EVGA Brings Out 2 GB and 4 GB GTX 670 Cards

We've seen the new NVIDIA GeForce GTX 670, we've written about the ASUS GTX 670 DirectCU II and DirectCU II TOP and now we are doing the same for EVGA's own boards. 

EVGA has, or will soon have, five versions of the GTX 670 graphics cards up for sale. One might also say that there are only three boards available. The number five is reached because two of them come in both 2 GB and 4 GB GDDR5 VRAM memory capacities. The third comes only with 2 GB. 

The regular EVGA GTX 670 is one of the models with two memory options. Obviously, it sticks to the reference clock speeds: 915 MHz for the GPU (980 MHz GPU Boost) and the 6008 MHz GDDR5 frequency. 

The second card with both 2 GB and 4 GB versions is the EVGA GTX 670 SuperClocked, whose GK104 GPU functions at 967 MHz (1,046 MHz GPU Boost). The third and final card, the one with only 2 GB VRAM, is the GTX 670 FTW. Its GPU is clocked at 1006 MHz (1,084 MHz GPU Boost). 

All three adapters have dual-DVI connectivity, HDMI and DisplayPort outputs. What's more, since NVIDIA now supports up to four monitors on a single video adapter, 3D Vision Surround doesn't need a SLI setup to work. 

Not that SLI isn't supported of course. GTX 670 has three-way multi-GPU capability, although it will take a big motherboard with many PCI Express 3.0 or 2.0 slots for such systems to be made. All cards have dual-slot coolers after all. 

Be advised that only the reference-clocked EVGA GTX 670 sticks to the $399 / 399 Euro mark, unless it's the 4 GB iteration, in which case the tag is $469 / 469 Euro. The SuperClocked adapter sells for $419 / 419 Euro (2 GB) or $480 / 480 Euro (4 GB), while the FTX needs a sum of $439 / 439 Euro.

Read more »

Asus GeForce GTX 670 DirectCU II

NVIDIA's GeForce GTX 670 was meant to be a more affordable alternative to the GeForce GTX 680, but ASUS didn't see any problem with increasing the tag a little, not it if meant it could heavily overclock the GPU and toss a beast of a cooler on top of it.

For those who haven't read our article on the original GeForce GTX 670, the card is priced at $399 / 379-399 Euro. ASUS's GeForce GTX 670 Direct CU II TOP wants $419 in exchange for its services, and probably just as many Euro in Europe.

The name comes from the well-known DirectCU II cooler, with two fans and a large, long heatsink. This cooling module is as much a deliberate addition on ASUS's part as it is a necessity: the company used a custom PCB and factory-overclocked GPU, which means that there is extra heat to deal with, heat that the reference cooler might not so easily cope with.

Nickel-plated copper heatpipes come in direct contact with the GPU and take heat to the aluminum fins. The 90mm PWM fans disperse it afterwards. All in all, a 20% extra cooling prowess is achieved.

It doesn't hurt that DirectCU II is also much quieter than the alternative. And now we can finally say what those clocks are: 1,110 MHz during normal operation and 1,137 MHz when GPU Boost comes into play. For the sake of comparison, the normal speeds are 915 MHz and 980 MHz, respectively.

A final special perk is the ASUS Digi+ VRM, which lets owners fine-tune card voltage through the GPUTweak software. In addition to the GeForce GTX 670 DirectCu II TOP, ASUS is shipping the GTX 670 DirectCU II, which bears the custom cooler but no alterations to the performance. Consequently, the price is $399 / 399 Euro.

Like all GTX 670 models, ASUS' creations utilize 2 GB of GDDR5 VRAM clocked at 6 GHz, as well as dual-DVI, HDMI and DisplayPort connectors.

Read more »

2048pixels - World’s Best iPad Retina Wallpapers

A trio of talented and passionate people have been so impressed with the new iPad’s screen that they decided to launch a web site designed specifically to fulfill the need for higher resolution artwork required by the new tablet computer.

