Tuesday Shortbread

7 Up

  1. CVG: Valve ‘frustrated by lack of innovation‘ in PC hardware, ‘jumping in’
  2. CPU World: Intel launches mid-class and budget desktop CPUs

    and Intel adds pricing for new mobile processors and

    AMD plans six-core FX-6350 “Piledriver” CPU

  3. Fudzilla: Haswell has on-die voltage regulator
  4. DigiTimes: Intel may need to adjust roadmap for PC-use Atom

    processors and PC DRAM inventory remains high, says inSpectrum

  5. TweakTown reviews Nvidia GeForce GTX 660 2GB
  6. ZDNet: Lenovo exec. praises Microsoft Surface tablet and

    Mac OS X overtakes Windows Vista in global market share

  7. Softpedia: Ubuntu 12.10 adds encrypted installation


Tuesday

  1. Fudzilla: ARM boss warns of trouble ahead
  2. X-bit labs: Fujitsu Semiconductor to transfer chip assembly & test plant to J-Devices
  3. Beer brings people together: Obama, homebrewers,

    and online communities – OpenSource.com

  4. Hardware Heaven’s Pimp My Rig competition with PowerColor HD 7990 Devil 13
  5. Dealzon’s deals: $120 off 15.6″ Toshiba C855 Pentium B970, $400 off

    14″ hp Envy Spectre i5-2467M, $40 coupon for 27″ ViewSonic 1440p

    IPS monitor, and $5 coupon for 90GB Kingston SSDNow V+200 SSD

Mobile

  1. The Telegraph: Russia snubs Google for Android-style tablet
  2. The Verge’s exclusive: Nokia Lumia 920 to include

    wireless charging, 32GB storage, and 8MP camera

  3. Fudzilla: ZTE announces an Intel phone
  4. Madshrimps review Ebode BT Sound BTS30 speaker
  5. The Boston Globe: Smartphone apps track users even when shut down

Software and gaming

  1. Neowin: Windows Server 2012 RTM to be available for purchase on 9/4
  2. Linux.com: Xfce 4.10, the sane Linux desktop
  3. Eurogamer: Halo 4 spoilers reveal main antagonist, among other things
  4. Guild Wars 2 game status updates

Systems, storage, and networking

  1. Phoronix: The ARM Cortex-A9 can beat out the Intel Atom
  2. KitGuru’s ASRock Fatal1ty X79 Champion motherboard review
  3. ocaholic’s Asus Maximus V Formula review
  4. Bits and Chips review Asus P8Z77-V Pro (in Italian)
  5. Guru3D’s 16GB Corsair Dominator Platinum

    DDR3 2133MHz memory kit review

  6. Hardware.Info’s 256GB OCZ Agility 4 review
  7. The SSD Review on 256GB Crucial V4 SSD
  8. Rbmods on TP-Link TL-WR702N Nano 150Mbps wireless N router

Multimedia, power, and cases

  1. HT4U reviews EVGA GTX 660 Ti SC 3GB (in German)
  2. ocaholic reviews Cooler Master Storm Sonuz headset
  3. Legit Reviews on Plantronics GameCom 780 headset
  4. Hardware Secrets reviews Gigabyte Aivia Osmium keyboard
  5. Techgage reviews Logitech G600 MMO gaming mouse
  6. techPowerUp! reviews 850W LEPA G850-MAS PSU
  7. ThinkComputers reviews Azza Genesis 9000 case
  8. TweakTown reviews Sentey GS-6070 II Abaddom case
Comments closed
    • WillBach
    • 7 years ago

