|
|
|||||||||||||||
 |
|||||||||||||||
|
|
|||||||||||||||
|
|||||||||||||||
|
|
|||||||||||||||
|
|||||||||||||||
|
|||||||||||||||
|
|||||||||||||||
|
|||||||||||||||
|
|||||||||||||||
|
|
|||||||||||||||
|
|||||||||||||||
|
|
|||||||||||||||
| 2015 © KickAss Gear | |||||||||||||||
|
|
|||||||||||||||
|
|
| KickAss
Gear News Archive: July 2000 July 31st 2000 Oak Tech Wants to Mess Up Your Next CD Drive. Even though the RIAA's (Recording Industry Association of America) attempt to kill Napster has actually made Napster more popular than ever, it's not smooth sailing for MP3 sharing technology. An article at SBN says that Oak Technology Inc., the folks who make controller chips for optical storage devices, is gearing up to mess up your new CDRW drive. In just the same way that the RIAA has fostered other copy protection schemes, they are fully behind the newest one, called SDMI. SDMI is the "Secure Digital Music Initiative", who's purpose is to define a standard method of encrypting CD recordings so that only "SDMI-compliant" players will be able to access the data. In addition, the new technology will prevent sharing the data over an internet connection. It is not certain if Oak Tech's method will be chosen by the SDMI group, they have had trouble agreeing on standards in the past. But this hardware solution will be harder to defeat than software methods, and if the industry can standardize the hardware, then this chip-based approach has a good chance of working it's way into CDRW burners in the future. My questions: Will a black market develop around un-SDMI'ed or "cracked" CDRW drives? (Yep) Will old drives be able to play new, protected music? (Nope) If not, how angry will the public be when they need to replace their unprotected burner with a protected one just to listen to new CDs? Very angry. July 30th 2000 The Rambus Wars are Over! An article showed up at EBN, written by the venerated Jack Robertson, that declares the memory wars to be over. Rambus has been captured trying to escape through enemy lines, and is on it's way to the Playstation gulag (the crowd goes wild!). How did the Intel/Rambus love-in, which was in full bloom just a couple months ago, fall apart so quickly? Three letters..... A M D. That's right, all you AMD haters, and Intel lovers should keep one thing in mind, if it weren't for AMD, you'd be paying lots more for Intel products, PIIIs would still be at 750MHz, and you'd all be paying through the nose for Rambus memory. Thank you AMD! A
couple quotes from the article are in order: Ahhh. Finally. Spoken like a corporation rather than a political machine. But
here's where the sphincter phactor rises for Intel: Always good to have a fall-back plan, but Intel was a little late arriving at that decision. Oh well, better late than never, I always say. July 29th 2000 Napster Lives! A Washington Post report indicates that two federal appeals judges have granted the company a temporary stay of the lower court's ruling. As I mentioned on Thursday, the ruling by judge Patel indicated a profound lack of understanding of computers and audio technology. Well, apparently the federal appeals court judges agree with me when they stated that "substantial questions" had been raised about the "merits and form of the injunction." This is a pretty obvious indication the the federal judges may have clued into the fact that computers are indeed home recording devices, and therefore fall under the provisions of the Audio Home Recording Act of 1992, which protects noncommercial copying of sound recordings. NVidia Tries to Lie It's Way Back Into Fans Hearts. I don't usually call people liars... unless they are. I'll be very direct here, the interview I read over at Hypothermia with Derek Perez of NVIdia makes me ill. How the NVidia PR people can spew such obviously false information in an attempt to gloss-over their misbehavior is beyond me. It's almost embarrassing to read. How about this line? 2.) Although I was instructed by an individual who worked for NVIDIA to remove competitors logo's, you say this is not a "standard" policy of NVIDIA. What is your position on competitors logo placement?? Answer: Again, your site is your site - whatever you want to put on that site or however you wish to design that site is up to you. In PR we do not ask anyone to alter their design of the site for any reason. This is a bald faced lie, as any of the folks who have dealt with NVidia to get hardware will tell you. The next lie? 3.) Competitors product reviews would seem to naturally fall under the same category as the one listed above, but your official policy on this would be?? Answer: Products do not fall in this category - promotions and reviews are 2 different things. Whatever you review is whatever you review…we've never asked anyone to pull a review for competitive reasons. Oh sure. But I love this one the most: 4.) I know you have