Launching with a bunch of wallpapers of their own, as well as allowing others to submit works, “2048pixels provides a venue for anyone to download original 2048x2048 iPad retina artwork by some of the best designers in the world.”

The site is truly breathtaking, featuring a clean design and super easy navigation between the works. If we didn't know any better, we'd be inclined to say it was crafted by Apple Inc. themselves.

“With the release of the latest iPad, there is a real need in the market for products that utilize this incredible resolution. 2048pixels creates a place for all designers to showcase their skills while leveraging this latest technology in exciting ways,” reads the description.

All wallpapers featured on 2048pixels must be original artwork, uploaded by their creators. Designs are assessed for quality, appropriateness, and originality, and finally offered up for grabs on the 2048pixels gallery.

“2048pixels is the best place to share your original 2048x2048 iPad retina wallpaper,” says the company. “We will post the best designs and images for the world to enjoy. Simply fill out all the fields in the form below and submit your design for consideration.”

Artists can’t sell their works via 2048px.com, and they must follow a few guidelines as well, such as a mandatory resolution of (you guessed it!) 2048 x 2048 pixels, JPG or PNG formats,  no adult themes, no watermarks, no faces of individuals, including prolific figures or celebrities, and only one submission per design.

The good news is that each artist that submits any works to 2048px retains full ownership of their artwork. The company specifically states “We do not claim any ownership to any of the artwork on 2048pixels. These terms do not grant us any rights to their artwork except for the limited right to share their artwork on 2048pixels.”

The same goes for you - the iPad owner downloading the goods. You can’t use them for commercial purposes, or anything of the sort.

Read more »

Gigabyte 15.6-Inch Full HD Gaming Laptop Debuts

There is a new gaming laptop up for order, or there will be as soon as Gigabyte finishes whatever loose ends have to be tied up in regards to shipments.

Called P2542G, the mobile personal computer is a gaming notebook with a Full HD screen (native resolution of 1,920 x 1,080 pixels) and a display diagonal of 15.6 inches. We've already run into people who feel that there is no point to Full HD on that screen size. There is no shortage of people holding the opposite view of course.

For our part, we feel that if tablets are using Full HD or better screens, laptops may as well do it by default. The 15.6-inch P2542G relies on a third-generation Intel Core i7-3610QM Quad-Core central processing unit (CPU).

For those who don't know, that chip has a base clock of 2.3 GHz and a Turbo Boost limit of 3.3 GHz. Up to 8 GB of DDR3 RAM back the chip up (two slots of 4 GB) and graphics prowess is brought to the equation by a 2 GB GDDR5 NVIDIA GeForce GTX 660M discrete GPU.

Needless to say, the Optimus technology, which automatically switches between the GPU and the one integrated in the CPU, is present and accounted for. That said, storage space is provided by a hard disk drive with a capacity of 750 GB. An mSATA SSD may or may not be present, depending on customer selection. If it is, it will play a part in the Intel Rapid Start Technology SSD caching.

What's more, a Blu-ray writer is included in the feature list, complete with the ability to scribe rewritable disks in up to BDXL format. Other hardware parts present are Gigabit Ethernet, 802.11b/g/n Wi-Fi, Bluetooth 4.0, a 4-in-1 card reader, two USB 2.0 ports, an eSATA/USB 2.0 combo port, a 1.3-megapixel webcam, audio/mic jacks and, of course, video outputs (HDMI, D-Sub). Finally, a Kensington lock keeps the notebook safe from thieves.

We're not sure about the battery life, but the price of the laptop should be around $1,100. European stores will probably go for 1,000 Euro or so.

Read more »

Gainward Launches the Cool Phantom GTX 670 Video Card

Well-known graphic card manufacturer Gainward has just announced the custom designed Nvidia GeForce GTX 670 Phantom along with the reference design, on its official website. This is a sleek-looking powerful video card with a pretty silent cooling system.

The card expectedly comes with factory pre-overclocked frequencies. The powerful cooling system will ensure reliable functioning of the GPU at 1006 MHz base frequency with a boost setting of 1084 MHz.