    [quote<]Phoronix: The ARM Cortex-A9 [url=http://www.phoronix.com/scan.php?page=article&item=gentoo_arm_x32&num=1<]can beat out[/url<] the Intel Atom[/quote<] It's true (in certain circumstances)! Atom has (in the realm of low-power, low-cost processors) a strong memory subsystem (cache, TLB), a reasonably fast pipeline (including decode and forwarding), okay SIMD support, and good branch prediction. It does also have weaknesses: [list<] [*<]Long divide latency. [/*<][*<]In order issue. [/*<][*<]Moderate SIMD throughput. [/*<][*<]Assorted other weakness as seen in the Atom optimization guide (can't find the link now).[/*<] [/list<] With the caveat that the LZMA compression benchmark may have been I/O bound in the netbook (256 MB input, netbook hard drive), I think that the A9's success in many of those benchmarks (which were skewed towards scientific/HPC computing) was mostly due to it's NEON unit. I have no idea what explains the N-Queens result, except maybe repeated non-aligned accesses or accesses across cache lines, which can make Atom stumble. Edit: I have [b<]no[/b<] idea why Atom lost in the LZMA compression benchmark unless there's a lot of random reads going on. Thanks to willmore for pointing out the lack of fast flash memory on the ARM boards.

      • willmore
      • 7 years ago

      The two ARM boards used SD cards for storage. Those’re not known for fast read speeds. Even a low end laptop HD is going to beat an SD card in read speed.

        • WillBach
        • 7 years ago

        Thanks, willmore. For some reason I thought they were using one of the faster flash varieties, like the kind for internal memory.

    • dpaus
    • 7 years ago

    Hey, Apple users: the FBI knows [url=http://www.techdirt.com/articles/20120904/07434620264/hackers-get-personal-info-12-million-apple-users-fbi-laptop.shtml<]where you live. And your phone number. And what Apple devices you have. And their Unique ID #s. And their Push Notification service tokens. And your favorite color. And your dog's name. And....[/url<] Aren't you glad that Apple 'cooperates' with the FBI to ensure your personal data remains private?

      • Srsly_Bro
      • 7 years ago

      It’s all part of our Lord Jobs’ master plan that we as humans aren’t able to understand. Questioning his will is blasphemy.

        • Deanjo
        • 7 years ago

        [quote<]Questioning his will is blasphemy.[/quote<] should be, [quote<]Questioning [b<]H[/b<]is will is blasphemy.[/quote<] You must use proper capitalization when referring to the Almighty.

      • Deanjo
      • 7 years ago

      FBI denies:

      [url<]http://news.cnet.com/8301-1009_3-57505925-83/fbi-finds-no-evidence-that-antisec-hacked-its-laptop/[/url<]

        • Srsly_Bro
        • 7 years ago

        Did you expect them to admit they got hacked? lol

          • Deanjo
          • 7 years ago

          Well they have admitted before when they have been hacked. After crunching the numbers of how large the csv file would be for 12 million UUID’s with the amount of information they supposedly got, I do have reservations about believing what they say they have. Turns out to be a pretty damn large file.

    • BIF
    • 7 years ago

    [quote<]Mac OS X overtakes Windows Vista in global market share[/quote<] Really? So soon? [eyeroll] This is actually pretty suprising to me that it didn't happen about six months after Vista was released, what with all the whining and moaning about Vista by .. well EVERYBODY. Except me...

      • superjawes
      • 7 years ago

      It’s almost as if the whining against that newest version of Windows was just hot air…

        • BIF
        • 7 years ago

        By comparison, yes I think you’re right!

        Oh wait…which new version are we talking about?

        😛

    • codedivine
    • 7 years ago

    [quote<] Phoronix: The ARM Cortex-A9 can beat out the Intel Atom [/quote<] That is the most ridiculous review I have ever read. But then, most Phoronix articles fall in that category. As an example, the writer had no idea what system he actually was testing. Kept referring to N450 as dual-core. A single Cortex A9 core is vastly slower per-clock than an Atom. While current Atoms may have a questionable perf/W, a dual-core Atom based system will beat a dual-core Cortex A9 in performance at the same frequency.

      • Hattig
      • 7 years ago

      That would be pretty unfair – putting a quad-thread Atom against a dual-thread A9 though.

      The article clearly states that it is a dual-thread Atom right at the top.

      Regardless, Intel sells a 1.6GHz dual-thread Atom for mobile devices, whilst you can get mobile dual and quad-core A9s.

      In the end it is the performance that matters, not how it is achieved (as long as perf/W and perf/$ isn’t made bad in comparison), and ARM has gone down the smaller, more cores route for the A9.

