Last night NVIDIA's CEO, Jen-Hsun Huang, stated that shipments of its Tegra 4 SoC to customers would begin in Q2. A few outlets incorrectly assumed this meant Q2 of NVIDIA's fiscal year, but I just confirmed with NVIDIA that Jen-Hsun was referring to calendar Q2 - in other words, the period of time between April and June 2013.

Jen-Hsun also confirmed what was announced at CES: Shield, the handheld Android gaming device (with PC streaming capabilities) based on Tegra 4, would also ship in Q2. Jen-Hsun did add that Shield will show up in the latter part of Q2, which likely points to a late May/June launch.

In short, there's no new news here. NVIDIA mentioned Q2 as the release timeframe for Shield at its press event at CES last month. Obviously Shield can't launch without Tegra 4, so it's safe to say that Tegra 4 will also be shipping in Q2. With customer shipments happening in Q2, I'd expect products (other than Shield) in late Q2 or early Q3.

The rest of the earnings call was pretty interesting. GPU revenues are still solid despite maturing processor graphics solutions (although growing slowly), and Tegra revenue was up 50% over the previous year thanks to the success of Tegra 3. NVIDIA is still struggling on the smartphone side, but tablets have been a huge part of the success of the Tegra business unit.

Just after it launched, we discussed how the Nexus 4 included undocumented support for LTE on Band 4 (AWS) which could be enabled simply by choosing the appropriate RAT (Radio Access Technology) under Phone Info (by dialing *#*#4636#*#* - INFO). Back then, I noted that it was highly unlikely this would stick around for very long without the proper FCC paperwork, and although it took a bit longer than I expected, today's 4.2.2 update does away with this unofficial support for LTE entirely. 

  
Post 4.2.2 OTA (left), 4.2.1 (right) on Nexus 4

The OTA update for the Nexus 4 includes both software changes to Android (4.2.2 build JDQ39) along with a new baseband software image in the form of a delta update (radio.img.p). I tested the Nexus 4 the same way as I did in the previous article, on an Anritsu MD8475A base station emulator, which enables me to test any configuration or band network, and setup a Band 4 LTE network to attach the Nexus 4 to.

Before the OTA update, with the appropriate "LTE Only" selection made in the aforementioned "Phone Info" menu for preferred network type, the Nexus 4 would quickly attach to Band 4 LTE. After applying the update, the handset no longer attaches at all. In addition, trying to select the "LTE Only" preferred network type now quickly changes back to "WCDMA Preferred," likewise choosing one of the other modes which include LTE results in a change back to "WCDMA Preferred" after exiting and coming back. Previously this setting would persist until a reboot took place. 

I'm not surprised that undocumented LTE on Band 4 was removed, I am surprised however that it took this long. This also settles any lingering questions about LG Electronics filing for a Class II permissive change for the Nexus 4 to enable LTE on the bands supported by the hardware. If having support for LTE on Band 4 on your Nexus 4 is important to you, I'd recommend holding off on updating with the OTA zip for now, no doubt people will also make their own update images without the radio update as well.

The Android 4.2.2 update includes a number of other small changes as noted by other users, including enhanced quick toggles that can be long pressed to toggle WiFi or Bluetooth. 

Source: Google (OTA .zip link)

Earlier this morning Apple announced a combination of price cuts and spec updates to its MacBook Pro with Retina Display lineup. The price cuts impact the 13-inch rMBP, while the spec bumps extend across almost all models.

The good news is the price of the base and upgraded 13-inch rMBPs have dropped to $1499 and $1699, respectively. The 15-inch model remains untouched. The upgraded 13-inch rMBP configuration has a slightly faster Core i5 CPU (2.6GHz base clock instead of 2.5GHz, I believe this is a Core i5-3230M). The faster CPUs are nice to see, especially since that's really the only way to improve UI performance at this point until Apple brings some more software tweaks to OS X.

On the 15-inch side, both configurations get a 100MHz faster base clock (i7-3635QM and i7-3740QM most likely). The upgraded 15-inch model now comes with 16GB of DDR3L-1600 by default.

While default storage configurations don't change, SSD upgrade pricing does. The 512GB and 768GB SSD upgrades drop in price a bit depending on what configuration you're looking at. For the upgraded 15-inch model, moving to a 768GB SSD is now a $400 upgrade. That's not a lot for a 768GB drive, but it doesn't take into account the cost of the base 512GB SSD you are paying for but don't get to keep.

Overall these are welcome changes to pricing and specs. It was clear from the start that the MacBook Pro with Retina Display would eventually fall down to more reasonable prices, and this is likely the beginning of that curve. As high DPI displays become more commonplace, we'll see continued decline in the pricing department. These price cuts do come several months before the introduction of Haswell based rMBPs. Haswell's impact on the rMBP should be greatest on the 13-inch model, where the improved GPU performance will be able to make up for the fact that there's no discrete GPU (assuming Apple integrates Haswell GT3e silicon). You'll also see modest gains in idle power consumption, but the big platform battery life gains really come with Haswell ULx chips which we won't see until closer to the end of the year and will be used in tablets/convertibles.