  

Memory frequency is also increased from the default 6008 MHz to a mediocre 6108 Mhz, resulting in a less than 3 percent overclock. It’s better than nothing we may say.

  

Unlike the GTX 680 Phantom, the GTX 670 version comes with only two fans underneath the heatsink fins. 

The price for the Phantom version has not been made available yet, but the standard version will retail at about 399 USD. That should be around 307 EUR for the European gamers, but we expect it to be priced closer to the 350 EUR mark.

Read more »

ZOTAC’s GeForce GTX 670 AMP! Edition Is Official!

Popular video card manufacturer Zotac has just launched the GeForce GTX 670 AMP! Edition custom Nvidia GeForce GTX 670 Kepler-based design, on its official website. It comes with a very powerful VRM and custom cooling system.

Zotac’s custom video cards have always impressed the PC enthusiast gamers, but the company’s GeForce GTX 670 AMP! Edition is actually bringing a record under the belt. Having a memory frequency of 1,652 MHz with a GDDR5 6600 MHz effective frequency, the GeForce GTX 670 AMP! Edition achieves a staggering 211 GB/s memory bandwidth. That is the highest memory bandwidth ever achieved on a 256-bit BUS.

Most of the GTX 670 AMP! Edition design is inherited from the company’s GeForce GTX 680 AMP! Edition that we’ve already presented here. Therefore, the GTX 670 AMP! comes with the same a large cooling system that has two 90-millimeter fans controlled by PWM, that are blowing air over a large aluminum heatsink containing four copper heat pipes that help the GPU distribute the heat evenly among the fins.

The frequencies are, of course, higher than the reference design. We have 1,098 MHz (core base frequency), 1,176 MHz (core boost active), and the record-breaking 6,600 MHz for the memory.

Nvidia’s reference, on the other hand, stands at 915/994/6008 MHz. Of course, the memory chips are cooled by special installed heatsinks being by the air flow coming from the two 90 millimeter fans.

Thus Zotac offers the slightly stronger VRM and the pre-overclocked frequencies that are around 20 percent higher on the GPU and 10 percent higher on the memory, with the hope gamers will choose the company’s product and pay the 439 USD price. That’s around 338 EUR for the European buyers.

Read more »

NVIDIA Launches GeForce GTX 670 Graphics Card

After weeks of leaks and reports, NVIDIA has finally made the formal introduction of the GeForce GTX 670 graphics adapter, powered by the GK104 graphics processing unit. 

The GK104 GPU is the same chip used in the making of the GTX 680 and the dual-chip GTX 690, constructed on the 28nm Kepler architecture. In the new video board, though, the processor does not have all CUDA cores active, nor as high a clock speed as on the flagship cards. 

The fully unlocked GK104 boasts 1,536 CUDA cores, but GTX 670 has only 1,344. Of course, this is still a nice number, but we digress. Meanwhile, the frequency of the GPU is 915 MHz, although the GPU Boost technology can take it to 980 when the need arises. For the sake of comparison, the GTX 680 features 1,006 MHz / 1,058 MHz speeds. 

The product's specifications continue as follows: 2 GB of GDDR5 VRAM clocked at 6.0 GHz, a memory interface of 256 bits, a memory bandwidth of 192.2 GB/s, a texture fill rate of 102.5 billion/s and four video outputs (two dual-link DVI, one HDMI and one DisplayPort). 

Furthermore, the Santa Clara, California-based company implemented support for pretty much every technology it has: 3D Vision, 3D Vision Surround, CUDA, PhysX, 3-way SLI, etc. Add to that OpenGL 4.2, PCI Express 3.0, multi-monitor capability (up to four), internal audio input for HDMI and HDCP certification, and PC Gamers don't really have any reason not to at least cast a glance in the direction of the GTX 670 when browsing for a new graphics beast. 

All the above fit on a 9.5-inch PCB (241 mm) and the price of the 28nm Kepler-powered newcomer is $399, which will doubtlessly turn into 399 Euro on the old continent, or 379 if we're lucky. Stay tuned to see what video controllers NVIDIA's OEMs have prepared.

