Intel Atom powered Lenovo IdeaPhone K900 benchmarks are astronomical
2GHz dual-core Intel Atom benchmark (left) and the XiaoMi Phone 2 vs the Lenovo IdeaPhone K900 (right)
1. XPERIA-KNIGHT (Posts: 2384; Member since: 08 Aug 2012)
NOOO WAY!! LOL that is crazy.....but also, does this mean the everyday usage will be smooth as butter in everything you throw at it? or is it "just" benchmark scores?
12. Retro-touch (Posts: 249; Member since: 24 Oct 2011)
Its just benchmark scores, it will just be as smooth as Nexus 4 (virtually no lag), beyond that Android itself will have to optimized further to eliminate lag completely and apps will then have to follow, Facebook still sucks till now even though its faster, for example
Benchmarks at this point are worthless, most of the high end phones even the octo core when it comes outs will perform similarly, with slight differences, seconds faster aren't really significant
25. hung2900 (Posts: 771; Member since: 02 Mar 2012)
About the score, the writer had a mistake "Snapdragon S4 Pro inside the HTC DROID DNA gets 14,000 points on average when benchmarked using AnTuTu, which is about as much as we get with the 1.6GHz Exynos 4412 inside the Samsung Galaxy Note II."
Actually these scores are Antutu 2.x.x, with Antutu 3.x.x, S4 Pro's score is about 19,500 - 21,000 and Note 2's score is about 17,500 - 19,000
61. Phullofphil (Posts: 800; Member since: 10 Feb 2009)
I just ran my Verizon note 2 once and got about 17500 wich is on par with other note 2's I took screen shots but I dont know how to post thise in my coments.
43. SuperAndroidEvo (Posts: 3766; Member since: 15 Apr 2011)
Yeah I al totally with you on that one. I have no idea who has a HTC Droid DNA that gets below 19,500 on Antutu.
I hit 20,000 to 21,000 everytime on my HTC Droid DNA & that is at the STOCK clock of 1.5GHz.
47. remixfa (Posts: 13902; Member since: 19 Dec 2008)
And this is where I call shenanigans.... not on you, on manufacturers and their fluffing of things for higher benchmarks. My Nexus4 which has a near identical hardware setup (krait quad pro @1500mhz, 2gig ram) and a lower resolution screen (720p vs 1080p) only gets about 16800 on Antutu.. I just ran it 3 times. All things being equal, I should have a higher score not a lower one since it takes a lot less power to push a 720p screen than it does a 1080p screen.
What do you get on Velamo? I'm running it right now.
51. darac (Posts: 2156; Member since: 17 Oct 2011)
Well, the Adreno 320 in Nexus 4 is proved to perform lower than the one in Optimus G due to some thermal throttling going on.
Maybe the whole SoC has the same issue.
It would explain the difference in Antutu compared to the Optimus.
My stock S3 hits over 16000, btw
59. remixfa (Posts: 13902; Member since: 19 Dec 2008)
you have to break it down by category to see the differences. Normally the integer or I/O is fluffed by the manufacturer to boost the score. This being a nexus means its unaltered by manufacturer fluffing and should be as true a score as you can get.
Are you using a US SGS3 with a dual core krait?
here's the breakdown for my Nexus4
CPU integer - 3791
Ram - 2127
cpu float 1822
database IO 395
SD write 111
SD read 200
CPU frequency - 1296 x 4 cores
The fun thing is that I set the cpu to run at 1500mhz for high and low, so its still being overridden by something. i keep getting results saying that the test was performed at everything from 1100 to 1296... but no higher.
Looking at the score ranks, the highest nexus at stock frequency's score is 20500. dont know if its a stock kernel though.
Anyways, compare your scores.
62. darac (Posts: 2156; Member since: 17 Oct 2011)
Here, from a test I did today.
It's the Exynos version
A fluffed integer?
I'm running the latest firmware with the multi window and stuff
15. 4.hour.uptime (Posts: 3; Member since: 11 Jan 2013)
but with 1 hour talk time
and 4 hour stand by time
30. abate (Posts: 43; Member since: 19 Dec 2012)
Look at this link.. Razr M (28nm S4 dual) vs Razr i (32nm Atom)
(Note all the hardware are same except the SoC(CPU).
01) Intel Atom entry offered a marginally faster response to most actions.
02) RAZR i booted in to Home Screen around 12 seconds ahead of the Snapdragon sibling.
03) On our battery rundown -- video playback, 50 percent brightness, WiFi and 3G on, social networks and email on sync -- Intel's iteration lasted nine hours, compared with eight hours for the RAZR M.
35. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
That's pretty much what I would expect, as the single thread performance of an Atom Z2460 should be a lot higher than the Dual krait, more taking into account the high CPU clock.