After we published our Surface Pro review I went back and updated it to confirm that the SSD inside was indeed of the mSATA variety. My particular review sample used a Micron C400 mSATA SSD, similar to what I reviewed here a while back. Today, the patient folks at iFixit tore their Surface Pro apart and offered visual confirmation: the Surface Pro SSD is mSATA and technically, if you can get inside the system, it can be upgraded.

While I applaud Microsoft for sticking with an industry standard SSD interface and not something proprietary, iFixit discovered that getting into Surface Pro isn't all that easy. While Surface RT was built from three discrete pieces of VaporMg (kickstand, frame and back panel), Surface Pro only has two (kickstand and the frame with integrated back). The downside to this simplification is that the only way into Surface Pro is to soften the adhesive that holds the display stack in place and pry it out. This tends to be a potentially destructive way of getting into a device, so if you're planning on buying a 64GB Surface Pro and upgrading to a larger SSD yourself you might want to keep this in mind.

If you're still undeterred, it looks like you should be able to at least physically swap the Surface Pro SSD with something bigger. There's a whole other discussion about compatibility - I'm not sure what has been tested and validated with Surface Pro, but from a physical standpoint the SSD is upgradeable.

The most interesting things at CES were the things that weren't at CES. While I'll discuss the others in due time, one notable absence from CES was the often rumored Intel set-top box. Erik Huggers, the head of Intel's new Media division, confirmed today that Intel was working on a complete IPTV solution: service, software and hardware, all to be sold direct to consumers.

Erik confirmed that Intel Media was an all new group within Intel, staffed with folks from Apple, Netflix and Google among other places. Erik himself came from the BBC's Future Media & Technology group. Intel seems to know that it can't pull something like this off using internal talent alone. Bring in smart outsiders then give them the support of Intel seems to be the approach here - and it makes sense.

Details are scarce, and I'd expect them to be for quite some time. Intel committed to launch the box and service this year and it'll support features like catch-up, video on demand as well as live TV - all delivered over the internet. A list of content providers as well as pricing are two notable bits of information that were missing from today's disclosure. Until we get closer to launch, I wouldn't expect to hear anything on either front. Intel did mention that this wasn't a value play, and there was a strong focus on content bundles, which means this is likely not a full á la carte solution that's going to drive cable TV providers out of business.

I'll refrain from speculating publicly here, but there's a lot that makes this interesting. Some questions to ask:

1) Why is Intel getting into the TV business? Why now?

2) If not competing aggressively on pricing, then how will Intel differentiate?

3) Intel's divisions traditionally require high operating margins to survive in the long run. 

Unlike Intel's smartphone aspirations, its work in the TV space won't have the time to eventually build up steam. Intel also won't have the burden of playing catch up in a quickly maturing industry. The established players in the cable TV space have done a great job of stifling innovation for quite a while. It's clear that a more revolutionary approach is necessary to liberate live TV content. The question is whether or not Intel can be the first company to succeed where others have failed.

Many expected Apple to be the first to tackle and succeed here. Perhaps Intel will beat them to the punch? (this helps answer question #1 above).

More thoughts on this later. 

It was teased at this year's CES demonstrating VoLTE calls, we saw it working in a reference design, and today, Broadcom is finally ready to make their first LTE-enabled baseband official, dubbed BCM21892. This is again Broadcom's eagerly anticipated launch LTE baseband, and supports a pretty impressive set of features for their first entry. 

Broadcom's BCM21892 ensquared in red, from CES reference design

BCM21892 includes support for GSM/EDGE, WCDMA up to HSPA+ 42.2 (DC-HSPA+), TD-SCDMA for China, and UE Category 4 (150 Mbps on the downlink) LTE-Advanced with both TDD and FDD duplex modes. There's no CDMA2000 1x/EVDO which is pretty much expected, but more on that in a moment. BCM21892 is a 3GPP Release 10 part, which is again the LTE-Advanced release that includes things like LTE carrier aggregation inter and intra band. Broadcom tells me they will support carrier aggregation modes for LTE, and WCDMA. MIMO support on BCM21892 is 2x2. In addition, there's support for VoLTE calling (which we saw demonstrated) and 3G fallback modes for migrating that call over to WCDMA and GSM as necessary. In addition Broadcom claims they have enough performance to run an operator's IMS stack entirely on their baseband. 

BCM21892 is built on 28nm HPm process, which is the TSMC High-K Metal Gate for Mobile. Broadcom claims that their total solution area is 35 percent smaller than their main competition. I asked for a total package size, but they're not quite ready to disclose that. Another interesting tidbit is that Broadcom is including their transceiver inside the baseband package. DRAM is also inside the package, which is pretty common these days. There's still an external PMIC/PMU, which is also Broadcom's own design. 