        • Narishma
        • 7 years ago

        [quote<]That would be pretty unfair - putting a quad-thread Atom against a dual-thread A9 though.[/quote<] Then don't do that. Just pick a dual-core Atom and disable HT in the BIOS.

          • Hattig
          • 7 years ago

          But what’s the point of that? Intel are selling single core Atoms with dual-thread enabled for mobile uses, so that’s what you want to compare to the ARM options on the market.

            • Narishma
            • 7 years ago

            In that case, I don’t see the point of underclocking either, since Intel doesn’t sell Atoms running at 1Ghz. The N450 is also not meant for mobile uses.

            • willmore
            • 7 years ago

            And that’s why they didn’t compare power consumption–it would be unfair to the N450 as it’s not a mobile part.

            The putpose of underclocking is to remove one variable from the test. If all processors work at the same clock, and we have data for those parts at other clocks, we can see what role clock speed has in performance scaling.

            It’s unfortunate that they didn’t have a wider array of systems available for the comparison, but they’re not a huge web site and they don’t have unlimited funding. Most of the hardware they have is donated. “Beggers can’t be choosers.”

            • Goty
            • 7 years ago

            Most A9 devices don’t ship at 1GHz, either.

            • Hattig
            • 7 years ago

            It still uses the same Atom core as the latest Atoms however. Indeed it probably has an advantage due to the non-mobile memory controller.

            However comparing power consumption would have been grossly unreasonable as it isn’t the actual mobile-phone-oriented Atom.

            • NeelyCam
            • 7 years ago

            How do you decide what’s fair and what’s not?

            If you normalize to performance/watt, Ivy Bridge beats everything.
            If you normalize to performance, Ivy Bridge beats everything.

            Why normalize to “per-core, per-clock” while ignoring power consumption? How about AMD modules? Are they dual-cores or single-core dual threads?

            The only fundamental metrics are performance/watt and performance.

            • BobbinThreadbare
            • 7 years ago

            Agreed, this focus on IPC is nearly meaningless from a consumer perspective.

            Edit: there is one other important metric, which is performance given a specific power envelope. It doesn’t matter if Ivy is more power efficient per performance if they can’t produce a chip that runs at 1 watt, but ARM can.

            • Deanjo
            • 7 years ago

            [quote<]The only fundamental metrics are performance/watt and performance.[/quote<] Those are not the only ones, especially when dealing with portable devices. Standby and idle consumption are also important factors especially when those devices are rarely running full tilt but spend a heck of a lot time just being ready.

      • WillBach
      • 7 years ago

      I don’t think Phronix was trying to make an SoCs core review per se, I think it was an examination of Linux performance on different cores. The Linux kernel is portable, but getting an entire distribution up and running (like Ubuntu did, which is what Calxeda uses (should I say claims to use? No one’s seen real, confirmed, industry benchmarks.)) isn’t always an easy proposition, especially when drivers and real-world performance are considered.

    • Arag0n
    • 7 years ago

    [quote<]ZDNet: Lenovo exec. praises Microsoft Surface tablet and Mac OS X overtakes Windows Vista in global market share[/quote<] Once again lenovo showing how to not be a dumb-ass CEO. Just because that comments make me want to buy a Lenovo tablet instead of surface.

    • NeelyCam
    • 7 years ago

    [quote<] III. Fudzilla: ZTE announces an Intel phone[/quote<] When the Lava phone beat everybody in SunSpider, somebody said ICS will fix this for ARM, while Intel won't gain any benefit because it's all optimized already. Well, Engadget's preview shows a healthy improvement for x86.. EDIT: Link to Engadget: [url<]http://www.engadget.com/2012/08/30/zte-grand-x-in-benchmarks/[/url<]

      • willmore
      • 7 years ago

      Link?

        • NeelyCam
        • 7 years ago

        Added

          • willmore
          • 7 years ago

          Looks interesting. Many phone reviewers with experience are careful to point out that there’s a lot to the performance of a phone (or even just the browser on a phone) than what CPU it uses. The OS and browser can make a huge difference.

