Intel's Pentium M clockspeed dilemma solved

A BUMPER FRIDAY letters page for you starting with Intel's Pentium M clockspeed dilemma...

Here's an idea for intel (let's hope they don't take it seriosly).
Why not add a 2/3 clock devider in the chip and increse the external clock by 50% ?

Intel *could* then claim the chip runs at 2.4GHz.
After all any CPU has idle clock cycles - just they are not intentional usually ;)

Rather than normallize the CPUs to a PR rating, normallize them to the clock speed.

Come to think about it, I'm surprised they didn't do that to begin with.

Assaf

MS-FUD " Can Linux break through to the commercial office
desktop? "

This Special Report rehashes many long-discredited claims which have been made by Microsoft-funded 'independent studies' in the past. I am inspired to "flame" you, (for my very first time), because of the particularly offensive statements which claim that a commercial firm may have more "Confidence in the Supplier" of Windows. These 6 paragraphs complain at length about risks with Linux, but don't present even a SINGLE SENTENCE about the risks in depending on Microsoft as a supplier of support. Here are a few of the most obvious issues:

(1) Microsoft is the sole supplier of such "support". They may charge you as much as they like, and they are free to provide whatever level of effort THEY choose to provide, because you can't go anywhere else. In contrast, you may obtain support for Linux and other GNU/GPL programs from a variety of suppliers, or perform your own S/W Problem Analysis and Resolution in-house.

(2) 'Look But Don't Touch': Even if you were to subscribe to the 'Shared Source' program, you are given no right to modify the code for your own use. In contrast, you (AND your third-party support vendor) areare free to modify Software covered by the GNU GPL as you like. You do not have to offer the Source Code for these changes to the public unless you distribute them externally.

(3) The assumption that the 'Shared Source code' which Microsoft chooses to provide truly built the BINARIES WHICH YOU ARE RUNNING is a leap of faith: You can't build and compare the Windows binaries from the provided 'Shared Source'. (And Microsoft officials have never attempted to deceive customers or Courts in the past, right?)

Overall, I feel that the following quotation provides a good indication of the author's competence on this subject, and his real objective (i.e., of spreading pro-Microsoft FUD, even if he needs outright LIES to accomplish this):

"The unfortunate downside of this is that computer companies such as IBM, Dell, HP, Sun and even Oracle others may "fork" the Linux code base and modify it to suit themselves....".

(4) These companies may NOT distribute such a "forked Linux" to other firms. The Author is obviously incapable of (or unwilling to) understand the GNU GPL and the change control procedures which are in place to manage GNU/Linux (tm).

I feel that much of the other 'recommendations' and 'analysis' in this article consists of similarly worthless trash. I have focused on only one section. Feel free to post or forward my complaints, and please feel free to edit them for clarity or length.

Your appreciative (and only seldom disappointed) reader,
Rick Stockton

Re: Can Linux break through to the commercial office desktop?

To the author of this article:

No, I wouldn't suggest you move to Linux, either.

Until you and your IT staff are educated, you'll be scratching your heads about Linux with fear and bewilderment.

But I suggest you 'watch your back' for some young upstart who DOES know how to run an organization with Linux. Then your rather vague points about why not to go Linux will become moot. So will your job.

In other words, don't let your company's competition get Linux first.
Email address supplied

Cooler confusion

At the time of the release of the new packaging for the athlon xp, with the bundled cooler, I emailed in to suggest that the cooler supplied may be dubious, because it only connected to one of the three clips on each side of the socket. I had seen an article on Tom's Hardware about thermal overload, and it had mentioned this as a potential problem.

Well, I take it all back.

I recently installed a Cooler Master X-Dream into a users PC, which uses a clip which connects to all three of the points. 'It's dead' comes the phone call the next day. 'It just turns on then off again seconds afterwards.' So I get the PC back - the cooler has fallen off. My mind flicks back to the whole 'thermal overload video' debate. I prepare for the worst. It's not to be - the PC is fine after having the cooler reseated.