Interestingly enough Broadcom claims it can do 8 bands (I'm assuming they're talking about primary receive) on its transceiver, which is one more than Qualcomm's 7 primary Rx ports on WTR1605/L. I asked about the low / mid / high / very-high configuration for Broadcom's transceiver, but they're not quite ready to disclose that information. 

There's no GNSS/GPS support on BCM21892 at the moment, Broadcom envisions customers using one of their discrete GNSS devices for this, since their WLAN/BT/FM/NFC combos also don't include GNSS support. It isn't too much of a stretch to picture GNSS moving onto the baseband with some future revision, however.

I saw a slide at CES which noted support for envelope tracking, which BCM21892 indeed includes support for officially now. Broadcom claims that support for envelope tracking among other things will give them a 25 percent power savings over competitors without it. Without going into too much discussion, envelope tracking involves varying the drive voltage for transmit power amplifiers depending on the amplitude (envelope) of the desired signal for maximum efficiency. We will see other players also start talking about envelope tracking soon as well. 

I asked Broadcom about what it views is BCM21892's market differentiation, and it comes down to three things. First is again package size, which is 35-37 percent smaller as noted, second is envelope tracking which I touched on, and third is perhaps the most interesting – improved coexistence and collaboration with Broadcom's combos (BCM4330/4334/43341/etc) on the unlicensed 2.4 GHz band and nearby adjacent licensed bands. This is particularly important or interesting for Dish Networks and AT&T for example, on the WCS band at 2.3 GHz. 

The last real big remaining question is what Broadcom will do in the USA where having CDMA2000 1x/EVDO is a relevant discussion point for operators like Verizon Wireless or Sprint. The answer is that for at least Verizon, the goal is to get LTE (and later, LTE-A) on Band 13 (and later 4) up to parity with existing CDMA2000 coverage profiles, and then use VoLTE for voice. For data, we're already seeing the result of some of the newer RFPs (Request For Proposal) which don't even include CDMA2000 as a requirement, for example Samsung Galaxy Camera on Verizon which is solely band 13 LTE. It isn't entirely surprising to see CDMA2000 support become something of a non-issue by end of 2013, early 2014 when Verizon finishes its LTE rollout. I noted it in my writeup from the CES VoLTE demo, but there's no surprise that Broadcom chose Verizon's Band 13 for its VoLTE demo given that situation. 

The rest of the discussion is operator and OEM support, which Broadcom claims is enthusiastic and already heading into field trials. The market at present for smartphone connectivity with LTE has really been dominated by Qualcomm, and Broadcom entering the fray with a competitive offering should make things interesting. Timing wise, BCM21892 is sampling right now, with production in early 2014. 

Source: Broadcom

We’ve reviewed quite a few Clevo notebooks over the years from a variety of vendors: AVADirect, Eurocom, Mythlogic, Origin, and Sager, to name a few. While there are certainly reasons to go with a branded Clevo notebook, it’s always good to have other options, and one of the most noteworthy is MSI’s G-series. CyberPowerPC has offered MSI-derived designs for a few years now, along with Clevo and Compal offerings. They’ve recently updated their MSI models to include support for the latest and greatest CPUs and GPU.

The new models now carry the FangBook X7 brand (also spelled FANGBOOK, FANGbook, and Fangbook depending on which email or web page you’re reading), with support for quad-core Ivy Bridge processors including the i7-3940XM and GPUs up to the GTX 680M. The combination makes for a potent gaming notebook, albeit with most of the usual caveats: large size, potentially loud fans, and less than ideal battery life (the X7 measures 16.85”x11.34”x2.17” and weighs “under 8lbs.”) The only other GPU option currently available is NVIDIA’s GTX 675MX, which is the Kepler-based replacement for the GTX 675M (a rebranded Fermi GTX 580M). We have yet to test the 675MX, so we’ll try to see about getting a review sample from CyberPowerPC.

MSI’s chassis also offers a few interesting perks that are worth mentioning. Dual 2.5” hard drive bays on 17.3” notebooks are nothing new, and even mSATA is quite common, but this is the first I’ve heard of a dual-mSATA notebook with dual 2.5” bays. That means you can configure an mSATA “Ultra RAID” RAID 0 drive for the OS and apps (up to 2x240GB in size)—interestingly, CyberPowerPC doesn’t even offer non-RAID mSATA options right now. Dual mSATA drives will provide potentially higher performance (up to 900MB/s read speeds), though the Intel 525 drives they’re using cost quite a bit more than some of the other options (e.g. Crucial m4 256GB mSATA can be had for $190 each compared to $722 for 2x240GB Intel 525). The FangBook also supports four SO-DIMM modules, so configurations up to 32GB (4x8GB) are still reasonably priced ($229 extra for DDR3-1333 or $317 for DDR3-1600). A matte 17.3” 1080p LCD comes standard, as does Windows 8 64-bit (though you can select Windows 7 if you prefer).

