Archive

Archive for the ‘Hardware’ Category

Seagate hits one terabit per square inch, compares self favorably to the Milky Way

March 20th, 2012

Seagate hits one terabit per square inch, compares self favorably to the Milky Way

You know that big new hard drive you just picked up? Get ready to feel bad. Seagate today is talking up the fact that it has managed to cram one terabit (that’s one trillion bits, for the record) into a square inch. That super-dense storage comes thanks to heat-assisted magnetic recording (HAMR) technology, a successor to the perpendicular magnetic recording (PMR) being utilized in current hard drives. The manufacturer sees the technology hitting the market later this decade, “doubl[ing] the storage capacity of today’s hard drives” in its wake. Just how many bits are we talking about here? Let Seagate put things into astronomical perspective: “The bits within a square inch of disk space, at the new milestone, far outnumber stars in the Milky Way, which astronomers put between 200 billion and 400 billion.”

Author:

Alleged Exynos 5 specs leaked in slide show spyshot

March 20th, 2012

Alleged Exynos 5 specs leaked in slide show spyshot

Did the murmurings about a quad-core Galaxy S III this morning leave you feeling a bit… meh? We don’t blame you. Sure, four cores and integrated LTE sounds great but (and this is a big “but”), Cortex-A9 is old hat. What we really want to hear about are those upcoming Cortex-A15 chips. Well, this should get your blood flowing — a slide has leaked, allegedly loaded with detailed Exynos 5 specs. Looks like initial entrants will have a pair of A15 cores running at 2GHz, backed up by a quad-core Mali T-604 MP4 GPU. All of that will supposedly be built with Samsung’s new 32nm High-K Metal Gate process. We’ve been unable to independently verify these details, but they are in line with what we’ve been led to expect. Still, until some Samsung PR comes down the pipeline we’re taking everything with a grain of salt.

SOURCE via Semi Accurate

Author:

Unnamed Samsung exec says quad-core Exynos inside Galaxy S III, LTE on-chip

March 20th, 2012

Unnamed Samsung exec says quad-core Exynos inside Galaxy S III, LTE on-chip

An unnamed Samsung exec hinted to the Korea Times that its upcoming flagship, the Galaxy S III, will include a next-gen quad-core Exynos chip that will incorporate LTE and WCDMA radios. Of course, such a revelation shouldn’t come as too much of a surprise since both its predecessors were built around home grown silicon… at least originally. According to the source, Sammy is trying to become more self-reliant and distance itself from Qualcomm which has provided single-chip solutions for a number of the manufacturer’s high-end handsets. The new AP appears to be the 32nm slab of silicon we heard about before MWC, which is sporting four A9 cores and not the more powerful A15. The executive said the development of the all-in-one chip is complete and it’s simply a matter of sticking them inside smartphones. Now, when exactly we can hope to see such a device hit the market is still a bit of a mystery.

SOURCE via Korea Times

Author:

AMD steals market share from Intel

March 19th, 2012

AMD steals market share from Intel

PC microprocessor sales climbed to $10.9 billion in the fourth quarter of last year, up 1.8 percent sequentially, and up 14.2 percent. IDC reports that for the entire year, PC microprocessor sales jumped by 13.2 percent to more than $41 billion. Much of the increase was due to higher average selling prices, which added about 9 percent.

Both Intel and AMD were able to gain market share in individual market segments, but it was AMD that came out on top overall. Intel ended the year with 80.1 percent share, down 0.6 points from 2010, while AMD gained 0.7 points to 19.7 percent and Via dropped 0.1 points to 0.2 percent share.

According to IDC, Intel suffered a 2.6 point loss in the mobile segment (down to 83.8 percent), as AMD added 2.7 points bringing it to 16.0 percent. AMD lost in the server and workstation segment where it is now at 5.5 percent, which was down 1.5 points. Intel is overwhelmingly dominant with 94.5 percent (up 1.5 points). In desktop processors, Intel came in at 73.8 percent (+1.7 points), while AMD was down 1.6 points to 26.0 percent.

SOURCE via IDC

Author:

Rumor: Radeon HD 7990 set to launch after Nvidia’s Kepler

March 19th, 2012

Rumor: Radeon HD 7990 set to launch after Nvidia's Kepler

Right before the New Year, we learned about the upcoming “New Zealand” graphics card, which is based on two “Tahiti” HD 7900 series graphics chipsets. With Nvidia set to release its first 28 nm based Kepler card later this month, Fudzilla reports that AMD plans to launch the Radeon HD 7990 just weeks after NVIDIA introduces its new GeForce GTX 680.