I've decided that computer land has no morals, so I will instead award points on an arbritary basis:

2,000 points to MSI for making motherboards that have never let me down
1,500 points to AMD for reacting to the thermal overload issue.
-1,498 points to AMD for making processors that require such huge heatsinks.
800 points to VIA for taking the 'fanless' route to thermal elegance.
-1,000 points to Cooler Master for making such a rubbish clip (flimsy imho).

I am definitely beginning to think that if VIA could position themselves a bit closer to the pack in terms of performance while maintaining their minimal power usage / heat dissipation, then they'd be on to a severe winner.
Roll on optical processors....

Matt

"My OS is more stable than your OS"

Since there's no such thing as a "best" OS, only a "least worst", I thought I'd deliberately, and with malice aforethought, throw some petrol on the fire, simply 'cos I could use a good laugh:

I have a Psion netBook which runs the EPOC32 (since re-released under the "Symbian" brand) operating system.

It has been running non-stop, without so much as a hiccup, for FOUR years, barring one single, (free) solitary OS update. I have not switched it off EVER, since it CANNOT be switched off without letting it run for 2-3 weeks to run both its batteries down -- it can only be switched to 'standby' mode.
It is absolutely silent in use. It doesn't have so much as a noisy fan in the case or a massive heatsink on the CPU. It has no moving parts, yet it can open MS Office files -- even dBase IV if I want to. And it doesn't even use an Intel, or Intel-derived, CPU.

(Oh yes, did I mention it's a notebook that will run for 8 hours between charges? It supports WiFi too. And it has instant-on, a touch-screen GUI, built-in office apps and even Opera 5 on ROM and more. And this is a machine launched nearly 5 years ago. Intel's Centrino chipset is merely playing
catch-up.)

It even has an FTP server running on it.

Windows? Linux? Pffft! Neither is fit to even lick EPOC's boot code.

:P

--
Sean

Re: White boxes set to take over the world

Jon Peddie is right. I am in the middle of poor, down trodden telecom valley north of San Francisco and poor, down trodden silicon valley. White box computers are booming. I build 2-3 computers a day and never sit on any inventory. I stay current on what hardware is stable and in it's prime so I have Zero support costs. I do not advertise. Word of mouth say's you can get twice the speed of dell at a given price point from me.

I recently bid on a bunch of dual xeon workstations for a small time special effects house and delivered 10 workstations at $10k per box. I won the bid because Dell did not offer 1TB of drive space on a desktop machine. Profit was slim considering the huge value of the deal but after that they bought the farm, literally. They bought 42 dual athlons at $2.5k per from me for their render farm, $1k each was pure profit.

Where does that leave me? If I did not read about how bad the tech industry was doing, I would say it was boom times. Maybe it is boom times, just not in an area where the big corporations can show it on their bottom line. For that matter, I've read that Dell had a pretty good year thanks to our weed smoker friend. HP is choking on Compaq, serves them right for squashing Alpha in favor of Itanium. AMD is not profitable but that could change for them come April. Intel is profitable even though they keep bleating tech recession to their shareholders. Sorry Intel, no tech recession as far as I see. Only AMD breaking Intel's monopoly and with it their $2000 top end CPU and their $500 "budget" CPUs. Oops, margin is getting tight, better tell the world the market is dropping. Dammed irresponsible thing to do if you are the market bellwether.

As far as why the whiteboxes are doing so well?

1. Building computers is easy. No more IRQ conflict nightmares.
2. Even VIA and SIS make pretty stable chipsets these days meaning no more whitebox lemons.
3. Availability of parts to the enthusiast market and plenty of good web stores to flog them.
4. Great tech journalism pointing us all to the holy light of "wibble wound-up"

Doesn't that make the inquirer feel important that 60% of all computer buyers open their wallets based on new Caesars thumb up to ATI or thumb down to non-toys-throwing-out-of-oran, Nvidia. Hey, they make the best AMD chipset even if their Graphzilla status is now Graphdustbusterzilla. If Intel would let them play, they could become onboardgraphzilla and Intel could get out of that biz, which they are obviously not very good at.

Email address supplied

Re: Fee, fi fo fum, I smell the blood of an Intel erratum

The E7205 / E7505 erratum re: AGP 8X timing violations is VERY old news (several months now). More to the point, the issue is resolved by a relatively simple board fix that (to my knowledge) is incorporated by all shipping motherboards for both chip sets. That is why Intel isn't doing a turn of either chip set to fix this.

