Intel multicores, sexist games, and a heap more

I would like to thank Charlie Demerjian for his article "Intel goes on major multicore bender" and to offer a couple of more insights along the same lines.

We are all going to need a second core in our desktops and laptops of the future just to accommodate the needed work which must be done to support the I/O and security improvements we will demand in the future.

It should not be lost on you that the x-box is pioneering anti-virus/anti-trojan technology in the form of total digital signature validation and verification of each and every operating system module as it is loaded (remembering that the x-box operating system is nothing more than a stripped down version of XP). This technology, while primarily designed to protect Microsoft's intellectual property rights will be the basis in the future for Microsoft to verify the secure loading of Windows but will cost a huge amount of additional time on a general purpose system (on a gaming system you load a single program which uses a single set of services, once, not many programs running every combination of features Microsoft has ever thought of all of the time).

A second huge user of system resources in the future is storage management. Although we now have the ability of storing 400 GB of information on a single desktop drive our current management structure is the equivalent of searching the library of congress for a particular quote based entirely on a manual card catalog. It will give you some place you might try looking but then you have to go check each book individually for the quote, probably missing it because the person who made up the card had different thoughts about how to organize than you do.

This is why Microsoft would sacrifice a huge amount of revenue and even more professional pride with respect to Microsoft SQL Server by decimating that team in order to create a new filing system for Longhorn, one that is inherently indexable and thus you have the foundation to properly organize and retrieve information in the future. But running what is essentially a full blown copy of MS SQL Server is going to a again take up a lot of processor time and a lot more memory (note also any SQL database program is inherently 64 bit and since we are going to have it as part of the desktop environment you now have Bill Gates waxing poetic about 64 bits. It really is a better idea, finally an integer size which is truly large enough for most applications).

The Chipster

Republicans

A check of Federal Election Commission disclosure reports through July 22 by WorldNetDaily confirm neither HCL Technologies nor eServe received disbursements from the RNC. Answerthink and Computech, two U.S.-based companies that have joint ventures with eServe, are also not listed.

The Business Standard did not name the Republican National Committee and reported only that the "U.S. Republican Party" had contracted with the Indian firm.

Everything noteworthy in the Asia Times article by Siddharth Srivastava is "strictly off the record." As far as the HindustanTimes, yeah make of it what you want.

The Kerry Campaign making an issue of it? After a few pages of Google I could only find his "Town Meeting" web site in which a member site's the Hinduan Times report, and Kerry chime's in with one post.

I'm scratching my head throughout all this wondering what's going on? Nobody in the US has talked about it. Well, I should say noone of significance in the mainstream press, most of whom would love to see Kerry in the White House. I doubt anyone directly associated with the Kerry campaign will be foolish enough to go mainstream public with it. I'll let you know if I see it in the Enquirer.

I enjoyed the article by Nick on the hackers hacked. It's a small taste of justice and doesn't surprise me :)

Steve

BeOS

Again, BeOS gets no respect...

"There are all of no consumer grade programs that are multithreaded."

There are/were PLENTY of "Consumer Grade" multi-threaded programs out there for BeOS. and they're mostly all set up to support SMP by the very nature of the tools used to make them, which forced you to support it. While you may say that BeOS is dead, I'd point out that the BeOS Community is very much alive and growing again around OpenBeOS & Zeta. BeOS is still being used in some Commercial equipment, and at least one BeOS program has continued to enjoy Commercial success (TuneTracker automated radio station software).

Also, don't forget that OS/2 is still very much in use throughout the world in spite of all IBM and Microsoft have done to kill it. Multi-threading was a big selling point with OS/2 and it's preemptive multitasking. There was also an SMP version. eCommStation, the OS/2 morph, also has an SMP version.

(I'd be writing this on my dual P3 BeOS machine, but it's early here, and I'm on a SuSE 9.1 laptop.)

Bob (OS/2, BeOS, Linux user)

net.wars and ID cards

Dear Wendy:

I really enjoyed your latest installment of net wars on the endlessly recycled proposal for biometric national identity cards. I can only imagine that as the WW II generation continues to expire that the average bloke will know less and less about the great battle to defeat fascism that killed tens of millions and was supposed to make the world safe for democracy. "May I see your papers?" -- spoken with a German accent -- anyone? Instead it appears we must now fight endlessly recurring battles to make "democracy" safe for the world.

In an era where the world's most powerful democracy pulls a Third Reich by inventing a pretext to invade and occupy an impotent but strategic country; where self same democracy routinely classifies ALL government transactions and says to its citizens "just trust us"; where it routinely 86s its own citizens, suspending indefinitely the writ of habeas corpus, it is not too surprising that the world's oldest extant democracy wants to follow suite. It would be so sad if it weren't damnably, laughably entertaining -- that is, if I were 96, had no offspring and didn't give a rat's ass about the fate of the world. Double plus good, what?

Your observations on Britain's increasingly obsequious kowtows to all matters American reminds of a sci-fi novel I once read where England had seceded from the United Kingdom to become the 51rst American state with the royal family fleeing to set up shop in Australia.

A most enjoyable read. Keep them coming,

Raymond Cranfill

More on Booth Babes

Here.

Ya. well. rather than point out to this guy that in the techworld "pimp" and "journalist" are functionally synonymous, I gotta agree with him and say that after 30 years around this biz the inflatable dolls are still one of the nastier accoutrements of the industry. They get in the way and their disinterest is palpable. It's almost like an intentional insult to the hacks (and the dolls) by a marketing industry that largly fears and detests them.

I dunno about comedians, but hiring the cheerleaders on their game-scores or tatoo count would probably yield more positive results than using a high silicone/carbon ratio as an employment indicator.

Dhu

All Computer having a laugh

Mike,

While I have no particular affection for Intel, upon reading the All Computer patent cited in your article, I have to conclude that All Computer is All Wet.

They appear to make two primary claims.

1. They claim to have invented the idea of hijacking a processor's I/O pins with a second (more powerful) processor. This doesn't seem very relevant for today's products, and it isn't clear to me that it was all that relevant years ago. If Intel ever did do this (did they?) it was certainly before the date of the patent -- 1996.

2. They claim to have invented the idea of synchronizing a processor's clock with a multiple of a "sub-harmonic" of another processor's clock to get them both into synchrony. This is done all the time. Anybody with a valid patent on such an invention could really rake it in (or trade it for valuable prizes.)

But the concepts behind this invention are well known and have been well known for 20 years.

First, we'll take the idea of using a PLL to synchronize one processor's clock with another processor (or a master processor/clock). This was the scheme behind the clocking system for the VAX 6000 product line (introduced in 1984 or there-abouts) and may well have appeared in products from Digital and other vendors before that. So, that claim goes out the window, unless the 1996 patent was an extension or modification of a really really old application.

Second, let's look at the idea of using a "sub-harmonic" (a really silly term) to drive the PLL. What they've described is a dividing prescaler on the reference clock input to a PLL. This is a feature of almost every frequency synthesizer design going back to the time of Moses. It was certainly old hat by the time Floyd Gardner published "Phaselock Techniques." (The second edition came out in 1979, I'm not sure when the first edition was published.) The application of a prescaler to a processor clock circuit was well known to any practitioner of the art, circa 1980 if not earlier.

Perhaps there's something in this patent filing that a lawyer could find to sue over. But I can't. It seems to me that this is yet another instance of some well meaning competent engineer patenting the wheel, since he'd never seen one before. The sad thing is that the patent office is apparently so overwhelmed that they routinely grant this kind of patent. The litigating attorney should know better, however.

Intel should spank All Computer's lawyers and send them home.

matt

E-Festival

Hi Mike,

May I say I'm disgusted with the organisers of E-Festival. I found out firstly from a journalist on the phone this morning and then from you via the Inquirer that the show to which I have already pumped thousands of pounds other than the costs associated with the stand space itself.

I've sent them an e-mail expressing my dismay at the lack of common courtesy shown and still no response. If the event is postponed, this complete lack of courtesy toward the financial backers of the event - that is the exhibitors, will I am sure not bode well when they come begging for the coffers next time around. The kind of taste this leaves in a manufacturer's mouth when they have spent out on budget, used many man-hours in preparation and have to turn around to their finance department and say there will be a definitive zero return on investment - is much the same as taking a mouthful of runny shit from the arse of a rottweiler.

I am sure that there are many more manufacturers with the same outlook as myself in respect of this. I spent a whole 30hours without sleep building a customised system that cost no less than £2,000 - and what for ???

UK Vendor
Name, address supplied

Booth Babes again

Charlie D. was correct in stating 'whoever hired them picked right and knew the audience' because the real audience is the geeks not the geek press (no offence, mates). Further, since the geeks outnumber the press at the show by 100:1 only the most shockingly self-centered and self-important journalist would believe they were personally impugned by Charlie's hangover inducing exchange with the nice young lady at the booth.

Cheers.

