Which smartphone/tablet owns the fastest processor?
Of course, with the kind of pacing we enjoy in this industry, the fastest processor today is often seen as slow tomorrow, but still, it's always interesting to know who's currently on top of the game.
There are two main factors in determining the speed of a processor - its clock rate and number of cores. Since we've come to the conclusion that increasing the core count brings more speed boost than simply upping the frequency a bit, we'll be looking for the CPU which has the most cores, and highest clock speed. Please keep in mind that the theoretical speed of a processor does not translate into overall speed of the device that uses it. There have been many examples where a smartphone or tablet with a slower processor performs way better than a higher-spec'd competitor. With that out of the way, let the CPU party begin!
Which smartphone/tablet owns the fastest processor? Fullscreen
More popular slideshows
iOS 7 release date and time are today (Sep 18), get ready to update!
18 Sep 2013, 04:00
Cyber Monday 2013 deals on phones, tablets and other electronics
01 Dec 2013, 10:55
Samsung Galaxy S5 rumor round-up: release date, price and specs
03 Dec 2013, 05:54
Android 4.4 KitKat update: release date, features and rumors
31 Oct 2013, 02:31
PhoneArena Awards 2013: Best cameraphones
03 Dec 2013, 11:01
Which smartphone/tablet owns the fastest processor?
1. #3 - Samsung Galaxy S III
2. #2 - HTC One X and LG Optimus 4X HD
3. #1 - ASUS Transformer Pad Infinity 700
1. Non_Sequitur (Posts: 1111; Member since: 16 Mar 2012)
Yeeeeeah! The Infinity wins! :3 Can't wait to get Jelly Bean!
2. 09wbd03516 (banned) (Posts: 140; Member since: 30 May 2012)
What about nexus7 PA ignore the nexus7?
5. Ray.S (Posts: 235; Member since: 19 Jul 2011)
Sorry, the Nexus 7 didn't qualify in the Top 3, as it "only" has a quad-core 1.3 GHz CPU.
3. loken (Posts: 462; Member since: 09 May 2012)
How about fastest dual core? because only dual cores in us
46. nnaatthhaannx2 (Posts: 817; Member since: 19 Oct 2011)
The S4 used on the One X usually out preforms the GS3 and other phones using it, but I still would say the GS3 is better than the ONEX
4. TheLibertine (Posts: 484; Member since: 15 Jan 2012)
Isn't the Exynos Raster than Tegra 3 despite slightly lower clock speed?
And please make such a comparison with graphics performance - with the iPhone and iPad dominating.
6. Ray.S (Posts: 235; Member since: 19 Jul 2011)
It might very well be, but those things are hard to measure, so in this competition here we're just looking at the number of cores and clock rates of the processors.
56. mohdr (Posts: 4; Member since: 29 Mar 2012)
ray ray ray come on !!! this is not fair and you know it.
what if i smoked your 7.2 L challenger by my 5.5 L E63 amg.
this does not mean your bigger engine is better.
31. Non_Sequitur (Posts: 1111; Member since: 16 Mar 2012)
The T3 smoked the A5X in most benchmarks. Besides, I'd rather play on a tablet with a widescreen display than a boxier tablet. The iPad isn't bad though for iOS fans, I'll give you that. I've played with my friend's iPad 2. It's pretty nice. :)
8. PhoneArenaUser (Posts: 4487; Member since: 05 Aug 2011)
But Samsung Galaxy S III, HTC One X and LG Optimus 4X HD use the same ARM's Cortex-A9 CPU architecture the performance difference becomes only because different SoC's can squeeze different amount of performance from the same CPU.
47. Mario1017 (Posts: 336; Member since: 04 Sep 2011)
The Infinity also has the same architecture
11. Ray.S (Posts: 235; Member since: 19 Jul 2011)
This venerable athlete was once on top of its game, but has been retired for quite some time now.
18. bobfreking55 (Posts: 866; Member since: 15 Jul 2011)
i think it could beat all those! it's even holds more hertz than the hammer of thor! lol
10. tedkord (Posts: 3409; Member since: 17 Jun 2009)
Actually, your analysis and reasoning are off. Simply increasing MHz or cores does not necessarily mean increased speed. There are to many factors to take into account. That's why in PC'S, a slower clocked quad core Intel Dandy Bridge is faster than a six core AMD at a higher frequency.