stated that each site is treated on a case by case basis, so how would you
characterize my situation?? It might seem strange to people reading this,
and they wonder how things like this could transpire over such a long
period of time and no one be any the wiser including me, how does a
situation like this occur?? Isn't that a gas? Can't you just see NVidia swarming with non-employees, who do nefarious things while the PR team sits back and eats Chop Suey from cardboard containers?! They let un-monitored "interns" do all the correspondence with web sites? Of Course! The NVIdia PR people have no shame. They blame all their nasty behavior on "interns", and hope no one catches on. Jeez, they ought to look into careers in politics. This really won't have any effect on most gamers who just want to buy a good video card. It's all internal politics at the consumer-end of the hardware industry. But I believe that the way a company acts publicly says something about the people making decisions there, and if the decisions are bad enough, it affects my respect level for the company in general. I will always think of NVidia now as a company with a very dishonest PR team. Don't get me wrong, I LOVE NVidia video cards. I just don't like the PR team. In fact, I'm calling them liars. And unlike the other web sites that have to suck up to these bastards, we buy just about all the hardware we review, so we can say exactly what we want to say (can't you tell?). July 28th 2000 Another DRAM Maker Falls to the Rambus Steamroller. An article at Bloomberg.com says that Oki, a small DRAM producer, has agreed to pay Rambus Inc. royalties for the SDRAM and DDR DRAM products it produces. This brings the number of DRAM makers that are paying Rambus royalties to 3. Intel has been working with the remaining DRAM makers to develop a defense against the dubious Rambus patent claims. Considering the clout that Intel and the large memory producers have, I'm surprised that everyone isn't just in wait-and-see mode. Rambus' patent claims may have been acquired in bad faith, and if so, the royalties that Rambus is enjoying now may be short lived. Let's hope they have to pay it all back after they loose the fight. Sony Working on a 37Gigapixel Graphics Chip. A story over at SBN says that Sony is working on a new professional graphics engine capable of performing at 10 times the level of the Playstation-2. They have plans in the works for 2 upgrades to the design, the first of which will provide 100 times the performance of a PS-2, and the next with 1000 times the power of the console's engine. This translates to a fill rate of over 37Gigapixels per second! The new graphics engine is being called the GS-cube. This high-end graphics system will not be for home use, but you can bet there will be home-PC spin offs with awesome graphics power. Looks like we might have another potential competitor in the 3D PC graphics market! July 27th 2000 Late Edition: Looks Like the Pentium 4 Gets DDR. The speculation is apparently over. It seemed obvious that Intel must be thinking about a DDR-capable chipset for the Pentium 4, since their server-level processor, Itanium (Foster) was to use DDR memory, rather than Rambus. If Rambus was so damn good, why was Intel opting for DDR DRAM with the Itanium? Well, we need not go into the contract that Intel signed with Rambus. Nonetheless, if Jack Robertson at EBN is correct, Intel will be "investigating" the possibility of a DDR-Pentium 4 chipset. In corporate-speak, that's akin to saying that it's been in development for months. And as I mentioned in yesterday's news, if Intel is making a PC-133 SDRAM P4 chipset, they would be nuts not to make a DDR-capable one. Jack Robertson put in this way: Many observers questioned why Intel would even consider a single-data-rate PC133 chipset for its next high-performance processor. Nathan Brookwood, an analyst at InSight64, Saratoga, Calif., said a PC133 chipset has a data rate of 1-Gbyte/s, which leaves the 3.2-Gbyte/s dual channel Pentium 4 starved for data. "A PC133 chipset simply doesn't make much sense for Pentium 4." See folks, rumors on the Internet can often turn out to be true! Napster Ruling Proves Court Clueless. By now, anyone who uses Napster has probably heard the news. A Luddite Judge in California has ruled that computers are not recording devices (oh really?), and that Napster is therefore not a legal method of making copies of music. The judge in this case, Marilyn Hall