Thread: Anandtech News

Since starting their GPU Technology Conference in 2010, NVIDIA has expanded into several events so that they can hold events in Europe and Asia. The next flagship GTC will be in San Jose in May, but NVIDIA’s #2 conference, GTC Asia, is occurring this week in Beijing. As with GTC America, GTC Asia serves several purposes for the company: a research symposium, a developer training program, and of course a platform for NVIDIA to announce new GPU computing products and initiatives.
The latter is the spotlight today, as NVIDIA is using GTC Asia to announce the latest versions of their development toolkits. Parallel Nsight, NVIDIA’s Visual Studio development toolkit, has just had its second release candidate for version 2.1 released. Meanwhile (and more importantly), CUDA 4.1 is also being released as a release candidate.
CUDA version numbering is based on features so don’t let the .1 numbering undersell CUDA, CUDA 4.1 is a major release for NVIDIA due to what NVIDIA’s been doing to the backend. Previously the CUDA compiler toolchain was developed entirely within NVIDIA as a proprietary product; developers could write tools that could generate PTX code (NVIDIA’s intermediate virtual ISA), but the compiling of PTX to binary code was handled by NVIDIA’s tools. CUDA 4.1 changes this in a big way: the CUDA compiler is now being built against LLVM, the modular compiler.
LLVM in the strictest sense isn’t a true compiler (it doesn’t generate binary code on its own), but as a modular compiler it’s earned quite a reputation for generating efficient intermediate code and for being easy to add new support for new languages and architectures to. If you can generate code that goes into LLVM, then you can get out code for any architecture LLVM supports and it will probably be pretty efficient too. LLVM has been around for quite some time – and is most famously used as the compiler for Mac OS X and iOS starting with Mac OS X 10.6 – but this is the first time it’s been used for a GPU in this fashion.
So why is CUDA's using LLVM a big deal? If nothing else it should result in shorter compile times and slightly faster performing code for CUDA developers. How much any individual developer will benefit, but from what we’ve been hearing the compile times are anywhere between 10% to 50% quicker. Meanwhile the nature of GPU computing means that application/kernel performance won’t improve nearly as much – LLVM can’t parallelize your code for you – but it should be able to generate slightly smarter code, particularly code from non-NVIDIA languages where developers haven’t been able to do invest as much time in optimizing their PTX code generation.

However the move to LLVM isn’t just about immediate performance benefits, it also marks the start of a longer transition by NVIDIA. As we mentioned previously LLVM is a modular compiler, and was designed that way so that developers could easily add support for new languages (what goes in to LLVM) and architectures (what goes out of LLVM) to the LLVM ecosystem. In NVIDIA’s case moving to LLVM not only allows them to open up GPU computing to additional developers by making it possible to support more languages, but it allows CUDA developers to build CUDA applications for more architectures than just NVIDIA’s GPUs. Currently it’s possible to compile CUDA down to x86 through The Portland Group’s proprietary x86 CUDA compiler, and the move to LLVM would allow NVIDIA to target not just x86, but ARM too. ARM in fact is more than likely the key to all of this – just as how developers want to be able to use CUDA on their x86 + NVGPU clusters, they will want to be able to use CUDA on their Denver (ARM) + NVGPU clusters, and LLVM is how NVIDIA is going to accomplish this.
Finally, with the move to LLVM NVIDIA is also opening up CUDA, if ever so slightly. On a technical level NVIDIA’s CUDA LLVM compiler is a closed fork of LLVM (allowed via LLVM’s BSD-type license), and due to the changes NVIDIA has made it’s not possible to blindly plug in languages and architectures to the compiler. To actually add languages and architectures to CUDA LLVM you need the source code to it, and that’s where CUDA is becoming “open.” NVIDIA will not be releasing CUDA LLVM in a truly open source manner, but they will be releasing the source in a manner akin to Microsoft’s “shared source” initiative – eligible researchers and developers will be able to apply to NVIDIA for access to the source code. This allows NVIDIA to share CUDA LLVM with the necessary parties to expand its functionality without sharing it with everyone and having the inner workings of the Fermi code generator exposed, or having someone (i.e. AMD) add support for a new architecture and hurt NVIDIA’s hardware business in the process.
Source: NVIDIA Press Center



More...