Sincerely,

Patrick Chase

A happy and successful user of an E7505-based motherboard and AGP 3.0/8X graphics...

Re: Intel app accelerator gives desktop PCs super boost

If you have been playing with ATA drives long enough, you will know IAA is a mimic (or the newer version) of the notorious Triones drivers. Basically, they are poorly written HDD drivers with cache, as if Microsoft is so dumb that it cannot supply a decent cache within the OS for 10 years that somebody else can do a better job in disk driver. The early version of Triones drivers get their fame by cheating in PC Mag Winbench, and after Winbench was revised most (if not all) the advantages are gone. Worse, these does not necessarily enhance speed, but definitely kill stability - they don't like IDE devices other than HDDs (and don't like some HDDs in certain combination either). From my experience, IAA + CD burning means immediate system hang. There is definitely good reason for Intel to not pushing these.

Geoffrey

I do like your comparison to cars, I have myself used it for years. More for software than hardware though. Mainly to show the cost difference for the car companies to recall and fix a large number of faulty cars, to the cost of a software vendor to put up a patch for download on their site...

Anyhow, I think the problem that the demands for better quality from hardware, and software, vendors are two. The first is safety; a computer that doesn't work properly will not injure or kill people, but a car might and thus its not as important to regulate as for cars. The second is that the technology is to complex for consumer organizations, or even government departments that are supposed to "protect the consumers, to even "dare" to try and do something about it.

I think though that it is possible to make a ruling guideline which vendors has to follow. Much like there is for the car industry. For example forcing them to make bug fixes and patches available for both current and previous version, and never say that the use has to PAY for an upgrade to get a bug fixed. Much like the car industry has to keep a stock of replacement parts even for cars that they stopped produce years ago.

The best way of putting pressure on the producers is probably that the governments and its institutions, including schools and such, then say that from a certain date they will only buy hardware and software from companies who promise to follow that guideline, let say two or three years from when it's launched.

Some might say this might hurt the Open Software movement. I on the other hand think it will be the opposite. Specially considering how rapid bug fixes are released, often overnight, compared to some companies who doesn't seem to care unless the media steps in and ask questions why "known" problems aren't fixed.

Something clearly needs to be done since the technology just get more and more complex and the vendors doesn't seem to have procedures to fix problems other than forcing people to upgrade, and thus pay these companies more money.

Regards,
/Thomas

I read Samuel Clemen's article "Fee, fi fo fum, I smell the blood of an Intel erratum." I assume he is referring to the second errata, which deals with the AGP prefetch cache. Just out of curiosity, why does this upset him since the specification update states that this errata does not affect system performance?

I could understand his discontentment if the BIOS automatically had the feature enabled on his system, and he obtained the motherboard as part of a system from an OEM such as HP, Dell, etc. If the BIOS automatically disables this feature, however, then I don't understand the harsh response seeing as how this errata affects neither the system's performance nor the user's experience. Moreover, if the BIOS automatically has the feature enabled, but he purchased the motherboard separately and built the system himself, then I still do not understand his resentment. I say this because if an individual has the technical knowledge to order and assemble the parts for a new system, then I doubt they would be fearful of making changes in BIOS.

Could you please shed some light on his perspective for me?

Thanks,
Jeremiah

Re: Microsoft DirectX killing innovation

Ironic...
I was pondering the differences,births, and reasons for these two API's,(Yes, I have no life. So what?) DX and Open GL. I've often wondered why anything besides the FPS genre has not really used Open GL, as the performance of it seems more stable, and just outright better, than it's DX counterpart. Perhaps, you would care to elaborate on that...

One thing I didn't like or agree with, was your open-ended suggestion of there being different cards, with different features, for the same games. Why in the hell would I want to have to buy 6-8 cards just to see all the features realised in a game? (Hmmm...maybe I do have a life) For instance, if I have my R9700 Pro card, which can do Open GL and D3D quite well,it would be a troublesome, not to mention expensive necessity to get another card that does this, or does that better than that one. Think of the Mod community. It would be splintered into oblivion as certain aftermarket mod makers would have to choose one, possibly two options at most to make a mod for, thus leaving the don't haves without any mods or enhancements for their game.