Patel, seems to be completely ignorant of anything to do with computers. As if making the completely erroneous statement that computers are not recording devices was not dumb enough (I'm sure she has no clue that hard drives are magnetic recording devices), she then goes on to order Napster to make some new software that disables their old software! Now considering that Napster is all over the planet, on hundreds of thousands, if not millions of PCs out there, that would seem to mean that Napster is now supposed to write a computer virus that would track down and delete all versions of Napster in existence!? Again, Judge Patel has shown she is wholly ignorant about the subjects she is ruling on. Judge Patel even went so far as to say that Napster had "created a monster", and it was up to them to put the genie back in the bottle. The absurdity of her orders indicates a profound lack of understanding of the technology she is ruling on. Obviously computers are more than simple recording devices, but recording is one of a computers prime functions. Hard drives are magnetic recording devices, and they work on the same principles as tape recorders. Every time you install a program or download a file, you are making a digital recording on the hard drive. If the download happens to be music, then you are making a digital music recording, which can be played back later. This logic is unassailable. Obviously, computers are versatile recording devices, and therefore subject to the rules of home recording for personal use. I use mine that way, so am I breaking the law? If the RIAA and MPAA have their way, all music and all videos will be so tightly controlled, that we we will all need to sign license agreements just to listen or watch. Can you imagine entering in 16 digit CD or DVD codes to play your music or movies? This nonsense has to stop. It's time to make some rules here that stick, I'm tired of rehashing this topic every time a new device or piece of software comes out. It's stupid, and it's a waste of time and energy. It may be decades before all the Luddite judges are gone, and replaced with younger, more tech savvy judges, so I expect this nonsense will continue for some time. But the ability to get and share files over the Internet is here to stay. Wake up and smell the Java, Judge Patel. You can't undo Napster, anymore than you can undo the Internet. July 26th 2000 Rambus in Deep Doo Doo: PC-133 SDRAM for the P4. In a complete U-turn move, Intel announced yesterday that they would indeed be making a PC-133 SDRAM capable chipset for the upcoming Pentium 4 processor. This announcement has, in one stroke, struck down all previous Intel announcements on the topic in one blow. Rumors have circulated for months that Intel was working secretly on a double data rate (DDR) DRAM chipset for the Pentium 4. And just recently, Intel licensed the Pentium 4 chipset architecture (Tehama chipset) to competitor VIA in order that VIA could make a DDR DRAM version, which Intel is barred from doing by agreement with Rambus. While this development makes sense in certain ways, it also raises many questions. Does Intel believe this new, as yet un-named chipset, will not violate their agreement with Rambus to make Rambus DRAM the major memory architecture until the year 2003? Intel can argue that this is just an evolutionary step up from the i815e chipset they just released. Another question; if Intel thinks a PC-133 SDRAM version of the Tehama chipset is going to be a viable product, then don't they also know that a DDR-capable version would be greatly superior? This makes me wonder if the rumors about a fall-back DDR-Tehama chipset being in development at Intel might be true. According to Intel, the PC-133 chipset will not be a Tehama chipset with a memory translator hub (thank goodness!), but rather, will be a totally reworked PC-133 solution for the P4. Rambus released a statement that said they were confident that Rambus would be the superior architecture for the Pentium 4. Avo Kanadjian, Rambus' vice president of worldwide marketing said, "I'm confident when matched with Pentium 4 and dual-channel RDRAM, RDRAM will be superior." Superior to PC-133 SDRAM? Big deal! By late Fall, VIA will have a Pentium 4 chipset with DDR DRAM support. All of this has got to be depressing for Intel. Especially since they say that this PC-133 SDRAM chipset for the P4 will not be ready until the middle of 2001. That's long after VIA will have a DDR-P4 chipset to market, making the Intel offering seem too little, too late. Intel obviously realizes now that Rambus was a big mistake, but their contract with Rambus prevents them from developing a double data rate