TweakTown NewsRefine News by Category:
Toshiba have just made quite the leap, announcing they have developed a low-power, 64-core System-on-a-Chip (SoC) for embedded application in some areas which could be destined to automotive products and digital consumer products.
The insane chip sports 64 cores, claiming to feature eight times as many as its multi-core predecessor unveiled in 2008 and operates at an insane 14 times faster. The 64-core chip sits on a die area of 209.3mm2, two 32-core clusters are then integrated with dynamically reconfigurable processors, hardware accelerators, dual-channel DDR3 memory controller and other peripherals.
In one cluster, we'll find processor cores that share 2MB of L2 cache connected through a three-based network-on-chip (NoC). The high scalability and low power consumption is accomplished by the parallelized firmware for multimedia applications.
Take a quick stroll down memory lane and find Larrabee. Larrabee was Intel's failed discrete graphics creation designed to go head-to-head with the likes of NVIDIA and AMD. The Larrabee project was canned but Intel decided to continue the work with a change of focus. Instead of being a discrete GPU, Intel decided to target the HPC and supercomputing markets.
This new chip is an in-development chip known as Knight's Corner and is designed to be massively parallel. The product has clearly made significant progress as Intel has officially attached a brand name to the product. That brand name is Xeon Phi and will also be attached to future many-integrated-core (MIC) products.
Intel chose Xeon Phi as they wanted it to be part of the Xeon family. As for the Phi part, Intel says it's because it "evokes many concepts in science & nature including the 'golden ratio' in mathematics." Along with the brand name, Intel has released some technical details of the upcoming MIC chip.
The chip will be in production "in 2012" and be fabricated on Intel's 22nm process. The chip will feature more than 50 cores and "8GB+ GDDR5 memory" on a PCI Express card interface. Intel has asserted that the chip is capable of 1+ teraflops of double-precision compute throughput in Linpack. Intel has built a Knight's Corner cluster capable of producing 118 tflops as a proof-of-concept.
AMD's Mark Papermaster, senior vice president and chief technology officer of AMD, has said that they will be making a major change to its manufacturing process in 2013. AMD is planning on switching from its current SOI process to a 28nm Bulk CMOS process for the 2013 line of products. This, however, should not affect GPU manufacturing.
The current 7000 series "Southern Islands" GPUs currently makes use of TSMC's 28nm process. The upcoming "Sea Islands" GPUs will continue to make use of TSMC's 28nm process. The Sea Islands GPUs have already entered tape-out and should begin being manufactured in the end of 2012. The product should release first quarter of 2013.
Additionally, Papermaster pointed out that AMD is not opposed to working with other manufacturing houses as long as they benefit AMD in product launches. Papermaster furthermore commented that the HSA foundation was not directed at any specific competitor and ARM and AMD are only working together to combine ARM's TrustZone into AMD's APUs.
At the end of the Fusion Developer Summit, Mark Papermaster, AMD Executive VP and CTO, produced some roadmaps which provided us a small glimpse into the future plans of AMD. AMD is looking to bring three new APUs to market in 2013 and 3 new server chips later this year. With the come new updated core architectures.
Let's start with the APUs. AMD is planning 3 different versions for various markets. The top APU is codenamed 'Kaveri' and will feature 4 Steamroller cores. Steamroller is the successor to the Bulldozer/Piledriver cores. This APU will feature fully shared memory with the GPU and have 15-35W thermal envelopes and be used in 13.3-15.6" notebooks with thicknesses of 0.83" or less.
The middle APU is codenamed 'Kabini' and will be the successor to the Zacate and Ontario APUs. The Kabini APU will sport Jaguar cores which are the next iteration of the Bobcat core architecture. Thermals will be in the 9-25W range and be used in 11.6-15.6" notebooks with thicknesses in the neighborhood of 0.71-0.94". The bottom APU, 'Temash,' will be similar and manage to fit into a thermal envelope of 3.6W. Wow.
DigiTimes is reporting that Intel are to begin the phasing out of their Sandy Bridge-based processors this September. This paves the way for Ivy Bridge to get slapped into various products such as the Ultrabooks Intel are pushing big time.
Intel have said that OEMs have already begun receiving shipments of Ivy Bridge CPUs, with customers starting to notice them in systems later this month. Intel have also hinted that they expect to ship some 50-percent more Ivy Bridge CPUs in six months that it did in the same time period with Sandy Bridge.