In terms of pricing, that’s one area where the FangBook X7 tends to make the most of the MSI platform. Even with the custom FangBook cover, the base model X7-100 includes an i7-3630QM, 8GB (2x4GB) DDR3-1600 memory, a GTX 675MX 4GB GPU, and a 750GB 7200RPM hard drive. All of that will set you back $1299, and while that’s quite a bit of money, a similar configuration using a Clevo P170EM chassis tends to start at $1500 (or more). $1499 gets you the base X7-200, which doubles the RAM to 4x4GB and adds a 64GB SanDisk SSD; the $1799 bumps the GPU up to the 680M and adds a Blu-ray combo drive. All of the models are configurable, of course, so you can mix and match components as you see fit.

The new FangBook X7 notebooks are available for order, with a current estimated ship date of 2/25/2013. If you’re not in a hurry, there’s a 5% discount available via the “NORUSH” coupon, which brings the starting price down to $1234. Also worthy of mention is that since the FangBooks all use NVIDIA GTX GPUs, the latest  $150 in F2P game currency promotion applies.

Last week AMD announced their Never Settle Reloaded gaming bundle, with several high-profile games available with the purchase of AMD GPUs. This week, NVIDIA follows suit with their own announcement...except this is a "gaming bundle" that's quite different from what we've seen in the past.

We've seen the transition from traditional gaming models to Free 2 Play (F2P) over the past few years, with many MMOs reporting increased revenue from the "free" model compared to monthly subscriptions. F2P has been so successful that quite a few high-end games have skipped the traditional model completely and launched as F2P. NVIDIA's latest bundle targets three of these titles: World of Tanks, Hawken, and Planetside 2.

Part of the reasoning behind the latest bundle is to convince gamers with lower end hardware to upgrade. Based on figures from Valve's latest Steam hardware surveys and NVIDIA's recommended settings for the above games, 36 million gamers don't meet the required hardware specifications for the above three titles. NVIDIA provided some numbers showing performance with their newer GPUs compared to an old 8800 GT as a reference:

In order to open up access to these and other games on NVIDIA hardware (note that Hawken and Planetside 2 both support PhysX while World of Tanks is a 3D Vision title), NVIDIA is offering up to $150 of in-game value with the purchase of a new GTX series GPU.

For GTX 650 and GTX 650 Ti purchases, buyers will receive a $25 credit for each of the games. Purchase a GTX 660 or above and the amount of in-game currency bumps up to $50 per title. While all of the games are technically free, the $25 or $50 credit is enough to get you jump started, and clearly the game manufacturers are hoping that after the initial taste gamers will be interested in forking over additional funds.

Putting things in a different light, the least expensive GTX 650 currently goes for $100 (with a $10 mail-in rebate available right now), so if you're actually interested in playing the above games that's potentially $25 towards the hardware and the rest towards the games. The GTX 650 Ti starts at $140 (with a $20 MIR available), and it offers twice as many CUDA cores with increased memory bandwidth for a fairly sizeable increase in performance. The base GTX 660 starts at $220 right now ($10 MIR), so that would be $70 towards the hardware and $150 towards the games. It increases the number of CUDA cores yet again and also comes with a 192-bit memory interface, effectively more than doubling the performance of the GTX 550 for a comensurate increase in price. (Note that it appears the above promotion also applies to new laptops with GTX 650M or higher GPUs.)

Keep in mind that both the AMD and NVIDIA bundles are delivering new games with hardware that is now several months old at best--in fact, AMD's bundle with the 7800 and 7900 uses hardware that's roughly a year old, and the GTX 680 is from the same era. This is one more way to try and entice users to upgrade, and there's the potential for new hardware to come out in the next few months that will make the current offerings look just a little less shiny. But that's always the case. If you've been sitting on the fence for a few months, this might be enough to push you over and get you to upgrade; at least, that's the hope. The full set of slides are included below for reference.

About ten days ago I posted about a problem with Samsung laptops getting bricked when trying to boot Linux using UEFI. There was a fair amount of commentary on the problem, and we later updated the article to indicate that it appears the problem is in the NVRAM and that removing the CMOS battery and power is enough to clear the problem. That's certainly still a major bug, and the best thing to do is to simply avoid using UEFI with Linux on Samsung laptops. Except, the bug appears to be quite a bit deeper than just a Linux driver.

According to Matt Garrett, the problem is that the "UEFI Black Box" is supposed to take certain actions when specific conditions are met, but the UEFI code or some other aspect of the firmware is misbehaving. Bugs in firmware can be some of the most damaging, and in this case the bugs appear capable of bricking a laptop, even when what you're doing is done according to written specifications.