DRAM capable chipset for any of their processors until 2003. In the chip business, that might as well be eternity. So Intel is hedging their bets, and covering the bases. Good idea, but will it be enough to keep them in the chipset business? For more on the P4 and upcoming AMD Mustang, listen to the Daily Rumor. July 25th 2000 Intel to VIA.... "HELP!" In an interesting run of events, Intel is turning to VIA technologies, the little Taiwanese chipset maker they were suing just a month ago, to save Intel from the AMD processor onslaught. This puts VIA in a very VIAble position, as the predominant computer chipset maker for both AMD and Intel processor lines. This somewhat odd situation has been brewing for months, ever since Intel's own PIII chipset, the i820, failed to meet approval with the public or motherboard makers. The i820 disaster, as it is now know inside and outside of Intel's hallowed halls, was Intel's best attempt at meeting contractual agreements with Rambus Inc. to promote Rambus DRAM as the predominant memory standard at least until the year 2003. That leaves Intel in the sad situation where they can't make a Double Date Rate DRAM capable chipset for their new Pentium 4 processor, lest they break their dumb deal with Rambus (awe heck, just break it!). Intel has decided, apparently, that not having a backup plan for the "Tehama" Pentium 4 chipset is another possible disaster waiting to happen. But since they can't make their own P4 chipset with anything other than Rambus support, they have given VIA the go ahead to develop a third party chipset for the Pentium 4 that uses Double Data Rate DRAM. This indicates extreme worry on Intel's part, since they had originally wanted to be the sole chipset maker for the P4 line. Also, it suggests they are not confident that they, or anyone else can make Rambus into the dominant memory platform with yields so low, and prices so high. And finally, if the Tehama chipset is not popular, or has a problem with the Rambus interface again, then Intel will have no way of selling all the P4 processors they make. So they need VIA to fill the chipset void left by that 1997 agreement with Rambus, which does not expire for another 3 years. Remember the Y2K Bug? I predicted in December of 1999, that six to nine months after the so-called Y2K bug hit at midnight on January 1st 2000, no one would even remember it. Well, I haven't seen a single story on the failed Y2K bug in over six months. I think we can safely say that this was the most overblown non-problem in corporate history. The entire purpose of the heavy-handed scare tactics was to drum up business for computer makers and software authors. Indeed, the same thing still happens in miniature every time there is a big computer virus scare in the news. You start seeing Norton Antivirus ads everywhere, with the new virus' name in bold letters. Of course, the folks who made the big bucks on using the scare tactics would tell you that it was all their hard work that saved your butt. But that just ain't so. Sure, old corporate severs needed updating, but that's not new or special. Happens all the time. But it's in the interest of the Scare-Meisters to keep you off balance, and worried. So they need to keep up the corporate line that they were the ones who saved us all from impending disaster. Hogwash. I'm just glad it's another thousand years to the next one. July 24th 2000 When is Smaller Better? Several interesting developments occurred in the Land Of Microprocessors in the last few weeks. First is the announced delay of the Intel 64-bit processor code-named Itanium. Second was the announcement that Intel would move it's Pentium III line of processors to the smaller 0.13 micron fabrication process, in order to boost speeds. That's right, with microprocessors, smaller is better, because they are cheaper to produce, and run at higher clock speeds. According to an article at EBN, the new Pentium 4 may not be ready soon enough to compete with AMD's upcoming Athlon-Ultra, so Intel is reworking the "Coppermine" Pentium III to allow speeds up to 1.5GHz. This is contrary to the original plan, which was to keep the PIII at 1GHz and below, so that power-users would go for the more expensive Pentium 4 with Rambus DRAM. Another interesting bit of news was the report at The Register which says that AMD's 64bit processor project known as Sledgehammer has not been delayed, and in fact, engineering samples of the new chip have been sent to several large PC manufacturers. So it looks like the smaller company is quicker on it's feet than the lumbering chip giant. Who would have expected AMD to beat Intel to the Holy Grail of microprocessors, the