Internally, you can have the width and length of the pipeline in bits, cache size and speed, the speed of the bus, the number of calculations per clock cycle, etc...externally, the speed of the ram, how many background processes the system is also running, etc...
Now, if you're talking about processors all from the exact same family, then you can make some generalizations like in the article. If you have two Intel Sandy Bridges with the same number of cores, equal cache, etc...the more MHz, the faster.
I know this sounds nitpicky, but these sorts of misconceptions kept people from buying AMD Athlons back when they were spanking Pentiums with 20% higher clocks just because the Athlon said 1GHz, and the Pentium said 1.3GHz. (That plus the brand name. Lots of parallels to smartphones today in that analysis) People were paying more to get less. (More parallels)
57. roscuthiii (Posts: 1602; Member since: 18 Jul 2010)
"Goddamned incorrect"? Heh... I think even your autocorrect got autocorrected. ;-p
20. NexusKoolaid (Posts: 242; Member since: 24 Oct 2011)
To tedkord you listen. Know what he's talking about, he does.
30. Non_Sequitur (Posts: 1111; Member since: 16 Mar 2012)
Lol, I thought this was pretty funny.
23. doublehammer (Posts: 75; Member since: 07 Aug 2012)
thanks for saving me some finger stress.
Seriously... WHAT? I'm sorry Ray, but you need to do a lot of research on chips before doing such a baseless comparison.
The exynos found in the SGS3 blows the snot out of the T3 chip at equal frequencies, lower frequencies, and even higher frequencies.
Under your assumptions backed up by no facts, a (if there was such a thing) quad core Pentium4 chip running at 3ghz would blow away a core2duo running at 2ghz. Which is simply not true. There are so many factors to take into consideration its not funny.. which is why we lean so heavily on benchmarks.
And the benchmarks show quite the opposite. the exynos 4412 obliterates the T3 in all shapes and forms. The ONLY chip that has posted a higher bench than the 4412 is the unreleased S4 "Pro" chip which barely aced it out because of its new GPU. Of course that will be tossed by the exynos 5250, but that is a different story for a different day.
25. ObjectivismFTW (Posts: 211; Member since: 03 Jul 2012)
You have to remember, the Tegra 3 on the Infinity isn't the same one found on the One X, per say.
36. doublehammer (Posts: 75; Member since: 07 Aug 2012)
The T3 on the infinity is the full T3 that they use in higher end handsets/tablets, so it should be the exact same chip with a different clock. The T3 that they use in the Nexus 7 is slightly different to offset cost which comes with a slight performance decrease. I know they use the lower cost T3 chip on the lower end Asus Transformer models and the difference between the low and high end performance is substantial considering its relatively the same chip with minor modifications.
42. KingKurogiii (Posts: 5426; Member since: 23 Oct 2011)
the only real difference between the Tegra III that's in the Infinity and the Tegra III that's in the Prime is that the Infinity's SoC uses DDR3 RAM.
48. KingKurogiii (Posts: 5426; Member since: 23 Oct 2011)
why the thumbs down? i'm not wrong, you can look it up.
52. doublehammer (Posts: 75; Member since: 07 Aug 2012)
I shall up-thumb you in defense of your correct assertions!
54. ObjectivismFTW (Posts: 211; Member since: 03 Jul 2012)
People thumbed down my comment as well =0
44. ObjectivismFTW (Posts: 211; Member since: 03 Jul 2012)
Yes and No. Firstly, it is not the "exact same" chip .. The Tegra 3 on the Infinity is the T33 models, opposed to the one on the One X(lol) -- the T30 model. The main differences are more memory bandwith (DDR 667 2.7 GB/s to DDR3-1600 6.4 GBs per second) , a higher-clock speed and a higher voltage that yields for a higher boost clock.
24. ObjectivismFTW (Posts: 211; Member since: 03 Jul 2012)
I can see what your talking about tedkord. But you're forgetting that were talking about ARM CPUs here. You can't compare desktop CPUs to Mobile CPUs because so many variables differ, and so many inconsistencies arise. Remember, each CPU, first and foremost is constructed by ARM, and then tweaked to what ever liking the OEM desires. Different mobile factors like Cortexes put a sort of theoretical speed limit on how fast a CPU can be on that given cortex. Things like cache, DMIPS, Floating points, FSB speed, RAM speed and channels and pipelines only increase performance by the amount that a Cortex could yield. CPUs can only be that much faster than the other when they are burdened by the same theoretical speed limits. Even something like manufacturing process do not yield as many improvements on a Mobile CPU as opposed to a desktop CPU, because of transistor leakage. There are so many variables that decide the speed of a desktop CPU that simply aren't present, or do not make much a difference on a mobile CPU.