Matt explains, "This is pretty obviously a firmware bug. Writing UEFI variables is expressly permitted by the specification, and there should never be a situation in which an OS can fill the variable store in such a way that the firmware refuses to boot the system. We've seen similar bugs in Intel's reference code in the past, but they were all fixed early last year. For now the safest thing to do is not to use UEFI on any Samsung laptops. Unfortunately, if you're using Windows, that'll require you to reinstall it from scratch." Additional details are available in Matt's post.

What's not clear is whether the data that's written that causes the laptops to fail to boot is battery powered (i.e. removing the CMOS battery will clear the error) or if the problem is being written to NAND in some cases. If the latter is happening, the only way to fix the problem would be to send the unit to Samsung for service (or buy a new motherboard). We'll update if there's any additional information.

We already covered the leak, and now we've got the confirmation we knew would come. HP has officially announced the HP Pavilion 14-c010us Chromebook. Quite the name, eh? Specs are unchanged from before, with 2GB of RAM, Intel's Celeron 847, a 16GB SSD, and that 14-inch 1366 x 768 display. The largest Chromebook yet comes with the same battery life as the Acer C7 (about 4.5 hours) and an extra pound of weight. Not much more to share till we get the hardware in our hands. 

Gaming on Android has been a hot topic lately. Purpose built consoles and handhelds have been rolling out along with plenty of excellent games. The WikiPad, announced in 10-inch form previously, takes a middle ground. A 7-inch Tegra 3-powered tablet is paired with a handheld dock that features two analog sticks, directional pad, a quartet of front buttons, two shoulder buttons and two trigger buttons, along with a pair of speakers and a grippy, contoured design. The software side brings Jelly Bean along with NVIDIA's Tegra Zone, MadFinger's Dead Trigger and Shadowgun: Dead Zone and Distinctive Games' Hockey Nations Tournament. The WikiPad is also PlayStation Mobile Certified, giving it access to the handful of Sony's library of games that have been ported to the platform. 

The tablet's specs bear a striking resemblance to Google's 7" tablet, and priced at $249 for the 16GB model it competes directly with the 32GB Nexus 7 or a 16GB Nexus 7 and a decent Bluetooth gamepad. No firm date for release was announced, and the 10-inch model is still in the works. That larger product had some manufacturing issues that presented a lengthy delay, and with the market moving smaller and cheaper the opportunity to bring out a smaller was too good to pass up. We'll see how thing work out come this Spring. 

The Atom D2700 series CPUs have proved quite popular in the SMB / SOHO NAS market. We saw vendors introduce models with HDMI outputs as early as December 2011. In fact, we covered QNAP's HDMI-equipped NAS models at CES 2012. Intel's GPU libraries for video playback and acceleration were not stable at that time and so, vendors were unable to show the true multimedia capabilities then.

Since CES 2012, we have seen a number of vendors trying to pitch a media center running on the NAS as a value add. I have personally tested XBMC on the Thecus N4800 and came away distinctly unimpressed (a though SmallNetBuilder also seems to echo). At CES 2013, I saw an implementation of the Boxee UI on the Asustor 6-series.

QNAP is the latest to join this trend with the TS-x69 Pro and L series, both of which are based on the Atom D2700 2.13 GHz CPU and come with HDMI ports. With the XBMC app, users to directly connect the NAS unit to a TV for media playback through HDMI. As most HTPC enthusiasts already know, XBMC also allows users to also manage their video library, music playlists and create photo slideshows. We will be evaluating one of the supported units in a couple of months, and hopefully, my experience will be better than what I had with the Thecus N4800.

Pricing and Availability

The TS-x69 Pro and L series come in 2, 4, 5, 6 and 8 bay models with prices ranging from $479 for the 2-bay models to $1,099 for the 8 bay models. The XBMC Media Center feature can be downloaded in the QNAP QPKG Center.

The push for smaller and thinner laptops, Ultrabooks, and tablets of late has come with some potentially undesirable side effects, namely the loss of flexibility. Of the Ultrabooks we’ve reviewed, I’m not sure any supported more than a single SO-DIMM slot for memory expansion, and many of them have all the DRAM components mounted directly onto the motherboard, all in the pursuit of reducing the z-height of the systems. In an effort to provide something of a middle ground, both Micron and TE Connectivity are offering alternatives that provide some reduction in z-height compared to standard SO-DIMMs while still maintaining the flexibility of an SO-DIMM slot.

The solution is quite simple and maintains full backwards compatibility with standard SO-DIMM slots, but to fully realize the z-height savings a modified SO-DIMM socket is required. In short, Micron is offering single-sided SO-DIMMs (with a standard 4GB capacity); since there are no components on one side of the SO-DIMM, it can lie flat against the motherboard. This is where the new SO-DIMM socket comes into play: it would have the module sit nearly flush against the motherboard so the connector would be the same but the housing would be slightly different.