first consumer level 64-bit processor? So there you have it again. When talking about chip manufacturers, it looks like smaller is better, at least at getting newer, faster product to market. July 21st 2000 NVidia Doesn't Play Nice. NVidia may have pushed a few web sites a little too hard in the last few months, and now they are reaping the rewards. Most web savvy users know that web sites often make deals with hardware and software companies to get products to review, in exchange for ad space (you will see no ads on any of our pages). What many folks don't realize is how radically different the approaches are that various promotional (PR) people use at the companies supplying products for review. The PR behavior these companies exhibit says oodles about the corporate climate and attitudes within those companies. July 20th 2000 T-Bird/Duron Overclocking Revisited. AnandTech has posted and article on how to modify an FIC socket-A motherboard to permit overclocking of T-Bird and Duron AMD processors. It requires lots of soldering, and the results were messy, but the new AMD CPUs are very overclockable, easily reaching 900MHz or more. The take home message is not to hone up on your soldering skills, but rather Anand was showing that dip switches or BIOS settings could be made to do the same thing. And that means that Abit, Asus and the others CAN make socket-A boards that overclock the new processors by increasing the so-called "multiplier factor". It had been reported all over the net, including here, that AMD would prevent multiplier changes via the motherboard. So if the AnandTech article is correct, then motherboard makers are probably working out those dip switches and BIOS settings right now. Tom's Hardware Rails Rambus Again! Another highly derogatory article about Rambus memory has appeared at Tom's Hardware Guide. Last time this happened, the stock price (RMBS) plummeted (hint!). Tom reiterated many of the things I have been saying about Rambus over the last few months, and also believes that Rambus' future is in question. Here's a fun sample. "Like a minion of the undead escaped from the grave, Rambus has managed to reanimate its lifeless corpse to roam the high tech landscape and terrify the populace of PC enthusiasts. Shouting curses promising eternal servitude to the Dark Lord of memory, Rambus has become more frightening than ever. Does Rambus really own the future of computer memory technology?" July 19th 2000 VIA Vows to Make a DDR Pentium-4 Chipset! An article at EBN yesterday indicates that VIA will be making a DDR-compatible chipset for the upcoming Willamette (Pentium-4) processor. This is huge news. Intel has it's hands tied by a 3 year old agreement with Rambus Inc., forcing them to steer clear of Double Data Rate (DDR) memory until the year 2003! This means that if Intel wants to avoid breaking their agreement with Rambus, then Intel can only make Rambus DRAM-capable chipsets for Pentium-4 motherboards. Their only loophole in this contract is if Rambus comes to within 5% of the cost of DDR memory (currently it is 3 times as expensive). And just by coincidence, Samsung announced that it has set a target for getting Rambus memory to within 5% of the cost of DDR memory by the year 2002. Will wonders never cease? So what can we expect this Fall? The chip wars in all their bloody glory. We will have the Pentium-4 processor versus the AMD Mustang processor. But now it looks like rather than the complete split between Intel and AMD on the memory front, we will have a choice between Pentium-4-based systems with DDR-DRAM support (VIA chipset) and those with Rambus DRAM support (Intel chipset). AMD will stick with DDR DRAM support for their upcoming Mustang processor. All this will begin sometime in September, so get your pop-corn and beer coolers ready! July 18th 2000 Intel to Drop Rambus?... Please!? According to articles at SBN and EBN, the relationship between Intel and Rambus is rocky, and getting worse. In a very interesting turn of events last week, Intel met with 5 major DRAM makers, excluding Rambus Inc. The topic under discussion was whether the DRAM makers were going to file an anti-trust case against Rambus for it's attempts to subsume all synchronous DRAM technologies under their Intellectual Property umbrella. Rambus did not comment on the situation, but it obviously has them nervous. Without Intel, Rambus will be relegated to toy systems like the Playstation2. All this talking behind Rambus' back is good news for fans of double data rate (DDR) memory technology. But don't get your hopes up any time soon for an Intel chipset that works with DDR memory. According to an article at SBN, a minor clause