Samsung Galaxy S II Skyrocket benchmark tests
0. phoneArena 04 Nov 2011, 15:59 posted on
Before running any benchmark tests, our mind tells us that it’ll probably be slightly below the level of its Exynos equipped counterpart. Rather, we’re pleased to find it performing on the same level, since we’ve able to get scores between 2,787 and 3,636...
This is a discussion for a news. To read the whole news, click here
1. remixfa (Posts: 14255; Member since: 19 Dec 2008)
Am I the ONLY one that can make the easy assumption that Samsung has found a way to fidget with score numbers? NO HTC phone running the EXACT same hardware configuration comes anywhere close?
Is it magic?
No.. its fidgeting with benchmark tests somehow... Otherwise, you would see the HTC Amaze get the same scores, and the Vivid and sensation get really close scores. Thats not happening, its ONLY samsung getting these kinda scores with this setup. Its being FAKED.
And this is coming from the guy that normally champions samsung phones.
2. snowgator (Posts: 3515; Member since: 19 Jan 2011)
As in the phones are like programmed to give a false reading? If this comes from anyone else, I would laugh at that. Do you really think Samsung would risk that in the face of all the legal challenges it is facing?
4. remixfa (Posts: 14255; Member since: 19 Dec 2008)
there wouldnt be any legal challenges... its not a big enough issue.
I know from past experience that the I/O score is easy to manipulate. The only way to really tell is to use the paid version of quadrant as it will break down each category and give you its individual score as well as the total score. Take that individual score and compare it to the Amaze and other Dual Core Snapdragon S3 phones with similar set ups. Id bet money that the I/O score has been manipulated and is higher on the Samsung hercules/skyrocket.
That is EXACTLY what samsung did. Here is the Antutu benchmarks from the Amaze which runs the exact same 1.5ghz and 1gig ram set up as the Hercules/skyrocket. That gray part of the benchmark IS THE I/O score (Antutu calls it CPU Intiger). Compare that screen shot to this phone's and you easily see the huge increase in the I/O marks.
e.php?m=Articles.Images&f=name &id=47294&name=1.jpg&caption=& title=Image+from+%22HTC+Amaze+ 4G+benchmark+tests%22&kw=&popu p=1
I love some samsung hardware, but I will call out dirty pool wherever I see it.
6. remixfa (Posts: 14255; Member since: 19 Dec 2008)
Best part is, if your breaking it down by score, the I/O score is the least useful part of the chip at this point. The chips are too powerful, they all score well in I/O.
If you want to know what chip is best look at the 2D/3D power, floating point, ram, database access, and SD READ... Basically everything but the I/0 (integer) and the SD Write.
I wish the OMAP processor was in there for a more accurate portrayal of the big 4 chips.. but oh well.. for reference it falls between the tegra2 and the exynos. the snapdragon unaltered falls behind the pack.
11. remixfa (Posts: 14255; Member since: 19 Dec 2008)
for reference here is the Sensation
28. vader11 (Posts: 5; Member since: 04 Nov 2011)
You are right, I have the paid version of qradrant, and Tmo-GS2 actually scores lower than Sensation XE in every categories except the I/O. Its I/O is way ahead of the the XE and the total score of the GS2 was ~3500s, and the XE was ~2600.
29. remixfa (Posts: 14255; Member since: 19 Dec 2008)
thank you for validating. Any chance of some screen captures? :) :) :) :)
33. Snap123 (Posts: 33; Member since: 04 Nov 2011)
Ok from comparing the score between the HTC Amaze 4g and the Samsung Skyrocket i've come to this conclusion. Correct me if im wrong because im no phone whiz.
HTC Amaze 4g:
CPU integer: 1238
CPU float-point: 423
CPU integer: 1952
CPU float-point: 772
It seems to me that the more ram is available for use the higher it can score for cpu integer and cpu float-point.
As we know HTC sense is a major ram hog and Samsung TouchWiz is pretty light on ram usuage.