To put things in perspective, a standard SO-DIMM is around 4mm thick; the new single-sided SO-DIMMs are able to reduce the z-height to 2.6mm. That’s not to say that they’re able to match surface mounted DRAM (around 1.2mm), but users and manufacturers would be able to choose between several memory configurations (generally speaking, 4GB or 8GB) and still maintain a thin profile. With surface mounted DRAM, you get the thinnest profile but completely lose out on upgradeability and if a company wants to offer two SKUs (e.g. 4GB and 8GB) it requires more effort in the manufacturing and assembly process. There’s also the potential for DRAM failures, which are simple to fix if you have a module but require a new board if you have surface mounted components.

From a high level, I’d just as soon see all modern laptops ship with 8GB standard, particularly the Ultrabooks with surface mounted DRAM, but manufacturers are always looking for ways to reduce cost and that has led to the existing crop of 4GB non-upgradeable Ultrabooks (ASUS UX21A/UX31A, Acer S7, etc.) One other item of note is that all of the reduced z-height modules from Micron will be reduced standby (1.35V DDR3L-RS). At least initially, the modules will only be shipping in 4GB capacities (currently, 8GB SO-DIMMs require dual-sided modules). Future higher density modules with monolithic devices (8x8Gb) should show up eventually, and of course all of the design elements are applicable to DDR4 when we see a shift to that some time likely next year.

This particular approach is only one of several that are apparently being tossed around in the industry, but thanks to the backwards compatibility with existing SO-DIMM slots it appears to have a better chance of succeeding. Other approaches that are being looked at right now include non-standard modules, which would require new connectors and modules and likely limited production compared to existing solutions. It’s expected other companies will also support the new connector, and availability of the new package (connector and single-sided SO-DIMMs) is expected this spring.

Arctic announced the availability of their MC001-XBMC passive HTPC today. Based on the Intel Atom D525 and the AMD Mobility Radeon HD 5430, the unit comes with OpenELEC pre-installed. The hardware platform supports HD audio bitstreaming and TV tuner / PVR capabilities.

The specifications of the unit are provided below:

One of the interesting aspects of the unit is the availability of a 1 TB HDD in the system for media storage and a DVB-T / ATSC TV tuner for PVR purposes. It is a bit unfortunate that OpenELEC doesn't officially seem to have ATSC tuner support for PVR purposes or HD audio bitstreaming enabled yet in the shipping version. Hopefully, a future OpenELEC update should resolve it (though one can use experimental builds / PVR version to get those features). That said, Arctic does have some interesting apps and features such as the Arctic Audio Relay which allows the unit to be used as a centralized jukebox when coupled with some accessories. The rest of the features such as a memory card reader, in-built Wi-Fi etc. are par for the course.

The MC001-XBMC is shipping now for 199 Euros / USD 229. Hit the source link for more information.

There has always been some level of lottery involved when buying an SSD-equipped Mac. The first Mac to start the lottery was the 2010 MacBook Air, which initially used Toshiba's SSDs but later it was found that some units were shipping Samsung SSD 470 based drives. Up until last year Apple kept using the same Toshiba and Samsung SSDs; the 2012 refresh finally brought SATA 6Gbps SSDs to Macs and, of course, another round of lottery too. The 2012 MacBook Air kept using drives from both Toshiba and Samsung, albeit both were newer generation (Toshiba's is SF-2281 based, Samsung is PM830). The rMBP, however, used only SSDs from Samsung, and the 256GB and 512GB SSDs in the MacBook Air were all Samsung made as well. 

Late last year it was discovered that some rMBPs might be using SSDs from SanDisk. At that time we only had system profiler screenshots and user reports to go with, which I didn't think was enough to go by (after all, the controller is the most interesting part of the story anyway). Today, I have some pictures to show:

Courtesy of MacRumors forum user Artagra

In short, that's a custom version of SanDisk's X100 mSATA SSD. The controller is Marvell's 88SS9174, which is also used in Crucial m4, Intel SSD 510, and Plextor M5S, each with custom firmware. That's coupled by eight (four on each side of the PCB) 24nm Toggle-Mode MLC NAND packages from SanDisk (SanDisk has a NAND joint-operation with Toshiba). Anand actually has an X100 in the house and just finished running tests on it a few days ago, so here's a brief comparison of the SSDs used in Macs at the moment (Corsair Force GS is the closest match to the Toshiba SSD):

Note that these are all retail drives; the drives Apple is using are custom and hence the performance may be slightly different (customized firmware). The X100 in the graphs is actually a newer 88SS9175 based X100 (it might be X110 as SanDisk wasn't very specific) so performance of the 9174 based might be slightly different (we shouldn't be anywhere close to U100 speeds, fortunately). The 9175 is essentially just a low-power version of the 9174 with only four channels, so I wouldn't be surprised if the 9174 based X100 was even faster. However, all the drives Apple uses are good performers. The SSD 830 is the fastest of them all but the X100 and Toshiba are trading blows as you can see above. 