Ivy, Sandy and Bulldozer are seperated by so many different manufacturing specifics, it is easy to see why the latter CPU, well, sucks. Also factor in given motherboards, VRMs and all that. But CPUs like Exynos 4412 and Tegra 3(infinity) are based off the same Cortex(although Exynos' Bus width was increased from 64 to 128, mirroring Cortex A-15), and can only differ from each other speed-wise(excluding GPU) in so many ways. ARM is the key here, they have the hand in virtually all the CPUs, so they can set that theoretical speed limit. Any slight deviation from ARM's "rule book" can cause significant defects in performance because we are also dealing with many physical space confinements, although that is not the only reason why. We can talk about batteries, and how Desktop CPUs are given ample amounts of power and circuitry to complete their calculations, but that would only bloat things.
Only when we move to Cortex A-15 can we see the theoretical speed limit increase, and see Dual-Cores outclass Quad-Cores of the previous Cortex.
27. tedkord (Posts: 3409; Member since: 17 Jun 2009)
I will readily admit my "expertise" lands more toward desktop processors. However, due to the fact that ARM cpu's are designed by ARM, but modified and built by others, I think there are significant differences. Just within the Cortex A9 line you can have processors based on 45, 40 or 32nm processes, which will affect mainly power requirements and speed ramping.
The exynos 4210 and Tegra 3 are both quad A9, iirc, but clock for clock the exynos smokes it. Or at least from the benchmarks I've seen.
But again, I put quotes around the world expertise in the first sentence because I am certainly no expert in core architecture. Our anything else, really. In fact, if the world were populated with people as smart as me, we'd all be living in caves still. We might not have even figured out the cave post yet.
43. ObjectivismFTW (Posts: 211; Member since: 03 Jul 2012)
I will commend you, however, on the basis that you did know what you are talking about your posts, and you did well to explain why this article did nothing more than proliferate the idea that more cores or higher clock speed automatically leads to better overall performance. +1 on each
Phone Arena users deserve more than articles like these.
37. doublehammer (Posts: 75; Member since: 07 Aug 2012)
oh.. and how the theoretical limits shall increase. We are so close to A15 its making me impatient. :)
I wanna see just how close they come to "console" quality in real life applications vs their number charts.
29. Phullofphil (Posts: 792; Member since: 10 Feb 2009)
You are right on the fact that processors with lower clock speed are better than some with higher clock speeds but you should read the story better and understood it more because if you did you would have not wrot your entire opinion or atliest in that way. The point of the the story was to list the top by the way of spped as in frequenc speed. What you wrote is comon knowledg from all the benchmark scores that have already been published. Although this story is pointles in regards too what processor is better it seems that aloot of people are confused and its a really easy concept. Its like you have two garden hoses one is running at 40 psi and the other is at 50 psi. The won that has 50 psi is going to spray alot faster but the one that has 40 psi is runing through a hose that is five times as wide. The wider hose wil be shooting the water out slower but alot more water is coming out. But to tie it in to the storie. Its kinda asking what hose is going to be shooing water at the highest psi not the one with the the most. Its not an artical about the best processor but the one running the fastest.
For petes sake i rember having a intel pentium ship inside mydesktop runing at 3.4ghz ten or so years ago, but that chip probably cant out perform a snapdragon duel core runing at 1ghz but it ran faster thats why they went to duel core and so forth and brought the speed of the chip down to save on the flow of electricity going through the chip to make it more eficient. I hope this helps sombody even though it will probably confuse people more if they did not understand the origonal artical. Oh well
45. tedkord (Posts: 3409; Member since: 17 Jun 2009)
The reason cpu manufacturers started going to multi core configurations is that, due to heat dissipation, they were reaching physical limits on how small they could shrink the die process. Shrinking the process is generally needed to reach higher clock speeds, because it decreases the power needed, and as a result didn't run as hot.
People got used to the frenetic pace at which computers were advancing in speed and power. They weren't going to be able to continue delivering those increases through clock ramps, because the larger die cores would run way too hot. The original Pentium processor generally ran with only a small heatsink. Today's need much larger heatsinks that are actively cooled by a fan being through the fins.