2. alterecho (Posts: 1059; Member since: 23 Feb 2012)
Whoa! Performance per core is astronomical indeed! Even if the Exynos Octa is more powerful, it will have a tough fight in single core tests.
16. 4.hour.uptime (Posts: 3; Member since: 11 Jan 2013)
there's a reason why there hasn't been any intel smartphone although intel already crying for 3 years.... performance over power requirement.
it's useless to have that fast CPU, if your phone only have 1 hour talk time and 4 hour stand by time.
23. linneti15 (Posts: 49; Member since: 26 Sep 2012)
intel has pretty much fixed the power consumption problem with the newest atoms... they are more power efficient than tegra 3 and exynos 5250 and just slightly better than snapdragon s4 but they are quicker!!! source --> anandtech.com/show/6536/arm-vs-x86-the-real-showdown. Sorry cant put the link in, i have done less than 30 comments :/ so just put: h ttp://ww w. without the spacs in from of it
36. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
A15 cores are more powerful (single cored) than the (current) Atom, so actually it won't be really any competition between the dual core Atom and and quad cores A15 like the Exynos Octa or Tegra 4.
The competitor for those will be "Bay Trail" (Atom quad core) and not "Clover Trail" (Atom Dual core), but it won't be out till the second half of this year.
3. noim1 (Posts: 297; Member since: 15 May 2012)
I always Knew Intel had the Raw Power ...they just have to utilize it for mobile devices !!!! Go Intel !! Bring in some Competition in the Processor Game...
8. MeoCao (unregistered)
if Intel somehow could use i7 for phones the benchmark would be through the roof.
The question for Intel is how about battery life?
13. eisenbricher (Posts: 970; Member since: 09 Aug 2012)
Agree. I am also waiting for AMD's specialization in low power graphics. I heard that they are working on graphics solution for smartphones!!
40. desijatt (Posts: 52; Member since: 28 Aug 2012)
I think Intel will get even more success in mobiles and tablets market than PC market....I'll wait for Bay Trail Processor in a android or windows tablet but still don't know how good will be the GPU with it....
4. mariosraptor (Posts: 108; Member since: 15 Mar 2012)
that is why it does not matter how many cores you have if they are not utilized prperly.
i didn't expect less from Intel, i expect more.
after so many years in PC CPU game they definitely know how to make chipsets.
now i want to see it's power efficiency and real life smoothness.
Good job Intel.
5. BadAssAbe (Posts: 434; Member since: 22 Apr 2011)
This sound too good to be true, If it was even remotely close to 25,000 Intel would have officially confirmed it while saying it could get better
6. Izzy_V (Posts: 216; Member since: 07 Jul 2012)
Kudos for giving the finger to all the skeptics that didn't think Intel would be serious competition from the Razr i's scores.
9. darac (Posts: 2156; Member since: 17 Oct 2011)
No doubt that this is a powerful chip:
..but I still think there could be something fishy about these results.
So far Atoms have been nowhere near those numbers in Antutu.
And Antutu itself has been proved as prone to faking results by manufacturers.
The GPU on this thing for example is only half the power of iPad 4' GPU.
Anandtech also compared the older dual core 2GHz Atom with Exynos dual, and the latter was way superior. Something just doesn't look right here.
20. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
Same as below... Old Z2460 is single core, that's how the Exynos dual was superior to it (actually depends for what task).
27. darac (Posts: 2156; Member since: 17 Oct 2011)
No, it was dual core, but clocked at 1.6, not 2ghz(my bad again)
29. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
I hope you were not the one that thumbed me down for correcting you, and if you are please have the decency to change it to positive.
The link you are posting is for a Atom N570 not Atom Z2460. So it's irrelevant.
49. darac (Posts: 2156; Member since: 17 Oct 2011)
Yeah I thumbed you down accidentally.. replying over the phone, hit the wrong field. Sorry about that.
Anyway, care to explain how it's irrelevant?
It's the same basic architecture I believe:
"Since 2008, Atom hasn't had any core architecture changes. Sure Intel integrated the GPU and memory controller, however the CPU still communicates with both of them over an aging FSB. The CPU itself remains mostly unchanged from what we first saw in 2008. Even Intel's 32nm Atom due out by the end of this year doesn't change its architecture, this is the same dual-issue in-order core that we've been covering since day 1. The 32nm version just runs a bit quicker and is paired with a beefier GPU. "
52. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
I find it irrelevant cause one of the differences is precisely the clock speed, that differs between the Atom Chipsets. So that the N570 is single core or runs at 1.6 MHz doesn´t bring anything to the discussion about if the Z2460 is single or dual core or what is it´s clock speed.
That´s all I meant with it.
10. RobotMan (Posts: 124; Member since: 13 Apr 2012)
Have to remember that some app cannot run using this cpu .
11. Techsplash (Posts: 41; Member since: 24 Oct 2012)
That is really insane.It is almost double of the performance of the note 2 and droid dna.