I'm not surprised by Apple's move. Apple is constantly buying more and more SSDs (and flash in general) and relying on just one supplier is risky. If one manufacturer has a hiccup in production, Apple still has the channel to quickly buy SSDs from another manufacturer. There's also a possibility that the availability of Samsung PM830 is decreasing because it's now a discontinued product. Samsung told me they still have plenty of stock for OEM and B2B customers but the volume Apple needs is fairly significant and Samsung may not be able to supply that anymore. Samsung does have a newer SSD 840 (Pro) available but it's possible that Samsung's pricing made Apple choose SanDisk.

Either way, SanDisk is now the third OEM to supply SSDs to Apple. I've only seen the X100 appear in the rMBPs so far but it's certainly possible that we will see SanDisk SSDs in other Macs (such as the Air) too. While I'm sure most of the buyers would prefer the Samsung SSD (including myself), the good news is that none of the drives Apple is using are exceptionally bad. Unless your workload is very IO intensive, there shouldn't be a significant difference in real world performance (except when dealing with incompressible data and the Toshiba drive).

Thanks Paul for giving us the permission to use your picture!

Although game bundles are nothing new for the video card industry, in the last couple of years they’ve come back into vogue in a way we haven’t seen in at least a decade. With digital distribution becoming the de facto distribution method for PC games it has transformed video game bundling by supplanting the need to include physical copies of games with video cards.  This has allowed GPU manufacturers and their partners to get a lot more creative in what games are bundled, particularly by allowing them to bundle newly released (or even yet to be released games) instead of months-old games that the logistics of physical distribution require.

At the same time of course we wouldn’t be talking about this had bundles not also proven to be successful for GPU manufacturers, with both NVIDIA and AMD having found a great deal of success in this new wave of video game bundling. By bundling games AMD and NVIDIA can not only latch on to the attention and the hype that goes with the launch of a new AAA game, but it helps their financial sides too. Through bundling they can add value to the product and prop up prices by including something that is valuable to the consumer (but relatively cheap for the manufacturer), rather than having to slash prices and repeat the bloody price wars that marked the start of this decade.

With that in mind, following the success of AMD’s most recent video game bundles – Gaming Evolved (summer) and Never Settle (holiday) – AMD will be orchestrating another bundle for the winter/spring timeframe, to coincide with the forthcoming launch of several Gaming Evolved sponsored titles. The new bundle, dubbed the Never Settle Reloaded bundle, will see AMD putting together a couple different bundles at different price points with a mix of old and new games.

This time around the 7700 series will be left out (no doubt due to the sub-$100 prices those cards now fetch), but the 7800 series and up will now come with at least two games. For most buyers the 7800 series will be bundled with Bioshock: Infinite (launch date March 26^th) and Tomb Raider (March 5^th), while for Asia-Pacific region buyers will get Devil May Cry (February 28^th) in place of Tomb Raider. Meanwhile the 7900 series swaps out Tomb Raider for Crysis 3 (February 19^th), an unlikely (and unexpected) addition to the Gaming Evolved stable. Finally, for buyers picking up a pair of 7900s (or a single 7990), AMD has a six-game bundle composed of the three new games (Crysis, Tomb Raider, and Bioshock), along with Far Cry 3, Sleeping Dogs, and Hitman: Absolution from their previous bundle.

Q: What the heck is an “ultimate reload?” A: Well, it’s like IDKFA.

As always, these bundles are being offered in conjunction with AMD’s major retail/e-tail partners, who will be distributing vouchers for any qualifying purchases. So anyone looking to partake in this bundle will want to make sure their retailer of choice is participating. The vouchers themselves are redeemable through AMD’s website, which will be distributing Origin keys for Crysis, and Steam keys for the other games.

On a final, more speculative note note, in the past we’ve seen NVIDIA run their own bundles opposite AMD bundles most of the time. Consequently we wouldn’t be surprised to see NVIDIA fire back sometime in the next couple of weeks. Their traditional partners Ubisoft, Gearbox, and Capcom all have AAA games due out in the next two months, giving NVIDIA several options if they do decide to run a bundle offer of their own.

NVIDIA put up a blog post recently that goes into some of the details of how Project SHIELD came together. There's plenty of marketing hyperbole in the blog post, but if you've ever wondered what it would be like to be one of the top engineers at a company like NVIDIA, this will give you some ideas. 14-hour work days for weeks or even months? Yup, it sounds like that's exactly what some of the people were doing, but that's what passion looks like I suppose.

The jury is still out on whether or not Project SHIELD will be a success, but you have to give credit to NVIDIA for at least trying something different. Plus, the blog post is an entertaining read, and if you need one more reason to like the idea: Project SHIELD can play Crysis (by proxy).

