In this episode Brian and I talk about BenchmarkBoost-Gate or whatever, CPU governor optimizations for mobile benchmarks, the new Nexus 7, Android 4.3 and TRIM, Chromecast and Moto X.

The AnandTech Podcast - Episode 24
featuring Anand Shimpi, Brian Klug

iTunes
RSS - mp3m4a
Direct Links - mp3m4a

Total Time:  2 hours 40 minutes

Outline h:mm

BenchmarkBoost - 0:00
CPU Governor Optimizations - 0:53
Nexus 7 - 1:10
Android 4.3/TRIM - 1:30
Chromecast - 1:44
Moto X - 2:04
Comments Locked

45 Comments

View All Comments

  • et20 - Sunday, August 4, 2013 - link

    Do you listen to yourselves?

    First you say that you must use benchmarks to quantify performance because manufacturers only pay lip service to user experience.
    And in the next breath you admit the tools you use are hopelessly inadequate at actually measuring real life performance.
    Who's just paying lip service again?

    You are not entitled to screw up the mobile industry like you did the desktop industry with your reductive tests and braindead charts. I applaud the manufacturers for not letting you poison consumers with ridiculous, reductive and irrelevant test results simply because your engineering background biases you to try.

    If you have any doubt about whether that actually happened you only need look at the recent state of the quality of input and output devices on desktops and laptops. The lack of clear cutting numbers and charts doomed critical parts of the user experience for the better part of a decade. We had to suffer through decreasing quality of displays and keyboards and stagnating battery life because they were at most an afterthought in the face of the almighty performance and price.
    The primary reason for that was because performance and price were the easiest to benchmark and chart.

    The bottom line is that user experience does not belong on charts. Never has and never will.
    Trying to put it there is the reason why most people hate or feel stupefied by desktops and love their tablets and smartphones.

    What's the point of you trying to replicate the benchmarking of components the manufacturers do internally? That is "simply wrong", not the people questioning your use of benchmarks.
    In mobile you are dealing with integrated products. The more you try to dig deeper and look separately at hardware or software components, the less useful your reviews will be for informing purchase decisions because mobile products win exactly by being much more than the sum of their parts.

    So stop. You are trying to make things worse for everyone except a ridiculously small minority of people who get their kicks out of feeling superior for being able to buy or recommend worse products that can score higher on some arbitrary test.
  • jimbo2779 - Sunday, August 4, 2013 - link

    What in the flying flip-flop are you talking about?

    There isn't a single site on the net that breaks down technology like Anandtech does and to try and shift some sort of blame for whatever imaginary problems you have with the industry is just insane. You are correct in that AT probably does the most quantitative testing out of all of the different tech sites but I do not see that as an issue when it is accompanied by the most qualitative tests and descriptions.

    You are really in the wrong place to be pointing fingers as this site really does give by far the most in depth reviews of all and although sometimes these can be full of intricate details of the inner workings of whatever device they are reviewing they ALWAYS talk about the user experience in just as much depth.
  • Razorbak86 - Sunday, August 4, 2013 - link

    "You mad, bro?"
  • Mondozai - Sunday, August 4, 2013 - link

    My thoughts exactly reading his long, incoherent rant.
  • This Guy - Monday, August 5, 2013 - link

    These guys would have focused more on the positive features of the Moto X if your marketing gurus stopped using jargon to lie about specifications.

    AnandTech always concludes a review of a deivce with comments on it's user IO. Many reviews have pages dedicated to the quality of it's display. Poor buttons/keyboards/trackpads and screens always get criticised.

    As for internals, SSDs, batteries, low voltage chips, capacitive screens, digital pens and higher tier wifi solutions can make a significant improvement to user experience in certain devices when compared to cheaper solutions.
  • Excors - Sunday, August 4, 2013 - link

    On the mobile benchmarking, there was a comment that you're limited in what you can do in two days before a phone is released - there's no chance of rooting the phone to get a better understanding of what it's doing internally.