14. darac (Posts: 2156; Member since: 17 Oct 2011)
Again, I don't know.. take a look at these 2Ghz dual core Atom Z2460 benchmarks:
Note that it got destroyed even by dual core 1.5GHz Krait in the Razr M.
It's as clear comparison as you could get.
Now what so magically happened in this chip to warrant such performance jump??
21. abate (Posts: 43; Member since: 19 Dec 2012)
Are you an ignorant?
Z2460 is a SINGLE core atom.
get your facts correct before post anything.
17. darac (Posts: 2156; Member since: 17 Oct 2011)
..The only difference between Z2460 and this Z2580 is in the GPU.
So yeah, the results don't make sense at all
18. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
Darac, don't post nosense please.
Z2460 is a SINGLE CORE chip, while Z2580 is a DUAL CORE.
Atom cores are somewhere in the middle between Krait and A15 clocked at the same speed. So the results make all of the sense of the world.
It's normal that Atom at 2 GHz get insane results, but at that clock speed I would worry about the battery life.
Also the GPUs intel use are always underperformers, Z2580 not being an exception.
26. abate (Posts: 43; Member since: 19 Dec 2012)
"Also the GPUs intel use are always underperformers, Z2580 not being an exception"
I don't know where you get than. But according to this link
Z2580's GPU is better than iPhone 5(PowerVR 543 MP3), Nexus 4(Adreno 320) and Nexus 10(Mali 604).
It matches the graphical power of ipad 3. So only Tegra 4/Mali 658/ Adreno 330 can surpass the power of that GPU.
But according to XDA-Developers. Tegra 4 will only match Mali 604. So z2580's PowerVR 544 MP2 @ 533 MHz will be a top performing GPU for quite a while, until Mali 658 or Adreno 330 make it to any smartphones.
28. darac (Posts: 2156; Member since: 17 Oct 2011)
If it matches the iPad 3' GPU, then it's only half the power of the SGX 554MP4 found in the iPad 4.
And it's NOT more powerful than Mali 604..or even Adreno 320
31. abate (Posts: 43; Member since: 19 Dec 2012)
Go to this link and see it yourself
iPad 3's GPU destroys both Mali 604 found in Nexus 10 and Adreno 320 found in Nexus 4.
iPad 3's GPU A5X is more than TWICE FAST of Adreno 320 and more than 50% fast of Mali 604.
I know you are a blind fanboy to accept the facts. That's why I mentioned the source links in the previous comment, yet to come again and say Adreno 320 and Mali 604 beats A5X like a blindfolded person.
Seriously you have to come back to the real world from whatever delusional world you are in right now.
33. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
He may not be the blind one, for GPUs the test that counts more is the GLBenchmark 2.5 Egypt HD. In that test BOTH the Mali 604 and the Adreno 320 fare better, showing they are more powerful than the GPU in the iPad 3 (PowerVR 543 MP4 @ 250 MHz). That's in your own post.
If you don't get something as basic as that, I guess it's not worth to discuss further.
Just to make a stronger case, check the glbenchmark website yourself:
38. abate (Posts: 43; Member since: 19 Dec 2012)
Man, it's Egypt HD. That benchmark totally depend on screen resolution. We all know that.
to confirm you,
just compare the iPhone 5 score vs iPad 3 in Egypt HD.
Where iPhone 5 will surely beat ipad 3.
But you know that iPad 3 has a better GPU (SGX 543 MP4 38.4 GFLOPS @300MHz) than iPhone 5 (SGX 543 MP3 28.8 GPLOPS @300MHz)
Moreover all the rest results on that page is all about GPU. CPU benchmarks are on the other page.
GPU => anandtech.com/show/6472/ipad-4-late-2012-review/4
CPU => anandtech.com/show/6472/ipad-4-late-2012-review/3
42. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
You just demonstrated how little you know about GPUs...
The iPhone 5 beats the IPad 3 int GPU any time for one reason: higher clocking. The SGX 543 MP3 in the iphone 5 is clocked at 266 MHz, while the SGX 543 MP4 in the iPad is clocked at 200 MHz (both estimated by Anand, not me), therefore the advantage of the iphone 5.
So yes, the graph for GLBenchmark 2.5 Egypt HD offscreen is correct, and yes, the Mali 604 and Adreno 320 are more powerful than SGX 543 MP4 @200MHz.
Please check the facts and learn how to interpret graphs and benchmarkt tests before posting again. It will save me some time and I'll be grateful for that.
53. abate (Posts: 43; Member since: 19 Dec 2012)
I could understand how desperate you are to prove you are right and seal you are a fanboy.
It clearly shows how little you know about GPUs..
GPU power is not only about higher clocks. It's more than that. It's about the architecture, how many GPU cores, how many supporting CPU cores and of cource clock speeds.