Intel released the SSD 335 in October last year but the initial launch included only the 240GB version. Today Intel is adding a 180GB model to the SSD 335 lineup. If you recall from our SSD 335 review, the SSD 335 is essentially just SSD 330 but with smaller lithography NAND (20nm vs 25nm). It's still based on SandForce's SF-2281 controller but with customized Intel firmware. The SSD 335 will also be sporting a new case design (see the picture above) later this quarter; drives available today will still have the old plain metal design.

Intel is reporting similar performance at both capacities, although I should note that Intel has always been fairly conservative when it comes to published performance specs (for example, most SandForce OEMs claim up to 90K random write). NewEgg is already listing the 180GB SSD 335 and it's currently priced at $180, which is actually $10 more than 180GB SSD 330 (although I would expect the price to drop once more resellers get the SSD 335 in stock). 

Last year, we looked at Fractal Design’s Core 1000, their entry level chassis that can be had for under $40. It performed “okay”, which is about all you can expect from a budget case. The newly redesigned Core 3000 looks to change more than a few elements of the Core 1000, and in fact other than the “Core” branding it bears little in common with the Core 1000. It’s bigger and heavier for one, measuring 200mm x 444mm x 480mm (7.87” x 17.48” x 18.90”), so this is no longer intended as a small PC box. It also comes with three fans standard: a 140mm fan in front, 120mm fan in back, and 120mm fan up top. Here’s the full list of specifications.

Other items not specifically listed in the above table include an upper HDD cage that can be rotated or removed, allowing the installation of even the largest GPUs (Fractal Design lists GPU clearance as 270mm with the cage or 420mm without—I’m not sure any GPU will ever need 420mm of clearance again, or at least I hope not). Dust covers are also included for the PSU and front fan positions, and the HDD/SSD trays are equipped with silicon vibration dampening grommets.

The Core 3000 USB 3.0 is currently available in some markets, but the US does not appear to be among them. We expect that to change within the coming weeks, and given the popularity of Fractal Design that will likely be sooner rather than later. Of course, the other big change is the doubling of the MSRP, which was inevitable given the improved features, but it looks like the Core 3000 should be competitive with other similarly priced offerings.

We typically don’t cover a lot of supercomputing news outside of the major Top500/Green500 announcements due to the fact that the launches of so many supercomputers are clustered around the official updates to those lists, but the following came across our desk earlier this afternoon from NVIDIA and caught our attention.

Italy’s Cineca computing center has unveiled their latest supercomputer today, a Eurotech created system built on a combination of Intel’s Xeon E5 CPUs and NVIDIA’s Tesla K20 GPUs. While the supercomputer itself is still considered a prototype and is fairly modest overall – its 110 TFLOPS would be good for roughly #250 on the most recent Top500 list – what makes this latest supercomputer particularly interesting (and just a bit snark-worthy) is not its peak performance, but rather its energy efficiency and cooling mechanism. In short, someone has finally paired a K20-based supercomputer with water cooling, with some interesting results.

Cineca's and Eurotech's Water Cooled Eurora Supercomputer

By using water cooling, Eurotech and Cineca have been able to capture the waste heat off of Eurora in a far more dense and more useful manner than air would allow, in turn allowing them to put that waste heat to productive use. By using that heat to not only heat their buildings but also to drive absorption chillers, they’ve been able to reduce the amount of heat that ultimately needs to be removed from the system, and consequently the amount of energy that is spent on cooling.

Adding an extra facet to this is the fact that due to the much greater heat capacity of water over air, using water for cooling – even warm-to-hot water as in the case of Eurora – results in lower processor temperatures than chilled air. In turn, thanks to the positive relationship in semiconductors between temperature and power consumption, these lower temperature processors consume less electricity and thereby produce less heat in the first place.

The end result is that between the use the water cooling and various other optimizations made by Eurotech, they’ve been able to produce a supercomputer significantly more power efficient than other K20 supercomputers, or any other supercomputer for that matter. At 110 TFLOPS performance for 34.7KW electricity consumption, Eurora has an energy efficiency of 3,150 MFLOPS/watt, almost 25% better than the current #1 system on the Green500 supercomputer list, a Xeon Phi system that operates at 2,500 MFLOPS/watt. This makes Eurora the most energy efficient supercomputer available by a wide margin for the moment, and while the next Green500 list won’t be published until June, if Eurora is accepted then will be a likely contender for the top spot on that list.

We won’t dwell on this too much more, but if you would like some more details there’s an interesting slide deck from Eurotech published over on NVIDIA’s website. Ultimately using waste heat for productive use is something of a one-trick pony – there aren’t too many productive ways to use that heat due to its relatively low temperature – but nevertheless it offers an interesting alternative into improving supercomputer performance within a fixed power envelope, particularly at a time where we’re beginning to see diminishing returns from smaller semiconductor manufacturing processes.