processors
Latest
NEC and Renesas looking to join forces against semiconductor evil
We're always up for a good semiconductor merger, and it looks like NEC Electronics and Renesas are prepping the biggest one we've seen this week. The two companies have agreed to team up, creating a combined force of $13 billion in yearly sales, and the largest chip company in Japan -- Renesas was already the product of a chip merger between Hitachi and Mitsubishi Electric. They'll still be behind Intel and Samsung in the overall game, but we won't hold that against them. Tokyo analysts believe this might lead to other "defensive" mergers by other Japanese chipmakers, but we'll have to wait and see. NEC and Renesas hope to finish talks by July and become a single company by April of next year.[Via Electronista]
Intel's upcoming mobile chips to squeeze 3GHz out of Penryn, bring high-performance ULV to the masses
Yeah, we've had just about all the Atom we can handle, and it looks like Intel's just about ready to help us back away from the difficult choice of sexy form factors for low prices and sexy form for exorbitant prices. Intel is working on Montevina Plus, which will push Penryn laptop chip technology past the 3GHz mark, while subsequently sending ULV chips into the mainstream, showing up in laptops ranging from $599 to $1,000, instead of the $1,500+ premiums they currently usually command -- great news for ultraportable lovers that actually want to get a few things accomplished on the road. Intel also sees 2009 as the year of the nettop, at least in emerging markets, and will naturally be pushing Nehalem all over the place -- with the way chip roadmaps are planned, the economic downturn naturally won't be messing with any planned rollouts for the time being.
AMD announces GLOBALFOUNDRIES spin-off, forgets to name it something awesome
AMD's finally dumped its fabrication facilities and technology onto a new spin-off brand, as expected. ATIC (Advanced Technology Investment Company) is on board as well, and the newly-formed GLOBALFOUNDRIES has $6 billion in investments to start out with, along with 2,800 employees. GLOBALFOUNDRIES will be primarily be building chips for AMD, just like usual, but will also be open to other gigs as a 3rd party chip builder -- its main rival TSMC just scored a deal with Intel to produce Atom chips on the cheap, a first for Intel who usually keeps production and processes in-house. First up for the new company? Churning out a 32nm process. We like the sound of that.
Intel's Silverthorn becomes the Atom, Menlow the Centrino Atom
Another brain-shattering announcement that you've all been waiting for is here. Intel has officially dubbed the Silverthorn and Diamondville chips "Atom," and the Menlow platform has become the "Centrino Atom." We know, it's hard to believe the day has come, but it's finally happened. Obviously, the 45nm Atom chips, and Centrino Atom technology will be targeted for MIDs, UMPCs, and all manner of small, internet-centric devices. Undoubtedly an unending vista of total awesomeness awaits us now that they've got some fancy new names.[Thanks to everyone who sent this in]
IBM creates a chip-sized supercomputer
Good news, everybody! Those super-geniuses over at IBM have whipped up a new form of CPU transfer which utilizes pulses of light instead of electricity to move data between cores on a chip. The new technology -- which is one-hundred times faster than current speeds -- is called silicon nanophotonics, and if implemented, could downsize supercomputers to laptop stature. The invention is unhindered by common problems with electrical chips, such as overheating and breakdown of data on short trips, allowing signals to pass unmolested over greater distances. Using this process, data can be moved a few centimeters, while requiring one-tenth as much power, resulting in lower operational costs for supercomputers. Will Green, a researcher at IBM, says that the company's creation will, "Be able to have hundreds or thousands of cores on a chip," and will result in huge speed boosts. Unfortunately, the project is on track to be carried out in 10 to 12 years, which leaves a lot of time to ponder if the chips will play Doom.
Intel roadmaps Penryn for mobiles
A recent leak from Intel has revealed the chip-maker's plans for its Penryn mobile processor line, detailing all the sweet news in a roadmap which has made its way into the waiting public's hands. On the sheet are the details of four new Core 2 Duo chips and one new Core 2 Extreme that Intel will be shepherding into the market in the near future (or so we hope). All of the chips feature the Santa Rosa-standard 800MHz frontside bus, with the C2Ds clocking in at 2.5GHZ and 2.6GHz with a 6MB cache, or alternately 2.1GHz and 2.4GHz with a 3MB cache -- the C2E delivers 2.8GHz with a 6MB cache. No word on pricing or availability, but you'll know as soon as we do.[Via Reg Hardware]
Intel Core 2 Extreme laptops in July?