But, there is a but here, folks. Ivy Bridge-based CPUs are much more expensive than their Sandy Bridge counterparts. According to DigiTimes, the price disparity is around $67-$100 per unit. This is a big problem for OEMs who are working on razor-thin margins on mainstream products that are the norm. If the Sandy Bridge chips do start disappearing in September, we should expect price cuts sometime this summer.
AMD has announced some more details on its upcoming Trinity APUs destined for desktops. AMD has stated that these chips will start to appear in all-in-one PCs by the end of the month. Curiously, this is contrary to what several motherboard manufacturers told TechReport while at Computex. They had said that desktop parts were delayed.
It looks as though AMD may be respinning the silicon for the retail desktop chips, so it probably won't be identical to the chips OEMs will be receiving this month. That aside, we do know the models which will be becoming available this month. The chips are almost identical to the mobile chips, except that they will be clocked higher.
All of the CPUs feature a quad-core CPU based upon the new Piledriver architecture and 4MB of L2 cache. They will support up to DDR3 at 1866MHz and slip into the FM2 socket. The K series denotes chips that are unlocked for overclockers, so if you plan to overclock, make sure to pay attention and get the correct model.
Computex 2012 - Overclocking is a huge thing here in Computex, where at the G.Skill booth they had some of the world's best overclocking champions facing it off on-stage. Young Pro, Hiwa, and Fred Yama were all there as we walked past.
If you remember, G.Skill are having their world-class overclocking invitational, and these great OC champs were going head-to-head. The presence and aura there was great, the crowd was loving it and the G.Skill booth babes were some of the best there. Yellow and black looks awesome, doesn't it?
Ahead of Computex next week, Intel have released details on 14 new dual-core Ivy Bridge-based processors. Core i5 and i7 processors will be among the first to drop, with Core i3 variants arriving later this year.
The Core i3 processors will also be joined with new Pentiums and Celerons, too. Out of the 14 new dual-core processors, six of them are desktop-grade, which means the remaining 8 of them are destined to mobile and ultra-low voltage. Ultra-low voltage chips sport a "U" in their name, which gives their identity away.
With Computex happening next week, we should expect to see a slew of Ultrabooks sporting the new ultra-low voltage Core processors. We should have a bunch of hands-on talk with various models, chips, Ultrabooks and so forth, so check back each and everyday next week for our coverage of Computex!
AMD's new APUs are pretty awesome. They combine great graphics with pretty good compute performance into a single chip. They are ideal for laptops, desktops, and embedded solutions. Today's award announcement, which brings them to their second consecutive win, is given by Taipei Computer Association after being judged by a panel of esteemed government officials, academics, research analysts, editors-in-chief of key media outlets and industry experts.
"We are honored to receive this prestigious award from the Taipei Computer Association for the second consecutive year," said Chris Cloran, corporate vice president, AMD Client Products. "The overwhelming market reception and industry accolades for the AMD A-Series APUs demonstrate that AMD is uniquely meeting the needs of our customers and a new generation of PC buyers."
According to the press release:
The second generation AMD A-Series APU combines the company's latest multi-core x86 CPU core architecture and award-winning AMD Radeon HD 7000 Series graphics to provide an increase in performance of up to 29 percent on productivity applications (as compared with the first generation AMD A-Series APUs), an increase of graphics performance up to 56 percent, and deliver up to 12 hours of battery life. Its unique features enable improvements in the areas that matter most to consumers today, including entertainment, gaming and mobility, through a growing ecosystem of APU-accelerated applications.
The AMD FX series of processors is known for overclocking pretty well. It currently holds the world record for fastest clock speed ever achieved and looks to be breaking that once again. A new CPU-z validation screen shot has shown up online. It details the feat even though HWBot rejected the the posting.
The rejection was most likely due to the fact that the overclocker didn't provide much in the way of detail about the cooling system used. When I say much, I mean any. What we do know is that the chip was clocked at a massive 8805.6MHz and was achieved by overclocker Ksin. Speed like that comes only when an overclocker uses extreme cooling such as liquid nitrogen or liquid helium.
To achieve the feat, Ksin used a base clock of 303.29 MHz and a multiplier of 29.0. To reach these clock speeds, an incredible 1.86 volts were pumped through the core. As such, it is recommended that you don't try this at home as that will kill your chip very quickly. Other components of the system include an ASUS Crosshair V Formula motherboard and 4 GB (2x 2 GB) of ADATA memory. The memory was clocked at 1417 MHz.
It appears that the posting has now been removed at the time of writing. I imagine this is due to the lack of cooling information.
UPDATE: The reason that the score was pulled from HWBot is because the Pi score indicated that the frequency could be bugged. This doesn't mean that it is, but that the possibility exists.