          So, unless you’re controlling for more variables than that quick preview seems to show, their results don’t tell us a whole bunch about the performance of Intel chip in the handset space.

          One quick example is for the iPhone 3GS. On OS 4.2.1, the Safari browser gets a sun spider result of 4769ms. Under Chrome, it baloons to 15537ms. That’s the same SoC, the same OS, but changing the browser.

          For CF bench, one of the reasons it may not have done well on this device is that this benchmark tries to run some native code (as well as some Jave code). Now, did they build in x86 native code? If they did, how does the benchmark determine if the device is compatable with some of the native code blobs it includes? A simple failure to use the native code blob could severely impact the performance score of a device.

          We’ll have to see what the ARM based Android world looks like when this phone finally makes it to market. It should be interesting.

    • NeelyCam
    • 7 years ago

    [quote<] I. Fudzilla: ARM boss warns of trouble ahead[/quote<] So, either the PCs aren't getting killed by smartphones/tablets after all, or ARM is losing market share to Intel. Either way, watch that insane ARM P/E ratio tumble

      • blastdoor
      • 7 years ago

      Or 3, the smartphone business is finally being hit by recession/depressions around the world.

      • MadManOriginal
      • 7 years ago

      Poor NeelyCam, still obsessed with ARM’s P/E.

      It’s kind of an odd statement by the CEO, much of it seems to be focused on their hiring intentions and specifically hiring new graduates versus experienced engineers. But what’s strange about it is that just last week there was news about Apple and Qualcomm looking to pay big to guarantee production from TSMC: [url<]http://www.bloomberg.com/news/2012-08-29/apple-qualcomm-spurned-in-bids-for-exclusive-tsmc-chip-supply.html[/url<] That makes me think things are supply constrained rather than demand constrained.

    • Voldenuit
    • 7 years ago

    [quote<]The Verge's exclusive: Nokia Lumia 920 to include wireless charging, 32GB storage, and 8MP camera[/quote<] I really hope they're wrong about that 8 MP camera bit, especially since the 920 is supposed to sport 'Pureview' branding.

      • Arag0n
      • 7 years ago

      Me too, but if finally is only 8Mpx instead of 16Mpx or 21Mpx that are rumored then we will need to wait for final shots to see whether it´s a real good camera despite low megapixels, but I won´t expect it to be as awesome as the Nokia 808 neither close to it. The megapixels myth was debunked some years ago because sensors with the same size and higher number of pixels at the end receive less light per pixel and every pixel is more noisy. However, Nokia acknowledge that but offset the problem and increased the final result by doing an inteligent downsampling. So, 8Mpx may not mean a bad camera, but not a top-notch next-generation one. Everything else in that phone still looks so sweet. Just missing some way to have TV-out for using the streaming apps with a TV.

        • Arag0n
        • 7 years ago

        well, seems i might have been wrong. Seems nokia found a way to compensate the size of the sensor with exposition time doing different tricks. Still, I will wait for reviews to do a final claim about if the camera is really revolutionary, but some demos are mind-blowing, just hoping it´s true. However, I tho the SD slot was a given and seems to be missing. That´s my only major disappointment. Everything else is just so sweet, I’m buying one as self-gift for Christmas most surely.

      • Voldenuit
      • 7 years ago

      Downvoted because I want a phone with a good camera? Welcome to Sparta.

        • kiyoku
        • 7 years ago

        I think you got down voted for not knowing 8 MP doesn’t prove anything here. See Arag0n’s reply.

          • NeelyCam
          • 7 years ago

          And whining about getting downvoted is a sure way to get downvoted more

          • Voldenuit
          • 7 years ago

          I didn’t think it was worth getting into phone camera sensor sizes and sensor pitch. You know, 1/1.2″ vs 1/1.7″ vs 1/3.2″, 1.40 um vs 1.75 um. etc.

          Anyway, the 920 has been announced and the specs list a 8MP camera. Looking at the camera bulge (or lack thereof), it looks to be a bog-standard 1/3.2″ sensor with LED flash, and the only real innovation is addition of OIS.

