Does Apple ever try to boost benchmarks with the iPhone?

Discussion in 'iPhone' started by Kendo, Jul 30, 2013.

  1. Kendo macrumors 68000

    Joined:
    Apr 4, 2011
    #1
    I'm curious by this report by Samsung boosting their S4 benchmarks by creative means such as using programs to overclock the CPU and GPU during certain tests. Has Apple ever done something like this? Normally I call bologna but this is AnandTech that uncovered the cheating.

    http://www.engadget.com/2013/07/30/samsung-benchmarks/

    http://www.anandtech.com/show/7187/looking-at-cpugpu-benchmark-optimizations-galaxy-s-4

     
  2. Lucille Carter macrumors 65816

    Joined:
    Jul 3, 2013
    #2
    Apple is a WAY more conservative company when it comes to hardware. Their phones will have ample power to run it's OS and apps. You really do not need more than that unless you are just a spec freak!

    Lucy!:rolleyes:
     
  3. cynics macrumors G3

    Joined:
    Jan 8, 2012
    #3
    I'm missing something or that article is BS.

    Its obvious that a processors clock speed adjust depending on task and in that case yes Apple does that in every device they have from iPhone to Mac. Game might use 1 ghz while a benchmark will max out the processor at 2ghz (example).

    However they don't "overclock". And I could go into the details about how simple it is to track real time clock speed on Android and bore you guys but I'll end with this.

    If you root and overclock an S4 even slightly your benchmarks get better then they were prior. If it automatically and secretly did this then it wouldn't change now would it?
     
  4. Merkie macrumors 68020

    Joined:
    Oct 23, 2008
    #4
    I doubt it. Only Apple sells iOS devices. For Android devices, there is a higher need for comparison because there are so many devices that are alike and hardly differ from eachother.
     
  5. gnasher729 macrumors P6

    gnasher729

    Joined:
    Nov 25, 2005
    #5
    Apple makes compromises between speed and energy consumption. An iPhone could run faster if Apple didn't care about battery life. One would hope that every company would do. As a result, benchmarks will not run as fast as the hardware could do, and benchmarks will run nowhere near as fast as hardware that Apple _could_ build. (It would be no problem for Apple to build an ARM chip that runs at twice the speed as long as you don't care that your battery gets empty very, very quick).

    If a device detects benchmark applications and runs faster with worse battery life when these apps are running, that would definitely be cheating. Since Apple doesn't use any benchmarks for marketing, I doubt that they spend any effort on cheating in that area.
     
  6. KenAFSPC, Jul 31, 2013
    Last edited: Jul 31, 2013

    KenAFSPC macrumors 6502a

    Joined:
    Sep 12, 2012
    #6
    Over the years, several graphics chip manufacturers have been caught running similar cheats with 3D benchmarks. I believe Anandtech even caught some of them.

    This is not necessarily some separate Android app that runs in the background that detects running software, which you can avoid by using a different ROM. These optimizations (e.g., those that overclock the device upon detection of certain benchmarks) are likely built into either the device's drivers or into the CPU firmware. You can't run the device without them. They may engage the overclock upon detection of the benchmark software (driver), or upon detection of certain routines associated with the benchmark software (driver, on-chip firmware).
     
  7. ucfgrad93 macrumors P6

    ucfgrad93

    Joined:
    Aug 17, 2007
    Location:
    Colorado
    #7
    Agreed. I don't think Apple is concerned with specs as long as the OS runs fine on a particular processor.
     
  8. lazard macrumors 65816

    Joined:
    Jul 23, 2012
    #8
    The GPU wasn't overclocked. The Exynos Octa uses a PowerVR SGX 544MP3 which is spec'ed to run up to 533 MHz. During the benchmarks Anandtech said the GPU clocked in at 532 MHz.
     
  9. scaredpoet macrumors 604

    scaredpoet

    Joined:
    Apr 6, 2007
    #9
    One would have to root their S4 and then re-run Anandtech's tests to see what happens.

    At the end of the day though, IF Samsung is doing this (and IF Apple were to do this), it's all really childish and silly. People who are running benchmarks on their smartphones and tablets (and even desktops and laptops for that matter) are by and large doing it solely to be able to brag to others about how much more amazing their device is because their bar on the graph is bigger than your bar on the graph.

    On the other hand, most people with actual work to do on their devices care more about whether those devices can perform those tasks well, and without getting the in way. If your device runs what you need it to run without stalling, crashing, or otherwise frustrating the user, that's really the only benchmark that counts.

    Maybe Samsung IS juicing the numbers, so that the fandroids can giggle giddily about how amazing their phones are at smoking iDevices. So what? Let them!
     
  10. teknikal90 macrumors 68030

    teknikal90

    Joined:
    Jan 28, 2008
    Location:
    Vancouver, BC
    #10
    I think the point was the s4 was made to distinguish benchmark programs and automatically overclock itself higher than normal when just those specific programs are running
    In other words, it's not running the way it would run programs under normal usage. Hence effectively "lying" to the benchmark
     
  11. cynics macrumors G3

    Joined:
    Jan 8, 2012
    #11
    532 MHz is not a clock speed higher then 533 MHz. Therefore not overclocking.

    How is that lying?

    If another app runs perfectly smooth and fine at 400 MHz why would you want it to run higher? Waste more battery needlessly?

    Some of you guys are way too into worrying about benchmarks. Just look at real world performance. This is the second thread about this going on in the iPhone section of an Apple forum....so really....who worries about benchmarks again?
     
  12. teknikal90 macrumors 68030

    teknikal90

    Joined:
    Jan 28, 2008
    Location:
    Vancouver, BC
    #12
    We're not worrying about benchmarks... but since the information is 'news' it is up for discussion. I don't know about everyone here but I'm not seeing this as an opportunity to back Samsung/Android... If Apple had done the same, I wouldn't be surprised.

    Now, I understand where you're coming from, but I don't think thats where the point lies.

    Here's the thing - the S4 has an 'overclocking' algorithm built in - as you said, it can go up 533MHz, depending on load.

    Depending on load.

    Hence, on demanding tasks, the system recognizes the need to bump up speed upon receiving demanding tasks and does so accordingly.

    The additional code on the S4 to boost speed for certain specific apps is cheating because it bypasses that dynamic performance system that in any other case would be in effect. It tells the phone to get ready and to boost up performance BEFORE the load is even put on it, completely bypassing the dynamic performance system, thereby artificially increasing performance. In effect, the S4 would 'hit the ground running' and would keep running for longer that it normally would. This isn't realistic as normally, the S4 wouldn't know to 'get ready' and boost up its clock until after the demanding task arrives and it may not keep that speed for as long as it does in the benchmark with the additional code.
    It's like giving a 200m sprinter an early start before the pistol and telling it to run at 100m pace for the whole of the 200m - if that analogy makes sense.

    my 2 cents.
     

Share This Page