Though the final specifications are not known, it is rumored the HD 7990 will feature a core clock of 850 MHz, memory clock of 1250 MHz (5 GHz effective), 6 GB GDDR memory (3 GB per GPU), with 384-bit (x2) memory bus. It will be based on a PCIe 3.0 interface and support for DirectX 11.0 and OpenGL 4.0. In addition, the card will include 62 compute units for a total of 4096 (2048 x2) stream processors, 256 (128 x2) texture units and 64 (932 x2) full color ROPs.

The one thing we do know based on past experience with AMD’s dual-GPUs, is the Radeon HD 7990 will be 1) expensive (suggested price around $850 dollars) and 2) a lengthy graphics card (over 11.5 inches). For those Nvidia fans, I wouldn’t worry too much, as it is rumored that Nvidia already has its own dual GK104 card ready to counter attack.

Author:

Marvell unveils new SSD 6Gb/s SATA controller

March 19th, 2012

Marvell unveils new SSD 6Gb/s SATA controller

Marvell announced mass market availability of its new 88SS9187 SATA controller that will find its way into SSDs for ultrabooks. The controller comes with a SATA 6Gb (Rev 3.1) interface and supports NAND flash data transfer rates of up to 200 MB/s per channel (up to eight channels are supported). Marvell boasts a sequential read performance of around 500MB/s even at dirty drive conditions.

According to the manufacturer, the device also features “groundbreaking correction capability” based on an ECC engine with Adaptive Read and Write Scheme, as well as on-chip RAID support. 88SS9187 includes a DDR3 DRAM interface that can carry up to 1 GB of memory.

Marvell said it is targeting both enterprise and mobile devices with the controller and claims that the 88SS9187 features the lowest active and stand-by power consumption rates of any SATA 6Gb controller available at this time.

SOURCE via CNET

Author:

HP announces Sandy Bridge Xeon Workstations

March 19th, 2012

HP announces Sandy Bridge Xeon Workstations

HP is updating the Z Workstation series with Intel’s recently released Xeon E5-2600 Sandy Bridge processors. The computers will be available with 14 different E5-2600 processors in DP configurations ranging from the E5-2603 (1.8 GHz) to the E5-2687 (3.1 GHz). Pricing will start from $1,169 for the Z420 with support for up to 64 GB ECC DDR3 memory, $1,699 for the Z620 (up to 96 GB ECC DDR3), and $2,299 for the Z820, which integrates 16 DDR3 memory slots for up to 512 GB of ECC DD3 memory. All systems are scheduled to be available sometime in April.

HP offers plenty of configuration options that cover up to six 2.5-inch SAS HDDs (up to 3.6 TB), up to five 3.5-inch SATA or up to five 3.5-inch SAS HDDs (up to 14 TB or 3 TB, respecitively). Customers can also choose to order up to six 300 GB SATA SSDs. The configuration sheet includes professional graphics solutions from AMD (up to FirePro V7900) and Nvidia (up to Quadro 6000, Tesla C2075).

A fully decked out Z820 workstation will cost more than $30,000 and will be substantially more capable and expensive than the current Core i5 and Xeon 1200-series based workstations the company sells.

Author:

Fujitsu announces fast Cycle-for-Cycle ARM CPU simulator

March 17th, 2012

Fujitsu announces fast Cycle-for-Cycle ARM CPU simulator

The company claims it now has the world’s fastest ARM core simulation technology.

Fujitsu’s simulation technology is based on a JIT compiler that enables hardware designers to run cycle-level simulations of multicore ARM processors that can be executed on a “standard PC” Fujitsu says that the system runs at “more than 100 MHz” and is more than 100 times faster than previous simulators. The margin of error is +/- 5 percent in respect of the actual hardware.

Since ARM cores are widely used for smartphones, tablets and future entry-level PCs, Fujitsu believes that its simulator will result in shorter development cycles as well as a greater variety of ARM-based systems. The company said that it intends to continue its development and make its simulator both faster and more accurate.

