vendor compilers are pretty good, and GCC is reasonable. When GCC 2.0 came out, GCC pulled ahead of many vendors, as I understand it, and in the meantime the vendors have caught up and/or have gotten ahead. I'm hoping that with GCC 3.0, GCC pulls ahead again.

PS. To give you an idea of how bad GCC is on some platforms, at least for awhile, the native build of the Alpha RC5 client was SLOWER than than the Windows version of the client running in FX32!. IIRC, the native version was a GCC build.

--Joe
--

Native Code

[JabberWokky]

.
I'd like to see a Linux Crusoe notebook and a gcc compiler that can generate native Crusoe code. That would be cool, I think.

-Sigh- OKAY... One Last Time -- there IS NO native code set for Crusoe chips. There IS NO point in coding a compiler for the layer of code under the morphing layer.

Two chips have already been released, and they have wildly different "native code" layers. Chips with the exact same model number might have different "native code".

Here's a little history to explain where they are going - Back in the Bad Old Days of computing, the computers had a circuit board covered with transistors, resistors and such for every single function. When you "programmed" those beasts, you were actually telling it very specific instructions on how that *one* computer should do things. Each and every model was different, often with significant differences between each *machine*.

Then Computer Languages came along, and made people happy for awhile. Although you had to write a compiler/interpreter for each machine you made, the programs could work across different machines.

Opcodes, and Integrated Circuits came *later*. Once they showed up, you could change the underlying circuitry and still maintain compatability, since you were programming to an "Instruction Set" that would stay the same across that family of processors.

Your beefy 750 Mhz Pentium III has a set of opcodes in it that mimic something called the Z80, used in CP/M machines (and several machines like the Commodore 64)[1].

Okay... so now, Crusoe is attempting to abstract the circuitry yet another layer with it's intelligent "code morphing" layer. In theory, the CM layer can do all sorts of loop optimizations, creating new structures that perform faster and faster the more times they are used.

At least, that's the theory, and what they present on their site and in their presentations. I don't know how real it is, as the benchmarks seem not to reflect what they claim.

But the point it - if you write natively, your program won't necessarily work on another machine *with the same chip*, and certainly won't work on any other Crusoe chips. Plus, you eliminate the advantages of any optimization that the Code Morphing layer delivers.

Think of it this way -- all the people clamoring for "Native Crusoe Code" are basically doing the same thing as people asking video driver programmers "Can you bypass those 3D and 2D accelerator chips so we can use the normal VGA registers? Yes, we know we'll have to write a seperate program for each video card."

[1] I ported over some CP/M utilities from an Apple ][+ with a Z80 card to an original IBM PC. I still have that PC, and years later looked at it's serial number. 512. Yeah, baby.

--
Evan

Why not PPC?

[tbo]

If you want a fast, low power-consumption chip to run Linux, why not a PowerPC? Sure, it's not x86, but there's a number of good PPC linux ports.

Yes, PPC has higher power consumption than Crusoe, but much lower than any equivalent x86. How do you think the new iMacs work with no fan?

From the article, it sounds like PPC would give you much better performance...

it's the form factor, stupid!

[webmaven]

I'm not waiting for a laptop.

I want a web-pad with a touch screen (optional keyboard) that has wireless Internet access form anywhere within my home.

The lower power consumption will make this sort of device a lot easier to hold and use due to a smaller battery.

If the connection was high bandwidth enough, it could even be running applications off a server in my home, and it could dispense with a hard drive entirely.
--

Dell says: "Linus who?"

[mwalker]

"Yahoo! has an article from ZDNet News that details how Compaq and Dell are shying away from Crusoe notebooks "for now".

"We just can't afford to anger our pimp." Said Dell's spokesperson. "Microsoft has always been good to us, and he doesn't like that ho Linus!"

"When I can give the customer full-size notebook performance," said Kyle Ranson, vice president and general manager of Compaq's Transactional Business Segment.

Kyle emphasized the importance of full-size notebook performance, noting that most customers expected a high framerate from their toy paperclip, and the ability to fry eggs on their CPU for all 45 minutes of battery life. When asked if he had considered the needs (and lower cpu requirements) of Linux users, Ranson responded:

"We're going to be taking care of those people shortly. Er... I mean... I'm not sure that Linux will be favorably recieved by Microsoft.Net."

Cool running processors

[Dungeon Dweller]

In my opinion, these processors are better suited to be the high power processors of yet smaller devices. The laptop industry has moved beyond the speeds of the crusoe processors, but smaller devices where more power is desireable are really where the 1st generation crusoe's belong. A couple months ago, though, the crusoe's would have really given Intel a run for their money (they still are with the whole power consumption deal, but I am willing to bet that the GHz notebooks will be out soon). Also, in cost, they seem to have Intel beat in value for the buck.

Does performance really matter?

[First Person]

Technically, the chips are intriguing and appeal to the nerd crowd (myself included). But even if they perform at 50% compared to the latest laptops, how many business travelers would want one solely for the increased battery life? I know that when I travel across the US, to Asia, or Europe, I'm nursing the batteries all the way. And I'm only running 'simple' applications.

Given the hype and counter-hype, these may be the most eagerly awaited benchmarks in recent memory. It's great to see competition at this level.