Thursday, November 15, 2012

AnandTech Article Channel

AnandTech Article Channel

Lenovo ThinkStation D30 Workstation Review: 16 Cores and 32 Threads Under Your Desk

Posted: 14 Nov 2012 08:01 PM PST

This month is a massive rush of new hardware. Users fond of high-powered portables are probably losing their minds; while Windows 8 and RT are of questionable value to desktop users, hardware designed to take advantage of them is flooding onto the market. Likewise, the SoCs powering smartphones continue to advance at a breakneck pace that hasn't really been seen since the dawn of the Pentium era. It's easy to forget that for how powerful portable technology has become, the potential for desktops and desktop workstations is downright monstrous.

For the foreseeable future, there will always be a need for CAD, video, and 3D rendering workstations. Basic desktop users see grossly diminishing returns on performance after about four logical cores (eight threads), but workstation tasks can still soak up every last ounce of performance you can throw at them. For major businesses where time very truly is money, that means needing the fastest hardware you can find and maintaining uptime for as long as humanly possible. That, in turn, means finding a workstation that's both reliable and easy to service. Lenovo hopes to address these needs with the ThinkStation D30, a dual-socket workstation capable of sporting up to sixteen cores and dual NVIDIA workstation cards (including the Quadro 6000 and Tesla cards for Maximus support).

Android 4.2 Hits AOSP, Broadcom Replaces NXP with NCI Stack for NFC

Posted: 14 Nov 2012 02:08 PM PST

Yesterday source code for Android 4.2 was pushed to AOSP (Android Open Source Project), paving the way for other developers to begin building their own Android images and looking through the source code changes made in Android 4.2 to the core of the Android OS. In conjunction with AOSP push, the Android 4.2 platform highlights page went live with more detail about new features in the point update from 4.1 to 4.2. We went over a number of the major user-facing features in our Nexus 4 and Android 4.2 review but there are a bunch of other noteworthy features outlined in the platform highlight details, including even more improvements made to the 2D rendering engine, WebView rendering engine (which doesn’t apply to Chrome unfortunately), and the new inclusion of GPU compute for Renderscript Compute which is an Android API for running code on CPU, GPU, and DSP. I’m sure we’ll be hearing even more about Renderscript Compute in the coming weeks. I’m burying the lead a bit here, because today Broadcom confirmed some observations I made about the Nexus 4 and Nexus 10 and their NFC solutions.

In the Nexus 4 review I noted that Broadcom’s BCM20793 appeared to have taken the slot of the up until now relatively ubiquitous NXP PN544 controller, which debuted with the Nexus S and has since been the go-to NXP controller for OEMs. In addition to taking the hardware slot, in the platform highlights page there’s a note about Broadcom also fully open sourcing its NFC NCI (NFC Controller Interface) stack which is a new NFC Forum specification designed to standardize the abstraction layer between NFC controller and SoC. Back when the Nexus S launched Google licensed and open sourced the NXP software stack for their NFC controllers, which has been inside Android ever since. I’m told that the NXP stack was previously tailor made to their hardware, and that with an NCI compliant stack OEMs will have greater flexibility when choosing an NFC controller. It makes sense that Google would want a stack with as much flexibility as possible and this gets changed up with 4.2.

I learned some details about BCM20793 as well. Broadcom’s NFC controller is built on a 40nm process and is compatible with both SIM and embedded secure elements (the Nexus 4 uses an embedded SE). In addition Broadcom believes its low power detection mode offers significantly lower idle power consumption than other solutions. Rather than inducing a field and polling for a tag or nearby device on a slotted cycle while the device is on, Broadcom’s “low power target detection” detects the change in antenna load caused by another NFC antenna being brought in proximity, and then polls. I’m told that this results in their controller inducing a field and burning current orders of magnitude less than other readers, and thus giving a fairly large power savings, in addition to the process advantage from 40nm. The obvious next step for Broadcom is to integrate this IP into a Wi-Fi, Bluetooth, FM combo like BCM4334, I wouldn’t be surprised to see something happen like that in the near future.

When I saw the BCM20793 in close proximity to the TI BQ51051b Qi wireless power receiver inside the Nexus 4, I suspected that part of the NXP shakeup might also involve some coexistence communication between the two. While there was RF planning involved (thankfully Qi works at a lower frequency than NFC), the NXP and Qi charging solutions still use different inductive loops and don’t need any collaborative communication. We haven’t talked a lot about the Nexus 10 or formally reviewed it yet, but interestingly enough has multiple NFC points of interest and is able to communicate through the front in addition to the back, and also appears to be BCM2079x based.