SOURCE via Japan Corp

Author:

Nvidia ‘Kepler’ GeForce GTX 680 specifications leaked

March 16th, 2012

Nvidia 'Kepler' GeForce GTX 680 specifications leaked

In a leaked slide coming out of the Chinese web-site PCOnline.com.cn, we see an official Nvidia slide listing the GeForce GTX 680 specifications. The GeForce GTX 680 is based on the 28nm Kepler GK104 chip and is said to launch around March 22nd. The GTX 680 features clock speeds of 1006 MHz (base), and 1058 MHz (boost). The memory is clocked at 6 GHz (1500 MHz actual), and 2 GB GDDR5 memory running across a 256-bit wide interface. As we know, the GK104 will feature 1536 Stream Processors, 128 TMUs, and 32 ROPs. As with the AMD Radeon 7000 series cards, the GeForce GTX 680 will be based on a PCIe 3.0 bus interface with support for DX11.1, Quad Way SLI and display outputs of two DL-DVI, a HDMI and Display port 1.2. The card will utilize two 6-pin PCIe power connectors and has a TDP of 195W.

Nvidia 'Kepler' GeForce GTX 680 specifications leaked

As we saw at the Game Developers Conference, Epic showed the infamous Samaritan Demo running on a single Nvidia’s Kepler GPU, which could be the aforementioned GTX 680. Now, we see a slide coming out of NGF community that shows the performance of the GTX 680 versus AMD’s top two cards, the HD 7970 and HD 7950. The slide uses the HD 7970 as the baseline for performance comparison. Looking across the board, the GTX 680 outperforms the HD 7970 by roughly 20 percent in the seven games/benchmarks utilized. The one benchmark that stands out the most is the performance increase with Battlefield 3 (4xAA). The GTX 680 shows over a 40 percent performance increase, which could be related to use of Nvidia’s FXAA technology.

Nvidia 'Kepler' GeForce GTX 680 specifications leaked

Back in early March, we saw the first images of the GK104 card thanks to a leak coming from Chinese web-site Chiphell. Today, Chiphell has provided us with other leaked images and this time it is a near-final version of the graphics card. From these images, we see the cards Quad Way SLI support, along with its unique stacked power connector setup.

There is no official pricing for the card but early reports has it falling around $500 dollars. Stay tuned for more coming soon!

Author:

Intel says Thunderbolt optical cables coming this year

March 14th, 2012

Intel says Thunderbolt optical cables coming this year

Intel spokesman Dave Salvator has reportedly confirmed to IDG News that the company will release optical cables for Thunderbolt later this year. Unlike the copper-based versions, these should provide more bandwidth and longer cable runs in the near future.

Co-developed by Apple, Thunderbolt was originally designed as a faster alternative to USB 3.0 using fiber optics to transfer data at speeds of up to 10 Gbps. First introduced back in 2009 and then launched on Apple Macs in 2011, Intel wanted to reduce the number of ports on a PC and Mac by running all data transfers, networking and display protocols (including DisplayPort) through a single optical port. It would even support PCI-Express 2.0 for connecting external devices.

Current Thunderbolt devices share a common connector, meaning they can all be daisy-chained one after another by connecting copper-based and the upcoming optical cables. But until this year, fiber optics has been far too expensive for the general consumer, thus the two companies decided to settle on a cheaper copper solution for the immediate future. The copper solution thus is now able to provide up to 10 watts of power, but can only pump data across six meters at the most.

But according to Salvator, the upcoming optical cables will allow transfers across “tens of meters,” yet devices will need their own power supply at greater lengths, as running power over longer optical cable will cause a impedance-induced power drop and thus be impractical. However optical cables will allow for more bandwidth as the technology develops, so that’s a plus.

Current Thunderbolt installations in Apple Macs are based on copper, but they will still be compatible with the fiber optic cables launching later this year. For consumers, this means they will be able to purchase existing Thunderbolt products on the market and switch over to optical cables without having to make hardware changes to their current rig.

Current Thunderbolt circuitry ensures that the cables are transparent to copper or fiber optics connections, but the technology could get expensive once it moves to all-optical. Jim McGregor, chief technology strategist at In-Stat, believes this could stymie its adoption — right now it’s still a “niche technology.” Thunderbolt really won’t make a dent in the way consumers function until it’s integrated into handsets, camera, MP3 players and more, he said.

Still, given that the industry is trying to shift over to an all-wireless desktop, Thunderbolt may be locked down to specific needs like adding an external GPU or something similar. Thunderbolt hasn’t even arrived on the PC platform as of this writing, but it’s expected to make a debut sometime this year from the likes of Lenovo and a handful of other PC manufacturers.

Author: