TechSpot

HTC boosts One M8 benchmarks, calls it a feature

By Shawn Knight
Mar 31, 2014
Post New Reply
  1. It's no secret that nearly all Android OEMs "cheat" with regard to benchmarks. By cheating, I simply mean that they inject code that is used to detect if a benchmark application is running and pump up the CPU clock speeds...

    Read more
     
  2. wastedkill

    wastedkill TS Evangelist Posts: 1,392   +329

    Ah here we go people will start calling it cheating for optimizing or overclocking because optimizing a device or overclocking a device is practically the same as killing a few people everyone just goes crazy over it.
     
  3. BlueDrake

    BlueDrake TS Evangelist Posts: 350   +103

    Moreso it's cheating being it looks for benchmark apps, and pushes everything to the max as a means to fake results. It's such a common occurance really that you have to question, the product makers as to why they do such things in the first place. Yes you topped the charts on a benchmark, but did you do it with more powerful hardware? Did you somehow make it more powerful all the while, making it more efficient to not drain the battery or did you basically rig the results?

    I doubt everyone looks at benchmarks, to figure out if something is good for them. Sure it's a good idea but as we seen, many Android based devices don't play fair. So you might think it's some super cool phone / tablet, but in the end it might not be nearly as good. All in hopes of pushing sales really, is all they want to do anymore. Obviously this stunt will likely drive sales, but then people will likely complain of battery life.

    Problem is with people these days, you want all your fancy toys but disregard battery life. Then complain half the time, when it doesn't make it through half the day without needing a recharge. Which then leads to people making a fuss, saying they were lied to among other stupid things. It's why the world is full of *****s, they blindly trust what they see or want things without a compromise. Oh well to each their own I guess.
     
  4. wastedkill

    wastedkill TS Evangelist Posts: 1,392   +329

    Your basically saying that all overclocking and optimizing should be made illegal... So optimizing hardware like what AMD/Nvidia/Intel/ARM etc. do should be made illegal by your standards right?

    If you want to push the limits of something you are allowed to there is nothing wrong with it, if you want good battery life on a device GO FOR A OLD NOKIA!!!!! If you want a device that has tons of power and IS ABLE TO PUSH ITSELF MORE then you get one of the new phones with 1-2day battery life and NO phones are not gaming consoles stop using them as such!!!

    Yes a phone can play games but if you expect 1month battery life out of it then you either a(stupid) or b(a genius that has invented a new battery) and no battery life isnt exactly improving the same way CPU/GPU's are improving sorry for your die hard old nokia brick phone mind set but this is 2014 not 1990.
     
  5. MilwaukeeMike

    MilwaukeeMike TS Evangelist Posts: 2,754   +1,107

    Any site that does reviews finds this stuff right away. Anandtech called their benchmark page on their M8 reivew 'Subtle Cheating: New Benchmark Optimizations' The funny this is, Anandtech says "All of our benchmarks are already immune to the optimization, so it’s really a matter of sacrificing integrity for no real gain."

    The benchmarks are far less important to most of us than actual descriptions of the phone's use. When Engadget says this...
    I pretty must just skip over the benchmark graphs anyway.


    Well... if it makes the device look more powerful in reviews, then it isn't a waste of time. And if you can manually turn on the 'high performance mode' that the benchmarks use, then it would have an effect of real-world performance (and battery life). I'd bet though that with it turned on you'll notice no improvement in performance and may notice a decrease in battery.
     
  6. EEatGDL

    EEatGDL TS Maniac Posts: 481   +159

    Dude, I don't think you get it. If you overclock your AMD/NVIDIA GPU or AMD/Intel CPU you'll be benefited across most applications, it's useful and you did it manually; overclocking automatically just for benchmarks -not for any other scenario which will represent the actual 99% of time usage- is giving a false impression and thus cheating, completely different, the user doesn't get a benefit from that.

    This case is more similar to when Intel gave Haswell CPUs to reviewers that could achieve higher frequencies than what the average overclocker would ever get, thus making a false impression of being decent for overclocking.
     
  7. Don't waste your time trying to explain HOW its cheating. Some just will never understand. You can say it in a way that appears so easy and simple to understand but they just don't get it.
     
  8. MilwaukeeMike

    MilwaukeeMike TS Evangelist Posts: 2,754   +1,107


    This is a tougher case though because of battery life; your computer CPU has no consideration. The phone makers turn down their processors to make the battery last longer. So what's more accurate... benchmarks that use a processor setting that you have to manually turn on? Or benchmarks that are based off of a processor speed the majority of users will be using? The latter of course, but I can understand why they do it. They do it because a) everyone else is doing it, and b) they're being 'graded' on benchmarks that are using an under-clocked processor.