According to what appears to be a leaked product roadmap, it's beginning to look like a hot summer if you're a laptop gamer, on-the-go video editor, or need to crunch some seriously large numbers while traveling. From the looks of the roadmap it appears Intel is planning updates and price cuts for laptop CPUs through September 2nd, including the introduction of a couple Core 2 Extreme chips (2.8GHz X7900 and 2.6GHz X7800, featuring 800MHz FSB/4MB L2 cache) as previously reported. All told, the processor overlords will apparently debut nine new chips and cut the price on four, with slight changes in pretty much every category. Of note is the addition of the budget Celeron M540 and M550, which according to "sources" are designed for the much-loved Santa Rosa chipset. Stay tuned, as we'll be bringing you more laptop-obsoleting news on this next round of chips as soon as we get it.[Via laptoping]
AMD Phenom FX, X4, X2 Stars to shine in Q3 / Q4
When it's been nearly six months since we've seen an elusive PowerPoint slide pointing to future AMD processors, it's been far too long, but it looks like the forecasting can begin again thanks to information about AMD's star-packed Q3. According to a roadmap slide charting out the future of AMD's "Stars" family of processors, a number of dual- and quad-core processor options will be headed to the desktop market sometime in the third and fourth quarters, and while the firm's "value" Athlon 64 X2 and Sempron chips will also see minor overhauls headed into Q1 of 2008, the focus seems to be on the Phenoms in the crew. The flagship quad-core Phenom FX (dubbed Agena FX) will clock in from 2.2GHz to 2.6GHz, sport 4 x 512KB of L2 cache, 2MB of L3 cache, and sport bus speeds of 3,200MHz or 3,600MHz. The X4 chips ratchet down in speed a bit, and only handle the AM2+ socket while the top-end FX plays nice with the 1207+. The X2 CPUs go the dual-core route, top out at 2.8GHz, but won't be ready for ordering until Q4 of this year. Be sure to hit the read link for a (larger) peek at the chart -- unless you just completed an order for a current-generation chip, that is.[Via RegHardware]
Xilinx, Altera showing off FPGA coprocessors at IDF
While Field Programmable Gate Arrays (FPGA) haven't benefited from a good deal of buzz just yet, things could be taking a turn, as both Xilinx and Altera caught a few eyes at IDF. We've already seen the unique, albeit highly specific chips in a supercomputer, but the dedicated coprocessor / accelerator modules could be landing beside your Intel Xeon CPU. Essentially, the devices plug "directly into the processor socket of dual- or quad-socket servers" in order to provide "high performance application acceleration ranging from 10x to 100x compared to processors alone, while simultaneously reducing overall system power consumption." The modules act as targeted CPUs, effectively computing very specific tasks in a much more efficient fashion than a general microprocessor can alone, which could boost the speed of scientific, financial, and life science applications that rely on very particular calculations. Of course, mainstream adoption still has quite a ways to go, but the quicker we get dedicated physics and AI coprocessors to go along with these snazzy new GPUs, the happier (and poorer) we'll be.[Via RobotSkirts]Read - Altera demonstrates FPGA at IDFRead - Xilinx demonstrates FPGA at IDF
Penryn CPUs benchmarked, don't slouch
So, 45nm sounds hip-cool and all, but what does that mean for performance? Intel's coming forward with some Penryn numbers -- with the help of Reg Hardware looking over its shoulder to make sure the tests are legit -- and it seems we're in for a decent speed bump, but not necessarily a revolution. Dual core and quad core 3.33GHz Penryns, each with a 1333MHz frontside bus, were pitted against the current desktop-leading 2.93GHz Core 2 Extreme QX6800 on a 1066MHz bus. Tests were run on top of a well decked test system, running Vista Ultimate. Naturally, the quad core Penryn took the top spot in all the tests, with varying margins depending upon which app was used -- no surprises there. The dual-core Penryn took third behind the spendy QX6800 for the most part, other than the single-threaded Half-Life 2 and the SSE 4-capable DivX test, which capitalized upon the media-friendly instructions Intel is adding to its new processors. Looks like we'll be seeing much more from these processors as more SSE 4 apps hit the market, but for now you shouldn't be feeling too bad about buying a QX6800 -- it's not obsolete by a long shot.
