• Samsung found to be continuing to cook the benchmarking books with the Note 3 by Ars Technica
    18 replies, posted
[url]http://arstechnica.com/gadgets/2013/10/galaxy-note-3s-benchmarking-adjustments-inflate-scores-by-up-to-20/[/url] [quote=Ars Technica]We noticed an odd thing while [URL="http://arstechnica.com/gadgets/2013/10/review-the-galaxy-note-3-is-big-and-it-pulls-some-benchmark-shenanigans"]testing the Samsung Galaxy Note 3[/URL]: it scores really, really well in benchmark tests—puzzlingly well, in fact. A quick comparison of its scores to the similarly specced LG G2 makes it clear that something fishy is going on, because Samsung's 2.3GHz Snapdragon 800 blows the doors off LG's 2.3GHz Snapdragon 800. What makes one Snapdragon so different from the other? After a good bit of sleuthing, we can confidently say that Samsung appears to be artificially boosting the US Note 3's benchmark scores with a special, high-power CPU mode that kicks in when the device runs a large number of popular benchmarking apps. Samsung did something similar with the international Galaxy S 4's GPU, but this is the first time we've seen the boost on a US device. We also found a way to disable this special CPU mode, so for the first time we can see just how much Samsung's benchmark optimizations affect benchmark scores.[/quote]
If it's running a special high power mode wouldn't that give it a significantly lower battery life? Isn't that a benchmark right there?
That's a dick move.
[QUOTE=sloppy_joes;42372031]If it's running a special high power mode wouldn't that give it a significantly lower battery life? Isn't that a benchmark right there?[/QUOTE] "kicks in when the device runs a large number of popular benchmarking apps" [editline]1st October 2013[/editline] aka not during regular use
At the end of the day, synthetic benchmarks are pretty silly.
[QUOTE=danharibo;42372325]At the end of the day, synthetic benchmarks are pretty silly.[/QUOTE] quantifying performance to something understandable to the populace is pretty silly? Well, Futuremark just went bankrupt then.
[QUOTE=Xmeagol;42372476]quantifying performance to something understandable to the populace is pretty silly? Well, Futuremark just went bankrupt then.[/QUOTE] Silly != unprofitable
I get that you want to show your phone is the strongest of them all but this is kinda scummy.
[QUOTE=Sir Whoopsalot;42372579]I get that you want to show your phone is the strongest of them all but this is kinda scummy.[/QUOTE] Samsung has scummy business practices In other news, dog bites man, next phone release has bigger screen and more GHz, sun rises More at 11!
It's not outright lying since it's simply using 100% of what the processor is rated for (not overclocking or anything) but if this mode is limited to literally only benchmarking apps it's dishonest as hell. Would also go a long way towards explaining how Samsung can so often dominate benchmarks but still have phones that lag to hell and back
I wonder how they'll try to explain this one. That is if they comment on it at all.
[QUOTE=Kybalt;42372301]"kicks in when the device runs a large number of popular benchmarking apps" [editline]1st October 2013[/editline] aka not during regular use[/QUOTE] ... Then you just run the benchmark apps and you can complain about how long the battery lasts.
[QUOTE=danharibo;42372325]At the end of the day, synthetic benchmarks are pretty silly.[/QUOTE] Synthetic benchmarks are perfectly fine when they're fair, and not being specifically optimized for, as is not the case here. They're valuable if you're getting actual numbers relevant to real world performance.
They're giving a synthetic environment to a synthetic benchmark. These benchmarks are supposed to measure the maximum performance of a device, not what the device is like in day-to-day use, thus there's no point in waiting for a governor to catch up when a benchmark is running. I'm fairly sure devices from other manufacturers (I heard HTC One being referenced elsewhere) also do this.
[QUOTE=nikomo;42373506]They're giving a synthetic environment to a synthetic benchmark. These benchmarks are supposed to measure the maximum performance of a device, not what the device is like in day-to-day use, thus there's no point in waiting for a governor to catch up when a benchmark is running. I'm fairly sure devices from other manufacturers (I heard HTC One being referenced elsewhere) also do this.[/QUOTE] No, they don't. HTC does not do this, it's just Samsung. Samsung's settings allow it to use clock speeds that can NEVER be used in any other scenario. It's not like they're just forcing it to 100%, no other applications except Samsung's built in ones can get the clocks to ramp up to these speeds. Just to quantify it, on the S4 the GPU gets ramped up 50MHz faster in benchmarking apps than it can ever be allowed to run in an intensive 3D game. Anyone justifying this is either misinformed or is very lenient with what they consider to be okay business practices. This is NOT a representation of what the phone is capable of in a scenario where the user is actually using it. It's scummy, and Samsung being shameless about it after already being caught doing it with the S4 just amplifies that. At least we have HTC for now.
I still don't get why it's so important to have super powerful phones, it's not like Angry Birds requires a very powerful system.
Solution to the slow gallery: rename it to Geekbench :v:
[QUOTE=Zeke129;42372697]It's not outright lying since it's simply using 100% of what the processor is rated for (not overclocking or anything) but if this mode is limited to literally only benchmarking apps it's dishonest as hell. Would also go a long way towards explaining how Samsung can so often dominate benchmarks but still have phones that lag to hell and back[/QUOTE] I always assumed Samsung Lag had to do with how much they bum fuck android with all their shitty stock apps and stupid themes
[QUOTE=JohnnyOnFlame;42374424]Solution to the slow gallery: rename it to Geekbench :v:[/QUOTE] Can someone try this? Stupid midterms..
Sorry, you need to Log In to post a reply to this thread.