in the 1997 agreement between Rambus and Intel prohibits Intel from making DDR-capable chipsets until the year 2003! How the hell did Rambus bamboozle Intel soooo bad? Anyway, that is why Intel is licensing DDR stuff to companies like VIA, because their own agreement with Rambus bars them from doing it themselves. I think the between-the-lines message here is that Intel is not satisfied with what Rambus has delivered, and is starting to think that maybe they (Intel) got the short end of the stick in the 1997 deal (boy did they!!!!). According to Jack Robertson at SBN, there may even be several factions working within Intel, and at odds with each other. Based on their recent behavior, this makes sense. I think Intel will release the Pentium IV with Rambus-only support in order to stay within the bounds of their 1997 agreement. But this may be the last time. If there is any indication of reduced sales due to higher cost, slower performance, or reduced confidence in Rambus, then Intel will dump Rambus like wet cow-dung. And when that happens, let's hope it puts an end to the era where lawyers do all the work in technology companies, and maybe then we can get back to where corporate suit marketing types let the engineers do their jobs. July 13th 2000 Will T-Bird be Overclockable on the Motherboard? As many of you already know, Dr. Tom Pabst posted an article about how it would be possible to overclock AMD Thunderbird and Duron processors. This involves cutting physical connections on the chip's surface, and then reconnecting other ones to change the core voltage and multiplier factors. Right now, you will need to do some microsurgery to sever the required contacts, making the operation risky for the patient (your chip!). The question is, will it be possible to do the same thing using motherboard settings if motherboard makers add the right BIOS features? The answer is probably, no. Changing the multiplier and voltage settings will probably always require that the surface contacts be physically altered. AMD has obviously decided to make it possible to alter the multiplier and voltage settings, by leaving the contacts exposed at the surface of the chips. This is something Intel could have done, but did not do. However, as with the golden fingers contact, AMD is forcing overclockers to make obvious modifications to the processor that would void their warranty, so AMD is covered. As such, AMD overclocking will most likely continue to be done with "microsurgery", rather than BIOS settings. This gives Intel a slight advantage in the overclocking wars, where simple BIOS changes can boost a system by over 200MHz. July 11th 2000 Memory Makers Contemplate Rambus Complaint. According to an article at EBN several major memory manufacturers are in discussions deciding whether to take on Rambus by filing an antitrust complaint with the Federal Trade Commission (FTC). If the memory makers give their lawyers the go-ahead, they will be adopting much of the Hitachi counter-suit against Rambus, which was settled out of court two weeks ago. The claim is that the patents which were recently granted to Rambus encompassed technology that was already "prior art", which is legalese for existing, non-patentable technology. Indeed, the Hitachi claims went further and suggested that Rambus improperly garnered information about synchronous memory standards at JEDEC memory standards meetings. It is well known that Rambus violated the JEDEC standards body's rules by not disclosing patent applications it had outstanding while a member of the committee. It is also known that Rambus quit JEDEC just before filing amended synchronous memory controller patents in 1996, which were just granted by the US Patent Office in the last 6 months. If the memory makers decide to take on Rambus Inc., they will be attempting to get the FTC to cry foul, rendering the patents at risk. The FTC does not have any authority to invalidate patents, but it can restrict trade if it determines that a company is violating the Sherman Anti-trust Act. Rambus' claim that it holds patents that cover all current forms of synchronous memory suggest they are trying to restrict free trade, which is a violation of the Sherman Act. The big unknown factor in this story is JEDEC, the standards body who's rules were trounced by Rambus. They have not commented on the possible action against Rambus, but it would seem that they would have to play a role in determining if Rambus acted improperly. This is not a done deal yet, and even if they do file a complaint, what if anything will come of it is anybody's guess. July 9th 2000 It's Official; Intel Admits Rambus DRAM Stinks. An Article at SBN shows