So. they started producing processors with more cores at lower speeds. Which of course had zero benefit unless the operating systems had smp support and the programs became multithreaded.
53. Hunt3rj2 (Posts: 396; Member since: 11 Nov 2011)
The Exynos 4 Quad undoubtedly wins by sheer memory bandwidth, and GPU power.
12. toyds123 (Posts: 22; Member since: 10 Feb 2012)
"Congratulations to ASUS and the Infinity 700! Their achievement will soon be surpassed..."
WHAT THE HELL? this made me laugh!
14. neutralguy (Posts: 1152; Member since: 30 Apr 2012)
rumored version of one x with 4 1.7Ghz will soon be released.
15. Reverence (Posts: 197; Member since: 16 Jul 2012)
You guys are looking at the clock speed like 1.4GHz or 1.5GHz or 1.6GHz...........Come on guys..........Its not the clock speed or the GHz....its the benchmark that matter the most and S3's 1.4GHz quad core processor smokes Nvidia's 1.5GHz processor in HTC One X and the LG 4x HD......not sure about infinity
17. Nikolas.Oliver (Posts: 1260; Member since: 01 Jul 2012)
benchmark scores depends on many factors like the ROm Driver, and something like that, benchmark is not everything
40. baldilocks (Posts: 410; Member since: 14 Dec 2008)
You are correct. PhoneArena is 100% wrong on this front.
16. Nikolas.Oliver (Posts: 1260; Member since: 01 Jul 2012)
just because it has a quad core processor doesn't mean it faster than dual core phones out there, S4 Krait processor has already proved it,
19. Mitchel (Posts: 228; Member since: 25 May 2012)
This article seems a bit off and unprofessional. Clock rate & cores seriously? Lol.
22. TylerGrunter (Posts: 704; Member since: 16 Feb 2012)
I thoght exactly the same. I wonder if when the A15 cores come to play Ray is not going to look at them unless they are quad cores and with higher clock rates than the A9. LOL
33. Mehrzad (Posts: 2; Member since: 10 Aug 2012)
I think benchmark results will be totally different.
34. pkiran1996 (Posts: 160; Member since: 22 Oct 2011)
What? So root a Galaxy S3, OneX and Optimus 4X and overclock them to 1.6GHz and suddenly they're all as fast as each other? Not all about clock speed and cores. Anyone could have written this. Pointless article.
35. E34V8 (Posts: 37; Member since: 16 Dec 2011)
I cant believe that phonearena knows that little about mobile chips to compare theire productivity just by clock speed.
Ray, you know better. What is this article?
38. E34V8 (Posts: 37; Member since: 16 Dec 2011)
What about manufacturing processes (40/32/28) , cache per Core, instructions and other supported features?
39. baldilocks (Posts: 410; Member since: 14 Dec 2008)
How can you rate the GS3 in 3rd place when it easily, EASILY outpaces the 2nd place holders??
PhoneArena gets under my skin sometimes..
41. KingKurogiii (Posts: 5426; Member since: 23 Oct 2011)
whoa, okay, this is wayyyy off.
the Galaxy SIII and the Galaxy Note 10.1 are #1 with their Exynos 4412 SoCs
#2 either goes to the Transformer Pad Infinity with it's newer Tegra III SoC clocked @ 1.6GHz or the S4 equipped devices like the One X for example.
49. an4564 (Posts: 35; Member since: 17 Jul 2012)
this is stupid im pretty sure everyone knows by now that cores and clock speed no longer determine how good the cpu run
50. joey18 (Posts: 486; Member since: 20 Jul 2010)
The phone is slow wify and data so sg3 is not iphone killer is copycat killer
55. kimoz (Posts: 11; Member since: 03 Jul 2012)
sorry but gs3 is faster and much better than htc one x . the htc one x has many problems . the browser is horrible and when you press the home screen button it takes time to load the ui also it doesnt offer real multitasking . while gs3 is top notch in everything
58. roscuthiii (Posts: 1602; Member since: 18 Jul 2010)
As soon as I opened up the article and saw who the author was I started reaching for my chest waders.
Seriously Ray, how do they maintain your employ here? Is your dad Mr. PhoneArena or something? Does he know you're sneaking onto the computer?
59. achangavi (Posts: 21; Member since: 25 Jul 2012)
Seriously, if you want to test which phone/tablet has the fastest processor, root them, and install asp with full driver support. That way you'll be able to take fair benchmarks and make good comparisons because roms make a HUGE difference.