Read more »

Windows 8 Introduces New NTFS Health Model and chkdsk

Among the various changes that Windows 8 will bring along when compared to previous platform versions, Microsoft included a new NTFS health model, complemented by a redesigned tool for disk corruption detection and fixing, none other than the chkdsk utility.

Whenever disk corruption occurs, Windows must isolate and resolve them, so as to continue to offer access to data on the disk. When launched, the NTFS file system would bring a simple approach to this, through looking at the file system volume as being either healthy or not. 

Provided that errors had been found, the drive would be taken offline to fix them, and access to files would be denied for as long as it was necessary to bring the volume back to a healthy state. The problem was that the process could take a lot of time, depending on the number of files on disk. In Windows Vista and Windows 7, chkdsk’s speed was increased, but it was offset by the release of higher capacity drives, which had a larger number of files on them. 

In Windows 8, a new health model of NTFS was introduced, as well as a new manner in which corruptions are fixed, so that downtimes would be minimized. Not to mention that the new ReFS file system does not need an offline chkdsk to repair corruptions.

The new health model comes with online self-healing, which involves a larger number of errors being fixed online, thus reducing the need for chkdsk. There is also a new online verification tool included in the platform, designed to check whether the corruption indeed occurred on disk and that it is not a memory issue. 

“This new service runs in the background and does not affect the normal functioning of the system; it does nothing unless the file system driver triggers it to verify a corruption,” Senior Program Manager Kiran Bangalore, Windows core storage and file systems, explains in a blog post. 

Additionally, Windows 8 comes with online identification and logging, which will also run in the background. Through precise and rapid correction (Spot fix), drives are taken offline only for a few seconds, at user’s convenience. 

“On Windows Server 8 systems with cluster shared volumes, we’ve eliminated this downtime completely. With this new model, chkdsk offline run time is now directly proportional to the number of corruptions, rather than being proportional to the number of files as in the old model,” Bangalore notes. 

File system health state in Windows 8

Windows 8 was designed to display the state of the file system via various interfaces:

Action Center – The health of the drive is most visible in the Action Center as the “Drive Status” (see figure below), which tells you when you need to take an action to bring the volume to a healthy state.

Explorer: The health state is also exposed in Explorer, under Drive properties.

PowerShell: You can also invoke the chkdsk functionality using a new cmdlet in PowerShell, REPAIR-VOLUME, which can be helpful for remote management of file system health.

Server Manager: In Windows Server, you can also manage the volume health states directly from the server manager utility.

In the new file system health model, there are four states in which a file system could find itself: Online and healthy, Online spot verification needed, Online scan needed, and Spot fix needed. The first three states do not require action from user. However, if a fix is required, the user has to restart the Windows 8 machine to resolve the problem. Due to the optimization of the process, however, only a few seconds will be added to the normal restart time of the machine. 

“For more advanced users who want to avoid restarting their system to fix a non-system volume corruption, they can open the Properties dialog for the affected volume, and on the Tools tab, they’ll see an option to check the drive for file system errors,” Bangalore explains. 

Basically, Windows 8 allows users to fix corruptions in drives that are not in use without fully restarting the system.

Read more »

Calxeda Presents First ARM Ubuntu Cloud Server

Calxeda, a company backed financially by ARM and which strives to make data centers as efficient as possible using ultra-low power technology, has revealed its Ubuntu powered EnergyCore reference server, in the presence of Mark Shuttleworth, at Ubuntu Developer Summit. 

Canonical, the company that developes the Ubuntu operating systems, has been working with Calxeda for a few years in order to make this server a reality. The solution presented at Ubuntu Developer Summy can easily deploy OpenStack based clouds using MAAS and Juju.

The server dubbed EnergyCore can house up to 48 Quadcore nodes at under 300 Watts with up to 24 SATA drives, which meas that it is possible to house 1000 server in just one rack.

