 |  |
 | |
|  |
|  |
|  |
|  |
|  |
|  |
|  |
|  |
|  |
|
| |
|
| |
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
  | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Microprocessor
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Top Stories for December 10, 1999 (details below) |
|
| EE Times | Intel process tweak pushes Pentium III past 800-MHz |
| Semiconductor Business News | Rambus aims to double data rates in 2000, pursue new applications |
| C/Net | Rambus aims for turnaround with reorg |
| The Register Files | |
| The Register | Oh no! It's Intel FDIV time again! |
| The Register | Dell last domino to fall to AMD Athlon? |
| The Register | Rambus Ink confirms big game deal |
| Today's Related Stories | |
| EE Times | Intel Process Tweak Improves Pentium III |
| Electronic Buyers' News | Rambus to expand beyond DRAM; names Mooring as president |
| Microprocessor
Headline News Collected By Robert R. Collins |
Week of December 5, 1999 |
Older News |
December 10, 1999 |
||
Intel process tweak pushes Pentium III past 800-MHzBy Peter Clarke December 9, 1999 |
Intel Corp. is manufacturing Pentium III microprocessors with better than 800-MHz clock frequencies by using a refinement of CMOS process technology described by researchers at this week's International Electron Devices Meeting.
The "notched-poly" process refinement described by Tahir Ghani, senior engineer with Intel's logic technology development group (Portland, Ore.), has already been included in Intel's 0.18-micron CMOS process and in circuits manufactured with transistor gate lengths of 0.1 micron (100 nanometers). At a session after his presentation, Ghani said Intel had "made Pentium III devices but not shipped [them] yet." He later qualified that statement, saying, "Some devices have shipped, but we haven't announced that yet." Regardless, "The clock frequency is better than 800 MHz," Ghani said. |
See Today's Related Stories |
Rambus aims to double data rates in 2000, pursue new applicationsDecember 9, 1999 |
Rambus Inc. here today revealed plans to double the speed of its high-bandwidth memory technology to data transfer rates of 1.6 GHz during next year. Rambus' new technology roadmap also targets development of memory module bandwidth of 6.4 gigabits per second beyond 2000.
"These advances in Rambus technology will further lengthen our lead in high-bandwidth chip connections over competitive approaches," said Geoff Tate, chief executive officer of Rambus while disclosing the year 2000 roadmap. "Rambus will be able to provide system companies the performance improvements required as advanced applications are developed," he promised. |
See Today's Related Stories |
Rambus aims for turnaround with reorgBy Joe Wilcox December 9, 1999 |
Rambus is trying to shake off a troubled past by retooling for the future.
The firm, which designs high-performance computer chip technology, today announced organizational changes that include a new president. The shifts come as it prepares to launch into new markets, namely communications and other chip-to-chip applications. Rambus also disclosed the 2000 road map for existing chip technologies. |
See Related Story |
The Register Files |
||
Oh no! It's Intel FDIV time again!By Peter Sherriff December 9, 1999 |
As if it weren't enough that you can't buy Coppermine Pentium IIIs for love nor money, the latest list of erratanotbugs on the Intel Web site (over 60 and counting) includes one guaranteed to send a shiver down the spines of everyone at Satan Clara:
"E56. It is possible that a negative sign bit may be incorrectly applied to the result of an x87 floating point operation if it is closely preceded by a Streaming SIMD Extensions (SSE) multiply operation... |
|
Dell last domino to fall to AMD Athlon?By Mike Magee December 9, 1999 |
If Dell ever starts using the AMD Athlon in its machines, then that would be a mighty blow for Intel.
There are rumours spreading like wildfire through the UK industry that Dell is just about to take such a step, which would really be an unprecedented move for the direct vendor, acting, as it does, like an Intel distributor. |
|
Rambus Ink confirms big game dealBy Mike Magee December 9, 1999 |
One or two of our long-time readers were rather surprised when we ran a story a couple of days back suggesting that Rambus Ink, maker of fast memory, was undervalued.
We said that the introduction of the Sony PlayStation II was sure to give the company a fillip, regardless of whether or not it was worth stuffing RIMMs into Intel i820 or i840 motherboards. |
|
Today's Related Stories |
||
Intel Process Tweak Improves Pentium IIIBy Peter Clarke December 9, 1999 |
Intel is manufacturing Pentium IIImicroprocessors with better than 800-MHz clock frequencies byusing a refinement of CMOS process technology described by researchers at this week's International Electron Devices
Meeting(IEDM).
The "notched-poly" process refinement described by Tahir Ghani, senior engineer with Intel's logic technology development group in Portland, Ore., has already been included in Intel's 0.18-micron CMOS process and in circuits manufactured with transistor gate lengths of 0.1 micron (100 nanometers). |
|
Rambus to expand beyond DRAM; names Mooring as presidentBy Andrew MacLellan December 9, 1999 |
Rambus Inc. today appointed long-time senior executive David Mooring as its new president, as the company strikes out in search of additional markets that demand high-speed chip-to-chip connections.
Mooring, who was formerly vice president and general manager of the Personal Computer Division at Rambus, Mountain View, Calif., will be tasked with overseeing technology development and marketing within the company's core markets. |
|
December 9, 1999 |
||
Intel: Itanium to exceed performance estimatesBy Ken Popovich December 7, 1999 |
The development of Intel Corp.'s new 64-bit Itanium processor has reached a major milestone with the
shipment of prototype systems to developers, according to a company official. In addition, early test
results show the new processor exceeding performance predictions.
"I can't underestimate the magnitude of the accomplishment of being able to deliver development systems for this within three months of first silicon [i.e., building an initial sample chip]," said Ron Curry, director of marketing for Intel's I-64 processing division. "I think that's a really strong indicator of how things are going." |
|
Intel Ships Itanium-Based Prototype SystemsBy Mark Hachman December 8, 1999 |
Intel Corp. has begun delivering an early version of its 64-bit Itanium processor to development partners, the company said Tuesday. Intel's Itanium systems feature clock speeds "comparable to today's production servers and workstations," the company said. Observers said they believe the chip will be further tweaked to increase overall performance, if not the clock speed, to a level suitable for testing in the second quarter of 2000.
Volume shipments are still scheduled for the second half of next year. |
|
The Register Files |
||
Further memory problems dog CaminogateBy Mike Magee December 8, 1999 |
Once the Intel VC820 boxed mobo becomes available, and that won't be until next week according to channel sources, people prepared to pay the exorbitant price of Rambus RIMMs also need to watch out for a number of factors.
Although Intel engineers solved the problem by shutting off one of the sockets with a terminator, there is an additional factor which you will need to take into account if you want to populate the beastie. |
|
Merced-Itanium PCs expected October 2000By Mike Magee December 8, 1999 |
Sources close to Intel's plans have said that the first workstation and server systems using its 64-bit Merced Itanium chip are not now expected to hit the streets until October next year.
It was hoped that some systems would be on sale by August of next year, but our source said that a more realistic assessment is now October. The reason for the delay -- if it is indeed a delay -- is because hardware vendors want to stringently test machines before they go out of the door. |
|
December 7, 1999 |
||
Intel shipping prototype Itanium computersBy Stephen Shankland December 6, 1999 |
Intel has started the next phase of making its upcoming Itanium processor a real product, shipping prototype computers with the chip to hardware and software makers for product design and debugging.
Intel has made thousands of the prototype chips and plans to ship hundreds of systems containing the prototypes by year's end to computer makers, said Ron Curry, director of marketing for the new chips. The Itanium, Intel's first 64-bit processor, will compete against high-powered processors such as Sun's UltraSparc III. |
|
Transmeta: Embedded Powerhouse Or Pentium Killer?
|
If you were looking to Comdex for the inside scoop on the highly secretive start-up Transmeta, you were sorely disappointed. In fact, you were probably misled to a large extent. Transmeta was founded in 1995 by Dave
Ditzel, the one-time chief architect of Sun Microsystems' Sparc processor family. Ditze didn't make an appearance, but Comdex did feature a keynote speech by Linus Torvalds, the creator of the Linux operating system. Torvalds also happens to be Transmeta's best-known employee, even though he signed up well after the start-up was formed. In terms of hardware expertise, many members of the Texas Instruments team that was disbanded after the company decided not to go ahead with a Pentium-class clone were among the initial Transmeta recruits. Nevertheless, Torvalds' high profile has everyone from programmers to pundits thinking Transmeta is a software-oriented company, rather than a fab-less chip vendor, and that Torvalds is the driving force behind the company's activities. |
|
New Rules Enable Faster Mobile CPUs
|
In previous columns (see MPR 5/11/98, p. 3), I have noted the growing performance gap between mobile and desktop PC processors. Intel has talked about closing this gap, but despite the latest 0.18-micron mobile processors, the gap remains at a full four speed grades, from the Mobile Pentium III-500 to the desktop Pentium III-733.
Recently, processor vendors have made two moves to help alleviate this problem. First AMD, followed by Intel, raised the thermal limit for low-cost mobile processors from 10 W to 16 W, creating additional frequency headroom. Second, Intel is preparing a variable clock-speed technology it calls SpeedStep, which boosts operating frequency when extra power is available. Both of these moves help the frequency problem but create new issues. |
|
Samsung to double Rambus DRAM output within monthsBy Robert Ristelhueber December 6, 1999 |
Samsung Electronics plans to double its output of direct Rambus DRAMs between now and February to meet strong demand from PC and server companies, the head of Samsung's semiconductor business said Monday (Dec. 6).
Samsung had stopped ramping production of Rambus memories in October after Intel delayed its 820 chip set, which was to support the use of Rambus memories. But the company resumed Rambus production last month. "Many customers came to us because not many companies can supply [Rambus memories]", said Yoon-Woo Lee, president and chief executive officer of Samsung's semiconductor unit. Although Samsung is building about one million 128/144-Mbit Rambus chips per month, there is still a big market shortage of the devices, he said. |
|
AMD adds faster versions of K6-2E embedded processorsDecember 6, 1999 |
Advanced Micro Devices Inc. continues to rev up the speed of its processors, today announcing four new versions of its AMD-K6-2E embedded processor family at speeds up to 350 MHz.
The new additions include the 350 MHz and 333 MHz standard power processors, and the 333 MHz and 300 MHz low-power processors. "These new processors give embedded systems developers a tremendous amount of performance for a very reasonable price, while taking advantage of x86 software compatibility, chip sets, board designs, and development tools to help enable a quick time to market," said Jerry Vogel, vice president and general manager of the Embedded Processor Division. |
|
The Register Files |
||
Free Merced simulator posted on WebBy Mike Magee December 6, 1999 |
The buzz on Intel's 64-bit Merced processor has lessened a little since the company's CEO Craig Barrett triumphantly announced silicon at the last Developer Forum in September.
But that doesn't mean that the whole world has suddenly gone quiet on Intel's flagship IA-64 chip. Instead, programmers and partners have been in a kind of Merced purdah, as they busy-bee away with their virtual screwdrivers and software wrenches. |
|
Seven Dramurai dwindle as PC-133 hopes kindleBy Mike Magee December 6, 1999 |
A quick look at the Rambus home page shows that the firm is claiming that its memory provides three times the performance of PC-100 synchronous DRAM.
That is highly debatable, given a PC platform like the i820, and given comparisons we've seen and written about on the world wide web. And even if it is true, a RIMM is currently over seven times the price of equivalent synchronous DRAM. Last week, a 64MB PC-100 piece cost around $70, a PC-133 piece about $88 and a PC-800 RIMM $490. A 128MB piece cost around $150, a PC-133 piece about $170, and a PC-800 RIMM $999. These prices may, almost definitely have, changed, but you can bet your bottom buck that the ratio hasn't. |
|
Gulag, SMP markets shiver at Coppermine StutterbugBy Mike Magee December 6, 1999 |
OK, maybe only one to two per cent of Coppermines might have the latest erratum which forced Dell to temporarily stop shipping some machines last week.
Last week, Intel said it would not recall the affected chips because of the erratum, which is No.50, instead saying it would fix the problem in the next stepping. The bug, which has already been dubbed Stutterbug, means that machines have to be switched on twice. |
|
AMD-Intel price war: hard figures emergeBy Mike Magee December 6, 1999 |
Intel will have to act fast to regain the initiative in the price battle it has waged with AMD during 1999.
According to Bill Henning at CPU Review, the latest set of figures he has collated on Athlon vs Pentium III prices show that AMD is winning the battle hands down. Henning states that AMD K7 Athlons are now cheaper than Pentium IIIs at all clock speeds, with a 700MHz Athlon coming nearly $140 cheaper than a Pentium III 700MHz chip. |
|
December 6, 1999 |
||
Intel continues to ship P3 CPU after bug is foundBy Will Wade December 3, 1999 |
Intel Corp. will continue to ship its most advanced microprocessor, despite the recent discovery of a minor defect.
The company has disclosed that up to 2 percent of its Coppermine chips — the Pentium III produced at 0.18-micron linewidths — have a bug that may interfere with the boot-up process. A spokesman said the glitch has only been seen in a lab environment, but that it could occasionally prevent the machine from switching on when the start button is pressed, requiring users to hit the switch a second time. |
|
Screen found for Pentium III bugBy Michael Kanellos December 2, 1999 |
Dell has begun to manufacture desktops containing Intel's "Coppermine" Pentium III processors again, as a test has been developed to detect a minor, recently detected flaw.
The flaw, more like a 24-hour flu, becomes apparent in the "boot-up" sequence and essentially forces computer owners to hit the "on" button twice, according to Michael Sullivan, an Intel spokesman. If a computer contains a chip with the flaw, it does not start until the second try. |
|
Coppermine bug stops Dell Optiplex PC shipmentsBy John G. Spooner December 1, 1999 |
A bug in the high-end Pentium III chip has forced Dell Computer Corp. to stop shipping its Optiplex GX110
corporate desktop.
An erratum, or glitch, in Intel Corp.'s desktop Pentium III "Coppermine" processors, which have been shipping since Oct. 25, can cause a PC not to boot every time its power button is pressed. The bug has not been found in mobile Pentium III chips, according to Intel. |
|
PC makers fixing 'Coppermine' glitchBy John G. Spooner December 2, 1999 |
The PC industry is getting back to normal after a glitch in Intel Corp.'s latest Pentium III processor put a short
hold on desktop PC shipments.
The glitch, which affects on/off functions of some PCs, caused Dell Computer Corp. and Compaq Computer Corp. to halt shipments of some Pentium III-based desktops. Dell, for example, halted shipments of its Optiplex GX 110 model. |
|
Gateway Refutes Claim It Uses AMD AthlonBy Mark Hachman December 3, 1999 |
A spokesman for PC OEM Gateway 2000 refuted reports that it plans to ship a system using Advanced Micro Devices' Athlon microprocessor.
Rumors of Gateway's commitment to the Athlon have swirled over the last months, and Gateway representatives have anonymously told other online news sites that an Athlon-based PC is in the offing. Compaq and IBM each ship Athlon-based PCs, and Gateway has previously built machines based on older AMD chips. |
|
The Register Files |
||
Non-Intel Processor Serial Numbers around for 20 yearsBy Pete Sherriff December 2, 1999 |
Chipzilla's much-maligned PSN is not the first system serial number.
Sun has had a hardware serial number coded into each system (as have most Unix hardware vendors) since the early 80s. Most large (aka expensive) Unix software is usually locked to a certain system serial number to prevent piracy. |
|
Coppermine bug a marchitectural disasterBy Mike Magee December 2, 1999 |
Now that the "respectable" news wires have picked up on our story yesterday (you know who you are) about the bug that means you have to switch your PC on twice, it's worth taking a longer look at the implications for Intel and for people able to buy Coppermine processors.
We note that several of these wires have run, verbatim, comments from Intel US PR spinners claiming that the problem only affects between one to two per cent of Coppermines going out the door. |
|
HP hopping mad at Merced-partner IntelBy Mike Magee December 2, 1999 |
Sources close to Hewlett Packard confirmed yesterday that it has expressed its displeasure to long-standing partner Intel over supplies of the ill-fated Coppermine processor.
But although HP is hopping mad that it, like other major vendors, was given very little advance information when Intel launched the Coppermine family on October 25th last, it is still striving to keep its temper, given the two companies' close, but increasingly strained, cooperation on the Merced Itanium processor. |
|
AMD likely to beat Intel, Alpha to 1GHz punchBy Mike Magee December 3, 1999 |
It now seems highly unlikely that Intel will be able to produce a 1GHz chip based on its IA-32 architecture before AMD, and possibly even before Compaq rolls out a 1GHz Alpha processor next summer.
According to information from a source close to AMD's plans, it can, as we have said before, produce an air-cooled Athlon K7 running at the magic 1GHz spot early next year, if it wants to. |
|
Intel starts to shuffle chip prices, specsBy Mike Magee December 5, 1999 |
Chip giant Intel will use next weekend to make some minor price reductions on its Pentium III family but will start to use its price fork extensively in January, as revealed here earlier.
While Intel will make some small reductions on Pentium IIIs across the range on the 12th of December next, the biggest cuts will come 23rd of January next. Our understanding is that Intel will also make a spate of announcements early in January, on the 10th. |
|
| Advertisement |