You have to understand iPad 3's SGX 543 MP4 is a quad core GPU, while iPhone 5's SGX 543 MP3 is a Triple core GPU.
You have to understand that,
Quad core SGX 543 MP4 running @ 200 Mhz is better than Triple core SGX 543 MP3 running @ 266 MHz.
iPhone 5 SGX 543 MP3 @ 266 MHz = 28.8 GFLOPS @ 300 MHz (if calculated)
iPad 3 SGX 543 MP4 @ 200 MHz = 38.4 GFLOPS @ 300 MHz (if calculated)
That's why iPad 3's GPU is powerful than iPhone 5.
You first go and study about hardware architectures before lecture others.
If you are still in doubt, go and ask some experts.
50. darac (Posts: 2156; Member since: 17 Oct 2011)
I understand why you got the impression of me being a fanboy. But anyway, I'm not..
I was referring to the 1080p offscreen test in Egypt 2.5HD, which is really the only fair comparison of raw performance, taking out the screen resolutions difference.
1. Mali 604 -33fps
2. Adreno 320 -29fps
3. SGX543MP3(iPhone 5) -27fps
4. SGX543MP4(iPad 3) -25ps
54. darac (Posts: 2156; Member since: 17 Oct 2011)
56. abate (Posts: 43; Member since: 19 Dec 2012)
Fool, look at the galaxy S3's score, HTC One X's Score and Crappy Razr i's score at your graph.
All three are same. So you accept that?
SGX 540 matches Mali 400 and Tegra 3?
More than 4 years old GPU matches the performance of 2011's GPU?
You ignorant, that benchmark is too screen resolution dependent.
55. abate (Posts: 43; Member since: 19 Dec 2012)
Man I told this already and told you again. But you fanboys won't accept it. I know that.
Egypt 2.5 HD scores are totally depend on the screen resolution.
If you are not sure, just look at Razr i 's score.
We all know that phone has one of the most out dated and crappy GPU at the moment. That is SGX 540.
We all know that will not match the (HTC One X) tegra 3 or (Galaxy S3) Mali 400 in performance.
But look at the Egypt 2.5 HD score.
It performs same as Galaxy S3 and HTC One X.
You people are delusional. More than that everyone trying to be experts, even though they are not in reality.
57. darac (Posts: 2156; Member since: 17 Oct 2011)
That's exactly why I posted BOTH the on-screen and off screen Egypt HD tests!!
In the on screen test you can see how much role the resolution plays.
In the off screen test you get a clear raw performance picture.
And you can clearly see that SGX540 is the weakest of all.
63. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
Thanks Darac, you saved me one post.
But I would no bother any more, the guy seems not to be smart enogh to understand the difference berween the onscreen and the offscreen tests. So trying to explain it´s a lost battle. :(
64. Berzerk000 (Posts: 3883; Member since: 26 Jun 2011)
He was referring to the OFFSCREEN tests, in which the results are completely *independent* from resolutions. And in this test which resolution is a constant (constant meaning invariable), the Mali-T604 and Adreno 320 (Mali being on top of Adreno) gets put above everything except the PowerVR SGX554MP4 of the iPad 4, which is expected.
Offscreen tests are the only tests in which resolution is completely IRRELEVANT in order to see the raw performance of devices, and it is in those tests that you can see that the Adreno 320 and Mali-T604 can ONLY be outclassed by the PowerVR SGX554MP4. The iPad 3 and iPhone 5 GPUs get left behind by the both of them.
Get your head straight.
32. TylerGrunter (Posts: 866; Member since: 16 Feb 2012)
Sorry, I must admit was thinking of Atom Z2760, not Z2580 as they didn't mention which one was used.
Still when the Atom Z2580 goes to market (intel anounces chips way in advance) is going to face off the iPad 4, Mali-624, Tegra 4 and Adreno 330, while it's perfomance is on the range of the iPad 3, Mali 604 and Adreno 320.
19. _Bone_ (Posts: 2104; Member since: 29 Oct 2012)
That's probably the new AnTuTu under which devices score 50% higher.
22. RapidCat (Posts: 351; Member since: 12 Jun 2012)
all about battery life
41. remixfa (Posts: 13902; Member since: 19 Dec 2008)
looking at the scores, their biggest leap comes in RAM and SD access. Gotta figure out what they are using on the internal set up.
That said.. its still a blah. Until they can prove 100% compatibility with all android apps and android OS out of the gate, I'm not interested. Otherwise, its just another fork in the android road we dont need.
44. Jonathan41 (Posts: 526; Member since: 22 Mar 2012)
Well damn. The high-end Android market might have a new candidate.
45. Shubham412302 (Posts: 298; Member since: 09 Nov 2011)
The main benebit of intels processors hyperthreading is that an extra core is only formed when its needes.
Uptill that single core can run at 1ghz