Calxeda EnergyCore processors are a family of SoC (Server-on-Chip) products that bring the power efficiency of ARM processors to the datacenter. Built upon the Calxeda EnergyCore Architecture, each SoC integrates a unique set of capabilities all within a single chip.

Ubuntu Developer Summit for Ubuntu 12.10 (Quantal Quetzal), the upcoming version of the popular Ubuntu operating system, is taking place these days in Oakland, USA, between 7 and 11 May.

Read more »

HP Says Its Spectre XT Is Nothing Like the MacBook Air

After launching the new ENVY Spectre XT Ultrabook this week, HP design executive Stacy Wolff was pressed to respond to a question regarding design, and a potential design-centric lawsuit from Apple.

The similarities between Apple’s MacBook Air and the Spectre are not as striking as with other laptops out there. In fact, Wolff defended HP’s design approach with a pertinent argument, stating the following (via Engadget):

“I would go back to the TC1000 Tablet PC from about 10 years, and that's a tablet. I think if you look at the new Spectre XT, there are similarities in a way, not due to Apple but due to the way technologies developed. Apple may like to think that they own silver, but they don't. In no way did HP try to mimic Apple. In life there are a lot of similarities.”

That’s not all. Wolff agreed to elaborate on the matter in an exclusive interview following the launch event. Wolff said “The thing is that you have to design what's right, and that is that sometimes the wedge is the right solution, silver is the right solution.”

Wolff said there are as many differences as there are similarities between the two laptop models, opining that “anybody that's close enough to the business sees that there are differences in the design.”

The HP executive then dished out some of those particularities, explaining that the Spectre is rubber-coated at the bottom, and that it uses magnesium, rather than CNC aluminum (Apple’s preferred material for laptop enclosures).

“We did a brush pattern on our product,” Wolff continued, “they didn't.” “We did a different kind of keyboard execution. We did audio as a component; they didn't.” “So there are a lot of things I can list off that are differences; but if you want to look at a macro level, there are a lot of similarities to everything in the market that's an Ultrabook today. It is not because those guys did it first; it's just that's where the form factor is leading it,” Wolff concluded.

Indeed, HP’s design buff makes a valid point, at least as far as the Spectre XT is concerned. The company would have gone out of its way to make it different enough not to resemble the MacBook Air, or anything that falls in the “ultrabook” category.

Plus, Apple's Jony Ive should feel insulted by the claim that his streamlined MacBook Air (pictured above) looks anything like the Spectre XT. However, HP and other computer vendors have released systems that have been clearly inspired by Apple’s. It’s one thing to implement features so that “form follows function,” it’s another to try and mimic the success of another company by using key design features.

Read more »

ASUS Unveils Zenbook UX32VS with Nvidia Graphichs and FullHD Display

We all remember how Nvidia bragged about three hundred Ivy Bridge and Kepler design wins. Well, the first Ivy Bridge notebook on sale was using an AMD GPU and many of Nvidia’s Ivy Bridge design wins are, in fact, Fermi-based, not Kepler. Asus reportedly briefly unveiled some small details about its new Zenbook UX32VD.

The new MacBook Air competitor form Asustek comes with a very high quality FullHD screen using an IPS display panel. To power that high resolution display, Asus decided to include Nvidia’s Fermi-based 40 nm graphics processing unit (GPU), called GT 620.

This part is based on Nvidia’s Fermi GF116 GPU that comes with just 48 CUDA cores, 8 TMUs and 4 ROPs. The GPU has an 810 MHz frequency, but considering this is a mobile chip, the frequency might be lower than the desktop part.

Just like all Fermi GPUs, the GT620 has a share frequency of 1620 MHz, double the frequency of the rest of the GPU. The effective memory clock can be up to 3592 MHz, but again this is a laptop part, so the rated frequency might be lower.

We don’t yet know what the memory configuration will be, whether 512 or 1024 MB on a measly 64 Bit BUS. The new Asus Zenbook UX32VS will come with hybrid HDDs instead of the expected SSDs and this might be the reason that the UX32VS is thicker than the UX31A.