Intel poised to unveil new UMPC platform?
If slides released today by hardware site HKEPC turn out to be accurate, Intel is preparing to launch a new UMPC platform and CPU on the 18th that promises a much smaller footprint than the current architecture while almost doubling the devices' battery life. Known only by its codename "McCaslin," the chipset will reportedly occupy a space of 975 square millimeters (as opposed to the relatively enormous 2915 square millimeters of today's models) and feature power-sipping 600MHz and 800MHz processors known as "Stealey" that are targeting four to five hours of run time. Also sporting additional comm ports for devices such as GPS, 3G, and DTV, McCaslin is said to be Intel's response to Via's small form factor C7M, which has been adopted into models like Samsung's Q1 and continues to provide advantages over offerings from Santa Clara. Hit the Read link for a few more deets and another slide, and obviously stay tuned, as this rumor should be validated or debunked within a matter of days.[Via The Inquirer]
Intel, IBM announce insulation "overhaul" for faster, cooler chips
Intel and IBM are acting oh-so-smug right now about their respective achievements in insulating 45nm transistors, allowing the chips to run faster, leak less power and generate less heat. Intel, which hinted at such improvements in June, knocked out the first press announcement, promising to start implementing the technology in the second half of 2007, and IBM scrambled to catch up, announcing that it was "on the verge" of a similar discovery, and was working with AMD to start building comparable chips in Q1 2008. Intel is calling this the "biggest change in computer chips in 40 years," and while that sounds like hyperbole to us, there's no denying that 45nm is the new hotness, and bully for Intel for making it work. The actual technology -- at least Intel's method -- replaces the silicon dioxide insulators which leak too much energy when sliced too thin, but when made thicker reduce the electrical charge that passes through, reducing performance. Now Intel is using a hafnium-based insulator, which improves conductivity and reduces leakage simultaneously. Just for kicks, Intel is also tossing in a new metal alloy for its gate electrode, to replace the old silicon gate. IBM confirmed it is also using hafnium insulators, but couldn't help getting in a cheap-shot metaphor: "It's the difference between can openers and Ferraris," according to Bernard S. Meyerson, VP and chief technologit for systems and technology at IBM. Man, those technologists can really bring the smack.[Thanks, Lee G.]
Intel fights AMD's 4x4 with new quad-core tech
There aren't many arguments that Intel has a winner -- however temporary -- with their new Conroe Core 2 Duo chips, but they aren't relaxing just yet. With the oncoming threat of AMD's 4x4 chips in the performance/gamer space, Intel is getting a bit of quad-core action of their own with the new Kentsfield and Clovertown setups for consumer and server use respectively. Originally due in the first half of 2007, near when AMD's "Conroe killer" K8L was set to drop, both have been bumped up to fourth quarter '06 status, which sure has Intel looking smug. There are differences, however, in approach: Intel's Kentsfield architecture squeezes four cores into one socket, as compared to the dual sockets used by AMD's 4x4. This means the Intel chips will have less memory and frontside bus bandwidth available to each core, which will hurt them some in the performance race, but it also means cost savings that could give them an edge in the bang for buck arena. If anything, this is an indication of the intense competition going on for our chip dollars, and even if the quad-core war is a bit of a semantic one, we'll accept all the Photoshop crunching and AI processing power these guys want to push our way.
Intel selling mobile chip unit to Marvell
The rumors of Intel searching for a buyer were circulating, and it looks like they were true: Intel is selling off their mobile chip arm -- responsible for the quite successful XScale group of processors -- to Marvell Technology Group for $600 million. With chips making notable appearances in Treo, Blackberry and HTC devices, and pulling in a reported $250 million in revenue last year, we're still scratching our heads as to why exactly Intel would want to get rid of such a successful portion of their business. We'd think powering smartphones would be a priority for Intel right now, but the word is that Intel will have an option to receive $100 million of the purchase price in Marvell stock, so they might not be out of the game entirely. The chip unit currently employs 1,400 people, and Marvell plans to retain the "vast majority" of them, so this move shouldn't be too traumatic for most parties involved. We just hope they keep up with the R&D to get us smaller, faster, cooler and cheaper chips on the regular.
The amazing, shrinking Cell chip hits 65nm in '07
As part of any successful company's strategy to reduce the cost of their product (in this case the admittedly expensive PS3), Sony's Ken Kutaragi has revealed the not-surprising news that they're planning to reduce the system's Cell processor to a smaller, cheaper, and cooler 65nm manufacturing process in 2007. They're currently manufacturing the Cell using a more reliable 90nm process in both IBM's Fishkill and Sony's Nagasaki facilities. Microsoft announced their intentions to use a 65nm processor several months ago as well, as part of their long term plans to continuously reduce cost (not price) of the console. It's unknown, but likely, that the PS3 will be met with similar concerns as the Xbox 360's overheating woes; a 65nm die shrink would ameliorate such concerns. [Thanks, Siraris]
Intel reveals Core 2 Duo info and ULV Core Duo CPU
Intel is using this week's Computex trade show to reveal various bits of information about upcoming additions to the Core Duo family, including details on the chipset that will support the new Core 2 Duo (formerly known as Conroe) processors as well as the existence of an impending low voltage version of the current Core Duo CPU. Senior Vice President Anand Chandrasekher announced that the 965 chipset will come in three versions (P965, G965, and U965) sporting either Core 2 Duos or Core 2 Extremes -- the latter chip coming in July in a 2.93GHz version, followed by a 3.2GHz model later this year -- along with DDR2 memory as fast as 800MHz. As for the new ULV Core Duo processor, the company claims that it will sip an average of just 0.75 watts (compared to the one to two watts consumed by the already efficient Core 2 Duo), and be available in a 1.06GHz configuration called the U2500 for $289 this summer. Release dates for the Core 2 Duo chip are still unknown, although a press conference being held tomorrow may contain that key piece of info, which is already being reported by some sources as July 23rd.
AMD's "4 x 4" platform pairs 2 dual-core CPUs, 4 GPUs
Gamers gather 'round, for we have a tale of a new platform from AMD that will surely get your pulses racing, at least if the idea of having two dual-core processors and four graphics cards in one machine sounds as appealing to you as it does to us. Known as the "4 x 4" Enthusiasts Platform (four cores plus four GPUs -- we know, it's a bit of a stretch), the setup will allow manufacturers to offer configurations which users can upgrade over time, so consumers could start out with one dual-core chip and and one or more video cards, for example, instead of laying down what will likely be loads of cash on a maxed-out system. Another interesting bit of info revealed at the AMD analyst day in Austin, Texas was the fact that dual ATI Crossfire configurations can be used with 4 x 4, which indicates that AMD's possible acquisition target is probably working on a quad-SLI-esque solution to match rival nVidia. So start saving your pennies, dear gamers, because things are looking to get mighty interesting this year.
The switch from Intel/NVIDIA to IBM/ATI
The "Dean" (har har) of Xbox education, Dean Takahashi, wrote a typically thorough and thoughtful piece for this month's Electronic Business magazine. What? You're not a subscriber? Tsk tsk. Lucky for you, the good folks at EBM have put the article online.Takahashi profiles Nick Baker, one of the 360 system architects who decided to switch chip camps, going from Intel to rival IBM and switching from NVIDIA to (bitter) rival ATI. Takahashi covers the fascinating project of building a next-gen game console in -- what we assume is -- just a taste of his upcoming book on the 360.One of the most interesting parts in the piece is Takahashi's details surrounding the shortage of the GDDR3 RAM. He writes, "Both Samsung and Infineon Technologies had committed to making the GDDR3 memory for Microsoft. But some of Infineon's chips fell short of the 700 megahertz specified by Microsoft. Using such chips could have slowed games down noticeably. Microsoft's engineers consulted and decided to start sorting the chips, not using the subpar ones. Because GDDR3 700-MHz chips were just ramping up, there was no way to get more chips. Each system used eight chips. The shortage constrained the supply of Xbox 360s."Microsoft has argued that it was a generalized component shortage, but we've known better. [Thanks, Joe]