When we took a look at Bigfoot's Killer Wireless NIC we had an opportunity to play with Cisco's E4200, a top of the line wireless router with 3x3 spatial streams and 450 Mbps performance in both the 2.4 GHz and 5 GHz bands. Not content with having one of the best routers available they decided to push the envelope a little further. The Cisco E4200 v2 Maximum Performance Dual-Band N900 includes an upgraded 1.2 GHz processor, designed to improve USB throughput when a disk is attached. Though we didn't test it with our sample, this is apparently a deficit in the original E4200, whose BCM4718 was clocked at 480 MHz. There's also been some optimizations to the antenna and amplifier configuration that should improve range and reduce dead spots. The router is available now at your favorite e-tailer for an MSRP of $199.



More...

Roku just announced a software update that is rolling out to its Roku 2 and LT lines, bringing several changes that should bring a smile to many users faces. The only additional functionality coming to the devices is the introduction of MKV support. This is constrained to USB support for now; there remains no official local network streaming channel available, though unofficial channels have been available for some time. Gaming, WiFi and audio performance improvements have been included, and the Roku Channel Store has been updated to include options for purchasing accessories along with Roku devices. Lastly, an official Roku iPhone app is being introduced, with an Android version to follow. We'll explore these upgrades further shortly and report in as we learn more.

In this release, we focused on several improvements and enhancements:

1. Support for more games – Added improvements to the 3D game experience on Roku 2. The first games to take advantages of these improvements—Downhill Bowling 2, Castle Warriors and Frisbee Forever—will be available in the Roku Channel Store shortly and are designed to work with the Game Remote.
2. Mobile support – Added enhancements to support an official Roku iPhone app that will be in the iTunes App Store soon. Followed by Android for our non-iOS loving friends.
3. MKV playback support – If you have a Roku 2 XS, you can use the USB Media Player channel to play MKV movie files from USB drives. Also, MP4 video playback has improved.
4. Improved Wi-Fi performance – Added optimizations to enhance streaming quality and reliability over Wi-Fi in Roku 2 players.
5. Improved Bluetooth Game Remote battery performance – Prolonged the battery life of the Roku Game Remote when it is used in conjunction with Wi-Fi networking on the player.
6. Improved Game Remote responsiveness – Improved wakeup responsiveness of Roku Game Remove over Bluetooth for quicker actions during video playback.
7. Improved audio performance – Fixed the cause of intermittent audio dropouts with some stereo receivers.
8. Shop Roku Deals in the Roku Channel Store – Recently we added the ability to purchase Roku players within the Roku Channel Store; now you can bundle accessories with your purchase too, like an HDMI cable.

More...

PogoPlug has been making appearances in a number of products for a few years now. The premise is simple: take a low-power ARM computer, give it a NAS optimized Linux build, add storage and network access and you have what's been termed the "personal cloud." Access your files anywhere from your PC, phone or tablet; and never beholden to a data center. PogoPlug's particular hardware has been integrated into branded docks and also to OEM branded storage solutions. PogoPlug recently improved its service to include 5 GB of cloud storage available free to anyone interested.
Today, PogoPlug introduces their Series 4 device, an update that adds one key addition. Like the original Seagate DeskStar dock, the Series 4 acts as a dock for Seagate GoFlex mobile hard drives and includes an SD card reader, Gigabit ethernet, single USB 2.0 port and two USB 3.0 ports. The USB 3.0 addition is new, and while the added throughput might be wasted to slow broadband and cellular access, the performance should help with local streaming.
We can expect more personal cloud devices coming our way. While the convenience of storing all your data in a massive redundant data center is great, having to contend with network difficulties and service fees can be less than ideal, particularly the high costs of expanding storage. Personal cloud devices remove the fees and storage barriers, and leave users with the task of optimization and network upkeep. If you're up to the task and interested in expandable cloud storage, this may be one of the best options around.The Series 4 is available now at your favorite e-tailer for $99.



More...