          Coming on the heels of the 808, this is pretty disappointing for the camera enthusiasts. I’m now sorely tempted to pick up an 808 the next time I’m in Asia (to get faster updates than the NA version).

    • jdaven
    • 7 years ago

    “Phoronix: The ARM Cortex-A9 can beat out the Intel Atom”

    Cue Chuckula and NeelyCam in 3.2.1….

      • Goty
      • 7 years ago

      You think it will take that long?

        • NeelyCam
        • 7 years ago

        No

          • Goty
          • 7 years ago

          Thank you for validating my faith in you, Neely.

      • Srsly_Bro
      • 7 years ago

      Don’t encourage them!

        • NeelyCam
        • 7 years ago

        Nothing to see here. Almost zero info on the tested systems. No power consumption data. And the ages old Atom was downclocked from 1.67GHz to 1GHz to accommodate the slower clocks of the ARM chips.

        Reminds me of that Caldexa benchmark that showed ARM beating Intel on power efficiency because they used TDP number for the intel system and guesses its memory power consumption (while using measured power for the Caldexa system)

        Sure, this article may make ARM fanbois feel warm fuzzies, but it really has no bearing in the real world

          • codedivine
          • 7 years ago

          Well the Calxeda benchmarks had various issues. For example, it was pointed out that the performance of the Intel Xeon based systems in those tests was bottlenecked by network chip used, leading to disappointing perf/W numbers for Intel.

          As always, will wait for (competent) third-party tests.

          • Hattig
          • 7 years ago

          It was a very valid test, comparing per-clock performance of a Dual-core A9 vs. Dual-thread Atom (with a single core A8 in there for comparison).

          You can get Quad-core 1.7GHz A9s, so that compares nicely with higher clocked Quad-thread Atoms.

          And in terms of the mobile space, Intel are shipping 1.6GHz single-core (dual-thread) Atoms, and now we can see that a dual-core A9 at 1.5GHz or any quad-core A9 will be a strong competitor to such a chip. And that as such, the Atom is not faster at all, contrary to the Intel sponsored advertisements.

          Yes, it would be nice to actually see some benchmarks of a mobile Atom vs. a mobile A9 in shipping products, and to see power consumption figures over a range of workloads (including many sparse usage scenarios to mimic how mobile devices are used), and then there is cost as well (ARM SoCs being very cheap).

            • Arag0n
            • 7 years ago

            You just resume all the points of the comparison.. Clock by clock so you can compare a single/dual/quad A9 vs an atom at a given frequency. Feels funny that NeelyCam wants to say that ARM works at 1Ghz while there is plenty products selling out there with ARM’s at 1.3, 1.4 and 1.5Ghz dual and quad core.

            • NeelyCam
            • 7 years ago

            And there are Atoms running faster than 1.6GHz. And there are other Intel chips running at way faster than 2GHz.

            The A8 used in the comparison was running at 1GHz; that’s why others were downclocked to 1GHz. Comparing the architectural per-clock performance is fine, but it’s mainly for academic purists, and doesn’t really matter in real life. What matters is how high is the performance for a given amount of power consumption OR how high is the performance period. Per-clock rankings are largely meaningless.

      • chuckula
      • 7 years ago

      Poor JDaven… a day late and a dollar short (already addressed this “news” article in yesterday’s shortbread). Bear in mind, the crippled Atom *still won some of the benchmarks* even in this ridiculous setup.

      EDIT: I’d just love to see JDaven’s reaction if TR decided to show Haswell’s superiority against an underclocked Phenom II system from 2010, which would be the rough equivalent of what happened in this “review”. Something tells me he would scream about how unfair that comparison is… and he’d be right. The difference is that JDaven has a very peculiar definition of “fairness” that only goes in one direction and that can be turned on and off at will to suit whatever outcome JDaven wants.

      • dpaus
      • 7 years ago

      [quote<]Cue Chuckula and NeelyCam in 3.2.1....[/quote<] At 3 minutes after midnight on the back-to-work-day following a long weekend? Quite the taskmaster, aren't you?

Pin It on Pinterest

Share This