Anyway, I think there needs to be one graphics API that can do all the effects under the sun and then some, and that API needs to always be in refinement. To an extent, Microsoft seems to be trying to do this.

Alot of this article just strikes me as the typical Microsoft bashing that always goes on. Microsoft doesn't make perfect products. No one does. This bashing is done mostly by the typical, mal-adjusted fat sweaty slobs, who have not learned how to communicate, but only to complain.

That's my two cents...
BMG

Doesn't Microsoft work with the hardware vendors, nvidia, ati, matrox, etc to come up with the next generation for directX?? Yes microsoft probably has a say in it, but its more of a collaboration between the various vendors and microsoft to come up with the new api.

Email address supplied

There is a problem with the article mentioned in the subject:

"Nvidia produced a card that had a strong DirectX bias, and was competing directly with the 3DFX offering. 3DFX, on the other hand, regularly beat the pants off nvidia in any OpenGL benchmarks."

NVIDIA has done a great job supporting OpenGL from the get-go, and beat 3dfx cards in both OpenGL and DirectX benchmarks. In fact, 3dfx got a lot of heat from developers for doing a "Quake-only" OpenGL implementation, whereas NVIDIA (which by the way has a ton of ex-SGI talent) did a very full-featured implementation. 3dfx also had other major problems like late, expensive products that didn't support 24-bit color.

For the last several years NVIDIA has had a deserved reputation for excellent OpenGL drivers. Even today, (AFAIK) NVIDIA is the only vendor who's Linux OpenGL scores are very even with the Windows scores. NVIDIA simply does a great job of supporting both DirectX and OpenGL.

No, I don't work for NVIDIA, nor do I have a business relationship with NVIDIA... just give a good company it's due.

Anon

You're right, the excitement is mostly gone from the gfx industry. I remember plugging in my brand new voodoo3 card and being blown away at how much faster and better it was. It's been a long time since I felt that way. Somehow, and extra few zeros on the max polygon rendering count just doesn't have the same effect.

While I'm not one to defend MS, I think this is more of an issue of the gfx market maturing, not necessarily the MS is taking over. Anyone could have seen - back then - that a standard *had* to be developed. If MS didn't do it, someone else would have. Even if OpenGL had really taken off, we'd all be following the beat of the OpenGL Consortium, or whoever was holding it.

It's sad that the excitement is gone, but PCs are becoming a commodity, and will soon be just as boring as pork bellies and plumbing. Luckily though, as PCs reach maturity, they will also become more reliable and stable - so you can count on them as much as taking a shower.

-- Brian

In a recent interview with CPU magazine, The project manager for Direct X 9 stated that it would rule the roost for the next 5 years at least, because Microsoft was slowing it's development cycle. Microsoft perceives an absence of competing specifications and chipmakers are being led by the nose. When OpenGL and Glide were in the race, Microsoft released a new Direct X every year. After going from Quake 2 quality graphics to Nvidia's fairy quality in the last five years, we can expect to be stuck at fairy quality for the next 5 years, albeit at ever higher frame rates or resolutions.

In a competitive market, if the fairy was Quake 2, 5 years would give us true photo realism.

5 years will give us 500 million transistor GPUs and two more generations of id game engines after Doom 3. John Carmack has shown a complete willingness to advance the OpenGL API any time Direct X did not meet his needs. Microsoft thinks they have no competitors, but could find OpenGL incrementally improving beyond Direct X performance and card manufacturers taping out chips with OpenGL performance at the top of the list instead of Direct X performance.

A hidden player here is DRM. While the stated purpose of DRM is to protect copyright, it could also allow Microsoft to completely close the pc platform to any competing technologies, just as the test bed for DRM, X-box, has shown.

It seems like a complete win for Microsoft except for two competitors, Apple and Linux. If developers begin to feel creatively constrained by Microsoft's platform and it's lack of photo realism, they may move on to greener pastures, to be found speculatively in the OpenGL of 2006. Linux and Apple are not big money game platforms so it is unlikely the large publishers will ever support them but as the mod community has shown, some of the best games are done for no money at all (initially). An interesting twist to this is that given id's track record of open sourcing older engines, Doom 3 should be open source before Microsoft advances to DX10.