Today Remedy announced plans to finally bring their psychological thriller Alan Wake to PC via Steam early next year. The developer will also be including the game’s two pieces of DLC: The Signal and The Writer.
Released in 2010 for the Xbox 360, Alan Wake is a Stephen King-inspired tale about a writer whose wife goes missing in the Pacific Northwest. The game was announced in 2005 as a PC/Xbox 360 release, but in 2009 Remedy confirmed that they’d limited development to the 360.
Though he couldn’t discuss specifics like release date, Remedy Executive Vice President Aki Järvilehto told Rock Paper Shotgun that the PC version will feature Steamworks and promised that “the experience of running it on a high-end PC, with a ninja set-up, [is] a big difference.”
Remedy’s also been prepping another Alan Wake project, the Xbox Live Arcade spin-off Alan Wake’s American Nightmare, which should release right around the PC version of the original. They’ve yet to confirm whether or not American Nightmare will make its way to the PC as well.



More...

After a long and arduous wait, Verizon has finally announced availability and pricing of its own Galaxy Nexus variant. Eager Galaxy Nexus shoppers will be able to get the smartphone either in Verizon Wireless stores or online tomorrow, December 15th, for $299 on two year contract. Online ordering will begin at 1:00 AM eastern time on the 15th as well.

Details about the Verizon LTE enabled variant remain the same as what we've known before, and comes with 32 GB of NAND as opposed to the 16 GB GSM variant we've seen abroad and are working on reviewing now. As suspected, we've also learned that the LTE enabled Galaxy Nexus on Verizon uses the same combination of basebands as the Droid Charge, a Samsung CMC221 for LTE and Via Telecom 7.1 for 1x/EVDO.
Source: Verizon



More...

Intel's haphazard ultra mobile strategy is finally getting some much needed consolidation. We got the first hints of this at IDF where Intel announced a restructuring that would combine its ultra mobile (read: Atom) and Core (read: everything else) design teams under one virtual roof. The goal there was to promote better sharing of technologies and improve execution efficiency. Today, Intel told Reuters that it is consolidating four divisions (netbook/tablets, ultra mobility, mobile communications & mobile wireless) under a a new mobile and communications business unit. The move makes tremendous sense as the line between netbooks and tablets blurs considerably over the next two years. Not to mention that the ideal SoC architecture for a smartphone, tablet and netbook is likely one and the same.
We're all expecting to see big news from Intel in the Android smartphone space next year. Whether it will be enough of an about face to bring Intel to a competitive point remains to be seen. On the tablet side, I'm expecting Intel to focus more on Windows 8 tablets rather than play too much in the Android space.



More...

Futuremark just sent out an announcement stating that they’ve updated 3DMark 11 to version 1.0.3. The results are still compatible and comparable with the earlier versions, and this is apparently a minor update in the grand scheme of things. Here’s the quick rundown of changes:

• Bullet physics library updated to version 2.79 to improve compatibility with future CPUs and GPUs.
• For most systems containing multiple GPUs (e.g. integrated graphics and an additional discrete graphics card), the user can now select the GPU to be tested using a new option on the Help tab.
• Improved error handling and user messages for the most common compatibility issues.
• SystemInfo component updated to version 4.5 for improved compatibility with current and future hardware.
• New "More" tab with information about other Futuremark products and services.

For the interested, you can grab the update here.



More...

GPU benchmarking on the smartphone side of things is continuing to heat up, and today RightWare, maker of both BrowserMark and the very popular Basemark ES 2.0 (which are regular fixations in our smartphone reviews), has announced its OpenGL ES 3.0 benchmark, named Halti. Halti is slated to launch in 2012 for testing OpenGL ES 3.0 on devices coming in the near future.
Halti consists of two primary benchmarking scenes, much the same way that Basemark ES 2.0 had two scenes (Taiji and Hoverjet), this time around they're named Rush and Car Race. The two tests will have triangle counts at around 100,000 triangles per frame, and be rendered front to back where possible.

Rush consists of an action scene on a city rooftop and aims to approximate PC-quality game rendering. Character animation with skinning and bones, dynamic shadows and reflections, per pixel lighting with 3 point lights, depth of field, light bloom, and SSAO are all featured in Rush.

The second test, per its name, includes a car racing around a track and effects common to racing games. General rendering features are similar to the main test, but doesn't include animated characters like the first one. Instead, features like volumetric light cones from car headlights, image displacement to simulate heat effects, motion blur, and dynamically added skid marks are graphical features in Car Race.
RightWare and Halti will be ready to run on devices as soon as OpenGL ES 3.0 swings around in 2012, and will no doubt be a fixation in our reviews when that happens.
Source: RightWare



More...