    I assume getting a review up quickly is vital for catching the buzz when a new phone comes out - but the phone's lifespan can be around a year (at least for the Galaxy S series) and most people don't buy it immediately after release. Wikipedia claims the S4 sold 10M after the first month and 20M after the second month, and the S3 sold 20M in three months and 50M in ten months - the sales don't drop off extremely sharply.

    So perhaps there is enough interest to justify a followup review of a phone, several weeks after its release, to inform people who are making purchasing decisions after that point? That would give time to go deeper - running more thorough tests, looking further into the internals, reading and verifying feedback from early adopters, etc - and then combine all the information into a new comprehensive review (not just a fragmented series of articles). That review probably wouldn't make a big splash but it'd be pretty useful in the long term for the majority of people who end up buying the device, so maybe it'd be worth spending the time on something like that.
  • xTRICKYxx - Sunday, August 4, 2013 - link

    I feel like you guys should develop your own benchmark for Android and possibly iOS. I understand your frustration and perhaps an in-house benchmark would be a good alternative and be supplemental to the 3rd party benchmarks that you currently use.
  • Krysto - Monday, August 5, 2013 - link

    That's actually a good idea. I'm sure they considered it before, but maybe not too seriously. But now AnandTech is respected enough in the industry and popular enough with readers, that if they have the money, they could start doing that. The readers could keep nagging OEM's to test their devices with AnandTech's benchmarks before releasing them, if they want to be taken seriously regarding their performance/battery life claims.
  • willis936 - Sunday, August 4, 2013 - link

    Hey guys. This is my first post on anandtech but I've been reading for a year or two. I'm a junior in EE and work at a lab run by my uni doing interop and conformance testing of networking equipment. My consortium actually tests the phy and protocol of mobile cameras and displays so my first glimpse into any industry is the mobile one but I'm also keeping a close eye on storage and WiFi. I've been to a face to face spec meeting and have been paying close attention to up and coming hardware and how it's embedded in phones. This podcast was really insightful for sifting through the bs, getting the big (hardware/kernel/os/application interplay in performance) picture and figuring out why things the way they are in phones today.

    I personally have a rooted N7 gen. 1 that's been slowing down. I ran fstrim through a play store app on the data, cache, and system partitions. I wouldn't have believed it was eMMC I/O performance chugging the UI if I hadn't seen the difference myself.

    It's nice to hear a level voice about this stuff and I look forward to more podcasts. The entire internet has been silent about 802.11ac the past month. It's due to be finalized in November and I'm sure official 11ac IP will be in products for the holidays. If you guys have any more to say about that, Intel HD5200 laptops, silvermont, or UWB I'd definitely listen in.
  • willis936 - Sunday, August 4, 2013 - link

    As a quick addendum: I've been excited about the moto x since I first heard rumblings about it a few months ago and most of my friends don't even know about it yet. The thought of that kind of very silent rumbling might have something to do with the marketing budget hadn't even entered my mind. It's an interesting thought.

    My opinion of the moto x right now was formed the instant I read the paragraph about pricing and availability in your preview. $200 on contract, carrier exclusives. I'm off the boat full stop. That's not what I was expecting from Google's divine intervention. Yes the gimmicks are cool and introduce innovation in an otherwise uncreative and stagnant market but when I first heard about moto x my impression was that it would bring nexus to the masses. I'm on a two year old HTC thunderbolt. I have unlimited LTE that I'm very closely guarding so an upgrade is out of the question. I'm now in the very tough spot of finding a decent phone on VZW. I'm not sure if anyone else has noticed but that doesn't exist, and it hasn't for years. Nexus is supposed to rip the carrier's hands off of the android experience because right now it's suffocating under VZW's massive bureaucratic and profit making weight. The moto x missed this vision and not by any small amount. To me it looks like another Motorola phone, not the first vertically integrated Google phone.

Log in

Don't have an account? Sign up now