The marketing cardboard next to the pictured sample says “7mm HDD up to 500G” so we believe they’re talking about WD’s new slim HDDs. There is also a "24G SSD for caching" mentioned. That is likely the NAND part of a hybrid drive or simply a mSATA SSD. There is no information on the price available yet.

Read more »

Intel Xeon CPUs Arrival Date, Prices and Specs

Intel's next batch of processors from the Xeon collection has been in the making for some time, and it's high time the official release finally happened.

A recent report offers not just the announcement date of the inbound chips, but also the specifications of some of them.

By all accounts, Dell's release of the PowerEdge C5220 microserver was evidence that the Santa Clara, California-based chip giant would soon unveil Xeon E3 and E5 chip models.

Keep in mind that “announcement date” does not necessarily mean arrival date. Even if Intel does formally launch fresh E3 and E5 models on May 14, there is no guarantee they will immediately be up for sale.

May 14 is next Monday, which means that we should brace ourselves for a week chock full of new server releases, maybe even data center inaugurations. We wouldn't be all that shocked to hear about some new supercomputers either.

Anyway, the Xeon E5-2400 is Sandy Bridge-based and handles dual processing on 4 to 8 cores. The clock speed can be of up to 2.4 GHz and the Level 3 cache of as much as 20 MB. The LGA 1356 chips will have 60 to 90 W TDPs and prices of $192 (148 Euro) to $1,440 (1,112 Euro).

Quad-core Xeon E5-4600 will be launched too, with up to 8 cores, 20 MB L3 cache, 2.9 GHz frequencies and TDP of as much as 130 W. Compatible with socket LGA 211, they will ship for $551 (425 Euro) to $3,616 (2,793 Euro).

And now we get to the Xeon E3-2100 v2, built for one-way workstations and servers. They use the Ivy Bridge micro-architecture and the socket 1155-compliant package.

Both dual-core and quad-core processing units will be available, each with clock speeds of up to 3.7 GHz and TDP (thermal design power) of 17 to 87 Watts.

Compared to the first-generation E3-1200, the E3-1200 v2 have a newer micro-architecture, better graphics (on the E3-12x5), lower power draw and, of course, higher clocks on some models. Oddly enough, their price has not been provided.

Read more »

Intel Core i7's Vain Struggle Against AMD R-Series

Advanced Micro Devices will have both desktop and laptop accelerated processing units for consumers to buy, but that doesn't mean it won't have other chips too. 

In fact, the Sunnyvale, California-based company is also going to launch the R-Series, which is basically the mobile “Trinity” series only for the embedded market instead of notebooks. 

It is this R-Series that appeared in an Advanced Micro Devices PDF not too long ago. A graph shows the AMD R-464L and the R-272F pitted against the Core i7-2710QE, Core i5-2520M and Core i3-2310M. 

All the chips were tested under 3DMark 06 and 3DMark Vantage v1.1.0, and the results can easily be seen in the picture. The R-464L did 206% better and the R-272F managed to strike a 145% higher score than the Core i7-2710QE. 

Sure, the Intel processor is a Sandy Bridge equipped with Intel GMA HD 3000, not the 4000, but one has to take into account that, if what the PDF shows is true, then the R-464L might be better than the Core i7-3612QE Ivy Bridge too, which does have 4000 graphics. 

All in all, AMD's chips seem to be slowly but surely gaining an advantage, even if Intel does maintain its lead in terms of pure x86 performance. As visible in the graph, the accelerated processing units have TDPs (thermal design powers) of 35W. We suspect that only special chips used in Sleekbook-type laptops will run on less power. 

And here is where we would express our regrets at not having any other information on the chips. Fortunately, CPU World prevented us from experiencing such grief. They claim to have located some details, albeit only about the R-464L. Somehow. They are as follows: four x86 CPU cores clocked at 2.3 GHz, Radeon HD 7660G graphics and 4 MB of L2 cache (similarities to A10-4600M can be noticed). The Turbo Core CPU/GPU frequencies are still unknown.

Read more »