Simply on technical merit, if a GPU is quadruple the clock speed and 4 times the transistor count but has to push the 8192x6144 native resolution of proposed LCD technologies, our fairy is stuck at 30 fps, which would require hardware and Direct X support for motion blur to make it playable. DX9 does not have that capability. Additionally, PCI Express opens up the possibility of multiple video cards interleaving frames in the same way that 3DFX executed on the PCI bus. Manufacturers will not be limited to a single AGP slot, also not supported under DX9. Whatever chipmaker is behind the power curve when PCI Express debuts will likely use this approach, breaking them from the Direct X fold in order to compete on frame rate.

Just as games drive desktop hardware performance (and recent mobile performance for that matter), games drive OS adoption in the home market. If Microsoft rests on their monopoly power in the gaming market, they will cease to be a monopoly. If everyone is using a Unix at home, and corporate servers are predominantly Unix, how could Microsoft expect to maintain a corporate desktop monopoly?

Email address supplied

One other thing you might also have mentioned in your article is that Microsoft recently removed itself from the OpenGL ARB. I'm not quite sure how to feel about that ;-) but it clearly shows that Microsoft would prefer OpenGL (just like every other open thing in the universe it seems) to die a quiet death. Microsoft would apparently like to hone DirectX and lure developers into single-platform APIs. That doesn't seem to make much sense for a lot of software developers, thank goodness.

I hope NVIDIA and other graphics vendors can keep OpenGL a viable cutting-edge graphics platform, even on Windows. Portability across many platforms, including Windows, Linux and Macintosh, is a nice benefit when doing OpenGL development.

Terry

Hey Mageek:

Please get the hell off the Iraqi War and back to coverage of technology news, where you folks belong. Your choices of what to cover on the war, and how to report it - well, frankly they stink.

I come to the Inquirer to get my 'daily ration' of technology news and commentary. If I want to hear about the war in Iraq, I'll turn on the 'boob-tube'. If I really want "bleedin' 'eart" crap like that bit about the Iraqi TV station, then I'll surf to some peacenik site.

Do understand, I'm not kocking you or your people; but you folks are technology journalists, not war correspondents. Stick to what you know best and do very well. Leave alone what you know damn little about and don't know how to handle.

Thanks for letting me rant.
Dave

Re: Canadian copyright levy

The last I heard, the jokers who have been collecting these levies had paid a grand total of $0.00 (Canadian) to the so-called "eligible" people, while they sat on a bundle of ill-gotten gains. A few months ago, they were making noises about paying some of it out "Real Soon Now" to stop the grumbling from the ranks. I didn't see "eligible" being extended to indie artists and others...

I strongly support having musicians (and other artists, which the Copyright Board missed, methinks) making a fair living, but the Canadian Copyright Board approach is not the right approach. It rewards the wrong people, using money taken from the wrong people.

I view the Board's approach as along the lines of this non-existent (so far) story: Canadian banks were having trouble with people stealing their money, due to poor locks on their vaults (and a introduction of socio-economic system that tries to emulate that of the U.S., but that's off-topic). So, the Canadian gummint, in their infinite wisdom, decides to solve this problem by placing a levy ("It's NOT a tax -- that makes it OK.") on all raw metal and paper sold or manufactured in Canada, giving the proceeds to the banks. The fact that the metal might be used to make
a nail, or the paper used to make toilet paper, was not considered, any more than the Copyright Board considers the use of said media.

The actions of the Copyright Board are extortion and piracy, pure and simple.

Yours in indignation,
Robert Bernecky

RE: Chris Davie's response to the Canadian Levy

To the INQUIRER staff,

In his letter, Chris and no doubt others, expressed outrage at the Canadian government for the the effectively doubling the existing levy on CD-RWs and other recordable media. It should be noted that "other recordable media" also applies to Mini-Discs, portable hard disk drives ( i.e. IBM microdives ), Smart Media ( flash memory ) and others. After filing a letter of complaint to the Canadian government, I later learned that music industry representatives met behind closed doors with government officials to discuss sales figures and statistics of album sales in Canada. The Commission in charge of the levy was to make all facts and arguments and letters of complaint public. This was clearly a breach of this trust.