Intel's own benchmarks prove that an i815e motherboard with PC-133 SDRAM is faster than an i820 motherboard with Direct Rambus DRAM. Click here for graphs from Intel's web site. You can also go here to see Intel's full benchmark results. Intel kept the results apart so a direct comparison, with it's embarrassing implications, would not be easy. Rambus Inc. can not be happy about this, and by the time the news spreads, I would not be surprised if they asked Intel to remove the benchmark pages. Now let's take a look at the consumer side of this equation. The 128MBs of PC-133 SDRAM and i815e motherboard cost under $250 US, while the i820 motherboard and 128MBs of 800MHz Rambus DRAM cost almost $700. So unless there is something I'm missing, Intel and Rambus expect us to pay over $400 more for a slower, less reliable system, just so their long-term deal with Rambus Inc. flourishes. Rambus has been telling us for years that the increased bandwidth and reduced "pin-count" afforded by Rambus DRAM would be essential for computer performance in the near future. Well, that future is still not here yet, and it looks like Double Data Rate DRAM will beat Rambus memory even more convincingly, at a substantially lower price than the Rambus offering. So the question remains, why doesn't Intel cut it's losses and drop Rambus altogether? It has to do with business deals and contracts between the two companies (see next article), and Intel's desire to make the PC platform more proprietary. The Secret Intel-Rambus Deal. Jack Robertson at SBN wrote an interesting piece about the secret deals between Intel and Rambus. It is based on a revised Rambus filing to the Securities and Exchange Commission (SEC) last week. The oddest part is that the deal between Rambus and Intel was originally signed in 1996. This deal was revised 10 months later, but Rambus just submitted the revised SEC filing last week! Hmmm. The few details that emerged shed some light on Intel's behavior over the last 2 years. First off is the chance for Intel to pick up 4 million shares of Rambus Inc. stock (RMBS) at $2.50 a share. The current price is around $100, so it would be a major money maker if they can pull it off. But the deal has strings attached, including the fact that Intel has targets to meet in the percentage of motherboards they ship with Rambus-capable chipsets. If Intel can not meet the targets, the deal is in jeopardy. This can explain part of Intel's thinking in reducing BX chipset production last year, and it's lack of enthusiasm for adopting PC-133 SDRAM support this year. Obviously, the i820 chipset disaster did not help Intel meet the market-penetration goals set by Rambus. Now the question is, are the secret deals between Rambus and Intel at risk due to Intel's new i815e chipset with PC-133 SDRAM support. Considering that Intel's own benchmarks show the i815e with SDRAM outperforms the i820 with the fastest Rambus DRAM available, it would seem that i815e sales will seriously hurt i820 and Rambus sales. Most people would rather not pay lots more money for slower memory. With Intel's release of the i815e chipset, and the benchmark pages that show SDRAM beating Rambus DRAM, Rambus must be thinking "Et tu, Intel?" July 7th 2000 High Chip Demand Brings on Fab-Construction Bonanza. Articles at SBN and EBN today discuss the future implications of increased fabrication plant construction by chip makers. From memory chip manufacturers, to Intel and AMD, everyone is scrambling to increase fabrication capacity to meet the high demand for integrated circuit chips during this prolonged economic expansion. And while the article only deals with comments from a few conservative industry analysts, it still offers interesting insights into the cyclical nature of fabrication plant construction in boon and lean times. If the analysis is correct, then we can expect a rather large increase in memory chip and microprocessor production capacity in 2001 and 2002, at a time when the economy may be winding down. In the past, chip makers have often overproduced new fabrication lines during good economic times, only to leave them idle during the next recession. What this analysis fails to take into consideration, in my opinion, is that all of that old fab capacity is in use now, and more is needed. There is no reason to think that this will not remain true in the near future. Even if there is a temporary reduction in chip demand in 2002 (which I doubt), the new fabrication lines will still be needed as soon as the next economic growth cycle begins. Also, in my opinion, the need for integrated circuit chips of all types will grow as more of the world develops an information