Name supplied

All Computers doesn't have a claim

You people should know that in my opinion All Computers Inc, of Toronto Ontario Canada does not have a valid claim and here's why, though I'm no expert in law, and one can see said company is looking for money for nothing.

Pentium cpus have been out since about 1993 I , and all the patent details, is an accelator board for putting a faster cpu, on to a motherboard such as a P4T for example. I read the patent, and no where does it say anythhing about technology used in a cpu.

Accelerator boards as described in the patent for putting a faster cpu on a slower motherboard, have been around for years, matter of fact they've been around since the time of the 80486 microprocessor. Companies like Evergreen Technologies have been making said accelerator boards, since the 80486 cpu have been on the market well before the patent was filed in 1993.

I honestly have to question, which you guys should be questioning too, why isn't All Computers going after companies like Evergreen Technologies, this patent of theirs has absolutely nothing to do with cpus, and everything to do with cpu accelerator cards.

Tracy Klein

All Again

I don't see how intel violated that patent.

That patent concerns a CPU using a faster bus plugging into a motherboard with a slower buss.

And a special synchronization has to be made so the signals coming from the fast cpu in in sync with the slower motherboard, and I guess provisions are made that force the fast cpu to talk to the buss slower (using wait, buss busy etc).

But in the past, cpus ran at the board speed (buss speed)

The only thing I can think about why intel is being sued, is for the circuitry inside its cpus that allow the cpu to work at a ratio from the board speed. If this is true, it doesn't only concern the P4, but the 486, Pentium etc.

The weird thing about the 386, is that this patent mentions a 48Mhz 386, but didn't a 386 have a 16Mhz bus?

This patent was issued at a time where there was a 386, and I remember my friend having a 40Mhz 386. (runs faster then bus).

So, I don't think this patent concerns a CPU running at a faster clock speed from the bus, but rather a cpu plugged into a daughter board, that is plugged into the motherboard socket. This daughter board utilizes a protocol to ensure the processor talks at a slower rate (by making the buss busy signals) and making sure the signal is in sync with the motherboard bus.

The only thing I can think of that fits into a daughterboard is a Pentium 2. But that daughter board runs at the same board speed and the cpu runs at that same board speed. SO no use of the technology is present here.

I think the patent is trying to use an aspect of its PLL (phase lock loop) technology to imply to intel, because the protocol inside the intel cpu works similar to the one described to the patent, but not exactly. (there is no pins inside the cpu and some of the signals are not required).

Name supplied

No more domains, please

I can understand the two issues he brings up of companies having to buy the other .whatever to protect their name, and the enforceability of the .xxx domain - considering that Larry Flynt will be all over anybody who tries to shut him up. The real problem is indecent whining about the problem without proposing solutions. Now your article didn't say he had any, so I'll assume he didn't. If you can't come up with at least an idea or two that would be better - then sit down and shut up. I sure hope he is the father of the internet in the parental sense, not the religious - otherwise with deconstructive criticism like that, we're all screwed (pardon the pun).

And just to make sure I don't do the same thing... how about giving the suggestions from ICANN a try? Test it in the court systems and let it work itself out... that is the American way after all - sue till you get what you want!

(you can print first name, please omit all other information - thx)

Tim

Gainward and FiringSquad

Hi Eva,

I was recently notified that a representative from Gainward used the opening two paragraphs from one of our articles as part of an open letter/press release. Normally I wouldn't be concerned, but I feel the piece that Mr. Tismer from Gainward chose was out of context.

Our article was a revival of an old FiringSquad tradition where two editors face off, proposing different points of view on a matter, and include a fair amount of trash talking in order to make the article matter a more entertaining read. Naturally, it appears as though Brandon is a staunch ATI fan and I am wholly pro-NVIDIA, but that's only for article purposes. I can assure you that neither I nor Brandon feel any particular loyalty nor bias towards either company, we just happened to choose those sides since I was more familiar with NVIDIA's current hardware than ATI's.

The problem, as I see it, is that it is difficult to tell where Mr. Tismer's text in the letter ends and mine begins, and that the post on your site only includes a plaintext URL, rather than a link, which makes it less likely that readers would discover the true nature of the article. The way the letter is presented makes it look like FiringSquad and Gainward are in seeming agreement about the superiority of NVIDIA hardware, when this is certainly not the case. Not only is it too early to tell, given the pre-release nature of the drivers and hardware, we're not in the habit of endorsing manufacturers in that fashion.

I would be grateful if you could clarify this situation in an updated post on your site.

Thanks in advance for your time,

Jakub