Source: Android Building, Android Platform Highlights, BCM2079x Family

The Xeon Phi at work at TACC

Posted: 14 Nov 2012 09:44 AM PST

The Xeon Phi family of co-processors was announced in June, but Intel finally disclosed additional details about the first shipping implementation of Larrabee. In this short article we'll go over the different Xeon Phi SKUs, what kind of software runs on it and how the Xeon Phi are implemented in a supercomputer.

We had the chance to briefly visit Stampede, the first Supercomputer based upon the Xeon Phi in Austin, TX. Stampede is the most powerful of the supercomputers at the Texas Advanced Computing Center (TACC).

Green500 Supercomputer List Released: Intel Takes Top Spot, Followed By AMD, NVIDIA

Posted: 14 Nov 2012 09:30 AM PST

Coinciding with the publication of the Top500 supercomputer list earlier this week, the Top500’s sister list, the Green500, was published earlier this morning. The Green500 is essentially to power efficiency what the Top500 is to total performance, being composed of the same computers as the Top500 list sorted by efficiency in MFLOPS per Watt. Often, but not always, the most powerful supercomputers are among the most power efficiency, which can at times lead to surprises.

Much like the spring Top500 list, the spring Green500 list was dominated by IBM BlueGene/Q systems. With the assembly of a number of new heterogeneous supercomputers since then however, not only has BlueGene/Q been dethroned from the Top500 list, but now the Green500 list as well. In its place on both lists are systems using co-processors from all of the big three: Intel, AMD, and NVIDIA.

For the latest Green500 list, the #1 spot goes to Beacon, a Xeon Phi based supercomputer running at the National Institute for Computational Sciences. At only 44.89 KW, Beacon is a much smaller installation than the likes of the major BlueGene/Q supercomputers or Titan (placing just #253 on the Top500), but Xeon Phi makes a very strong showing here as Intel’s first retail MIC co-processor. Altogether Beacon hit 2499.44 MFLOPS/W, nearly 400 MFLOPS/W higher than the BlueGene/Q computers it has surpassed.

Meanwhile at #2 on the Green500 is King Abdulaziz City for Science and Technology’s SANAM supercomputer, which is the only computer on the list powered by AMD GPUs. SANAM uses 420 of AMD’s recently announced FirePro S10000 cards, which are in turn each composed of 2 of AMD’s Tahiti GPUs. While AMD has had a significant showing in the Green500 list for several years now from the CPU side of things, they have never been a contender as a co-processor vendor, so this is a significant breakthrough for AMD and their first modern GPU compute architecture, GCN. Though much like the Xeon Phi powered Beacon, SANAM is a relatively modest supercomputer; its 2351.1 MFLOPS/W efficiency being spread among only 179KW of total power consumption (making it #52 on the Top500).

Finally, at #3 is Titan, the recently launched Tesla K20X based supercomputer at Oak Ridge National Laboratory. Titan is the current #1 computer on the Top500 list and larger than any other computer on the Green500 list, so along with their top showing on the Top500 list NVIDIA can now also claim to be powering one of the most efficient supercomputers in the world, a significant boost in prestige for their Tesla division. At 2142.77 MFLOPS/W for power efficiency Titan can’t quite match the top Intel and AMD systems, but it’s enough to push past BlueGene/Q just as it did on the Top500 list.

What’s interesting from all of this data is that of the top 10 computers on the Green500 list, the top four computers on the list are all heterogeneous systems using co-processors; the previous 3 systems being Intel, AMD, and NVIDIA, followed by NVIDIA again at #4. Just 6 months ago the CPU-only BlueGene/Q dominated the list, so for Intel, AMD, and NVIDIA to rocket to the top is a significant achievement for GPUs and GPU-like processors. The use of co-processors means that more traditional x86 CPUs are also along for the ride at the top of the list, with Intel and AMD each splitting the top of the Green500 list at 2 each. Ultimately this isn’t the first time heterogeneous systems have had a strong showing on the Green500, but this is the first time they’ve swept the top of the list like this, and marks a major leap in power efficiency for heterogeneous systems that finally puts them on-par with (and beyond) BlueGene/Q.

No comments:

Post a Comment