infrastructure. So the bottom line is, I doubt any company that is currently producing 0.18 micron or 0.13 micron fabrication facilities will be sorry they did so. Indeed, considering how fast computer technology is still improving, I do not predict a serious downturn in chip demand for the foreseeable future. Intel Processors Trickle In. For the first time in almost a month, our distributors are starting to show inventories of Intel processors. Most speeds are still absent, but a few models are actually in good supply. This is a welcome change, and hopefully signals the beginning of an end to Intel CPU shortages. July 6th 2000 VIA Settles Suits with Intel. It was reported at EBN today that VIA and Intel announced the settlement of their outstanding legal cases. This came as a bit of a surprise to me, as I assume it must have for others who have kept track of the case. On the surface, the relationship between Intel and VIA has not been good, but apparently, behind the scenes negotiations were quite fruitful. VIA has agreed to pay Intel an undisclosed lump sum, and royalties, for the privilege of producing Intel-compatible chipsets for the PIII and Celeron lines of processors. Intel, in return, dropped it's law suits against VIA for patent infringement, and it's complaint against VIA submitted to the International Trade Commission. It would seem that this is a good move for both companies. Over the last few months, while Intel and VIA were not on the best of terms, Intel found that it was unable to meet market demand for chipsets that support it's "Coppermine"-core processors. If it were not for VIA's alternative chipsets, Intel customers would have not had enough motherboards to pair up with Intel CPUs. It is possible that Intel figures that it's upcoming Pentium 4 and Tulloch chipset will be the next cash cows in their product line, and therefore it makes sense to license the PIII chipset to VIA. Intel keeps control, still makes money, and ensures a supply of PIII-capable chipsets for motherboard makers. It will be interesting to see if Intel goes the next step, and licenses the P4 chipset (Tulloch) to VIA in the near future, or if they try to keep that nugget all for themselves. July 3rd 2000 Computer Parts Shortages Continue. In the last two weeks, companies from Intel to 3dfx indicated that parts shortages were hurting production, and the bottom line. In the case of Intel, they have stated that shortages of components needed to make processors will keep Pentium and Celeron supplies tight through July at the least. We can confirm that supplies of all Intel processors have remained virtually absent from distributors warehouses, as Intel fights to keep Dell and other large computer makers supplied. In the case of 3dfx, low availability of certain circuit elements has reduced their production of the new Voodoo5 5500 cards, and may be leading to a delay of the upcoming V5 6000 cards. This has hurt 3dfx's profits, and could spell serious trouble for the video card maker. 3dfx got their new cards to market about 6 months later than they had originally hoped. Now, GeForce 2 cards are everywhere, and the new low-cost, GeForce2MX is on the way. Both are faster than the V5 5500 card, putting pressure on 3dfx to hurry the release of the V5 6000. However, if the parts shortages 3dfx is experiencing include memory chips, or VSA-100 chips, then producing V5 6000 cards would cut the number of cards they can manufacture in half. That's why I'm thinking that the V5 6000 cards may be delayed until the parts shortages are history. Other electronics shortages have reduced hard drive and DVD drive production, making the new higher speed models scarce indeed. We have been having severe problems finding good quality 7200rpm hard drives larger than 15GBs for about 6 weeks now. When these shortages might end is anyone's guess. Intel says that part of the problem was an underestimation of demand 2 to 3 years ago when fabrication plants were being built. Now we are seeing the results of these decisions in the form of insufficient component manufacturing capacity world wide. New NVidia Drivers. After spending some time working with the NVidia 5.30 detonator drivers, I can say they have fixed numerous problems we were having with GeForce cards, especially the flashing texture and menu problems. NVidia has also just leaked the latest 5.32 detonator drivers, but we have not had a chance to put them through tests yet. The 5.3 drivers seem to cut off a few frames per second in games, but they really seem solid compared to the 5.22 driver set. We will also check out the newer 5.32 drivers this week. You can get them both here. Copyright 2000, KickAss Gear |