    It reminds me of MPG ratings in cars. They're to be used 'For Comparison use only.' If all the smartphone makers are turning their CPUs to max for the benchmarks (which they are), then we can still use the benchmarks for comparison. It's not like we're timing our app launches in milliseconds....

    Honestly, I don't care... I have no idea how much faster a webpage is going to load between two phones with different javascript benchmark scores. This phone has the latest and greatest processor out and all reports say it's snappy, responsive and has terrific battery life. That's good enough for me.
    Now I just need to check under my couch cushions for an extra $600 in loose change.
     
  9. GhostRyder

    GhostRyder This guy again... Posts: 2,191   +590

    Well this all depends on how you look at it to see whether its cheating or not.

    If you would take a normal GPU into account, they benchmark and run things with the boost clock which is a best case scenario so long as the temperature is within a recommended range. So essentially it would be like a program overriding the temperature threshold and allowing the boost clock to run no matter what (Within reason) just for the benchmark.

    In all technicality, this is showing a best case scenario for the CPU on the phones by allowing the phone to run at its peak performance to show what it could be capable of. Now the problem before was that a mode like that would only be achievable on a phone while its plugged in because of battery meters and the system would not allow it to kill the phone battery. With the HTC M8, they allow this mode to be open to the developer which is nice if they want to make something graphically intensive that requires alot of power so this is more than fine.

    I really don't see it as cheating, more of a best case scenario.
     
  10. RebelFlag

    RebelFlag TS Addict Posts: 135   +67

    Seems like if everyone is increasing performance for the benchmarking apps, then the information is still relevent since everyone is doing it. Kind of like giving every student 10 points on a test. Is anyone really buying phones based on CPU benchmarks anyway? If you are worried about CPU and GPU benchmarks on a phone, you should probably get a game boy and a non-smart phone and save some money.
     
  11. MilwaukeeMike

    MilwaukeeMike TS Evangelist Posts: 2,754   +1,107

    On the plus side, having the story about HTC 'cheating' on the benchmarks with the M8 does get it out and in the news again. Nothing wrong with a little free exposure for your phone, especially when many of those stories will have links in them to reviews. HTC will need all the help they can get to compete with Samsung's enormous marketing budget.
     
  12. Jad Chaar

    Jad Chaar TS Evangelist Posts: 6,477   +965

    At least they opened up "high performance mode" to the developers rather than being secretly locked up like before.
     
  13. BlueDrake

    BlueDrake TS Evangelist Posts: 350   +103

    I think you're missing the point entirely, that was stated in the article. It looks for a "benchmarking" app in particular, to give out false information about it's true performance. I've said nothing about overclocking or optimizing, to get the best out of a battery in that respect.

    So would you like to know, if it falls below certain others based on it's hardware? I sure would but the whole issue is, Android devices see the apps and push everything to the max. Which is misleading in the end, because those specific apps don't purposely max out the hardware. It's a fake rating because of the detection, that's all my general point was.
     
  14. wastedkill

    wastedkill TS Evangelist Posts: 1,392   +329

    Really? I just saw that it overclocks & optimizes the hardware but if it somehow fakes the data then ye it sounds bad.
     
  15. hahahanoobs

    hahahanoobs TS Evangelist Posts: 1,631   +432

    Sounds legit. Flagships typically use the same SoC's anyway, so phone makers have to do something to make their product stand out.
     
    Last edited: Mar 31, 2014
  16. Its like if a car manufacturer published "This car will get 200 MPG" and then in the small print it says "you will never achieve this, this was done by going downgrade 2 degrees for 200 miles."

    Could you get this 200MPG? yes if you found a stretch of highway where it was downhill for 200 miles. Its it a realistic representative of what the car get 100% of the other times it was driven? No.

    False advertising 101.
     
  17. St1ckM4n

    St1ckM4n TS Evangelist Posts: 2,920   +627

    To be honest the whole phone benchmark thing is now obsolete. Just like frame pacing, and minimum fps before it, hugely affected GPU benchmarks recently, it's time to re-write how phone benchmarks are performed.

    I always found it odd that well-established sites that are experts in phone tech (esp. Android) post simple graphs for a single CPU test. Absolutely no mention of the CPU governor used, nor the clock speed attained.

    Phones need to be benched with standard out-of-the-box settings (with no cheating by OEMs), and also once again in full-performance mode.
     

Similar Topics

Add New Comment

You need to be a member to leave a comment. Join thousands of tech enthusiasts and participate.
TechSpot Account You may also...