These levys are more far reaching than one can imagine: it punishes independent music artists, companies and consumers making backups of data on computers, consumer and professional photographers ( because flash memory is deemed to be levyied ). It doesn't matter if you buy one CD or thousands, the money from the sales will go to the Canadian recording industry unless one has a permit to import. To date, 10 blank CDs cost approximately $10 CAN with the cases; less of course for the "unmarked" brands. With more levys in place the cost will increase to about $20.

It should also be noted that some $20 million dollars have been raised in the last few years from these levys, to date as reported in the media, none of the monies raised has been distributed to artists. What then, is the purpose of these levies?

Regards,
Jason

Re: US corporations possess less than zero imagination

I was a "formal objector" in the proceedings for the 2003-4 levies--this has been around since '99, they're "reviewed" bi-annually--and there hasn't been a decision yet, it's been almost a month since the hearings ended. My focus was to try to keep the levies off of anything with a hard disk, as there are proposals intended to levy certain portable computers referred to as "MP3 players."

Perhaps the most offensive development of late in our levy system is that the private entity--it's not part of the government, but it was created by the law that set this up--that collects and administers the levies ( www.cpcc.ca) realizes that the concept is fundamentally bogus and untenable, so to keep it going at all costs they are going to offer exemptions to pretty much anyone who asks(who could afford a lawyer,) if
they don't object to paying a fee and opening themselves up to being audited and being told where they can obtain their small plastic disks--much like anyone who works with deadly microorganisms or restricted weapons. See http://www.cpcc.ca/english/zeroRating.htm.

Jim

Re: Blueyonder trials 2Mb cable modem service

Blueyonder 2 MB trial

Signed up for the service last night.

Christ on a bike this is fast, smooth and packet loss free.

Check out the trial newsgroup, there are some disgruntled folks that didn't sign up in time (ah well)

But I was all set to go just 10 minutes after registering.

We had a fairly detailed marketing questionaire to fill out. Do we use Kazaa regularly? Do we download software and games regularly. How this squares with copyright wasn't quite explained...

Then there was a bit on pricing. How much would we be prepared to pay ON TOP of the existing price. Well 1 MB is thirty five quid and in the options box the absolute minimum we were allowed to tick was an extra 14.99 on top of the existing price. It doesn't take a rocket scentists to work out the price is going to be 49.99 per month. (you watch, I'll be wrong now)

Is it worth it? Hard to say... depends how much I start kicking and screaming when they try to take it away from me in a month's time. I must warn BY I may be armed by the time that day comes around.

Email address supplied

I do not know how they handle cable TV in the UK, but here is what we are getting in the States.

In my town (Framingham, MA, USA, thirty miles West of Boston), one of the two local cable companies, RCN, is offering a three megabit upgrade service for free to people who have been using their bundled voice/data/CTV service for more than one year. Their normal cable data service is 1.5 megabit.

When they built their network, they put in a ton of extra capacity- I believe it is OC-3 to the local poles with a multi-layered OC-48 backbone. According to them, they are providing two voice lines (up to six lines per home), digital CTV (250 channels at current) and data (1.5/3 mbit) and are only using about 6-7% of their network capacity.

Cost is about $190 per month w/o taxes.

The other cable company, Comcast (formally AT&T formally Cablevision formally another company I can't recall) is in the process of upgrading their network to provide similar service.

Both companies will start sending HDTV/6.1 surround over their networks in the next few months (about eighteen) on those networks that support it.

On a tangent note, they are actually running out of US networks to broad cast, so they are staring to add networks from other countries. I can subscribe to networks Japan (NHK), India, Pakistan, the Philippines, a Farsi based US network (owned by the son Shaw of Iran), Greece and a few others. The BBC (time delayed so that the shows run at their equivalent times), Univision, Telemundo, Galavision, and the national network of Portugal are free. (our town has a large population of Spanish and Portuguese speaking people from South and Central America).

Jeremy

µ