Anyone with with an HTC Amaze will find usally you have only 300-400 ram available for use. With Samsung Galaxy S2 phone you can get about 700-800 available ram.
So maybe this is why Samsung phones with the same Snapdragon s3 chip while always score higher than HTC phones
35. remixfa (Posts: 14255; Member since: 19 Dec 2008)
for over all system function sure, but the CPU I/O(intiger) is a baseline test. It isnt affected by the software. Its a test that is basically measuring how fast the chip can calculate answers... say to PI for example. As to why this test in particular is the one that is so easy to manipulate, that is beyond even my pay grade, but it is the test that people use to cheat and buffer over all scores on. Ram may or may not make a difference, but that is why I chose to compare it to the Amaze.. they run with the same amount of total ram and same chip configuration. The sensation however gets similar scores to the amaze and it has a good chunk LESS ram than the amaze does as well as the lower 1.2ghz clock.
Like I said in a previous post, this benchmark for I/O is nearly useless now exept for baseline bragging rights in super computers.. lol. Computer chips are just so fast now that we cant throw enough at em to make a real world difference for 99% of us.
If your not into tech, this wiki blurb might make your head spin, but its a good discription of I/O and floating point scores.
53. vader11 (Posts: 5; Member since: 04 Nov 2011)
I dont know how to post pics but here's the results:
Also note that the scores of the GS2 are very unstable with ~2500-3900, while the XE are quite stable with ~2500-2800
41. jinwons (Posts: 95; Member since: 24 Nov 2010)
what's wrong with that? I think it's probably the Sense UI that bogs down performance of HTC phones. Sense is notorious resource hog. But it seems Sammy did pretty good job on the new TW 4.0 to get those scores with better file I/O and hardware acceleration.
57. Forsaken77 (Posts: 552; Member since: 09 Jun 2011)
C'mon man... are you serious? You still cannot accept that the phone and processor are much better than you thought? You're provided with PROOF and you're still making excuses. "Samsung faked it." Ok, so they would fake this benchmark to look on par with the Exynos but not fake the Exynos to look even more powerful? ACCEPT IT... IT'S A GOOD PHONE! As for your HTC comparisson, maybe it's not just the processor. Maybe it's the whole package, the way all the components interact with the processor. Maybe Samsung knows how to squeeze more juice out of it then HTC. Maybe Qualcomm stepped up their game. Whatever the reason.... you bash the phone/processor and even when you're provided with proof you cry conspiracy. Think about what you're saying and how it sounds.
59. remixfa (Posts: 14255; Member since: 19 Dec 2008)
concidering i provided an educated opinion, proof to back up my claims and worries, and people that know about chips agree with me that its probably messed with... id say I'd sound like an individual that has done his homework and has a valid concern.... not someone who is screaming it must be true on a bunch of "maybes".
63. KiretoX (Posts: 1; Member since: 05 Nov 2011)
Are you entirely sure that the IO result comes ONLY from the CPU? Well, what really differs in OCZ Vertex 3 and V3 Max IOPS is the MEMORY (being synchronous vs. asynchronous), not the CPU. And likely firmware optimisation. And the firmware between HDDs for Servers and others. So more and better ram - more IPS. It's very likely that Samsung knows better than HTC how to use memory...
In truth I've thinking about the same issue - that Samsung could have manipulated the score, but if that was the case someone would have been done enough research @ XDA to make some noise. After all most of the androids is open source, aside of some hw drivers.
67. remixfa (Posts: 14255; Member since: 19 Dec 2008)
The I/O test itself is a CPU test that (and i explained this in another post) basically measures how fast it can come up with answers to an equation... like say, the first 500 digits to PI. Since its basically a FLOP test (how fast the internal "on/ off switches" can move) Im not sure if its affected by RAM or not. the I/O is a measure of MFLOPS on the cell phone scale.. Super computers go to PetaFLOPS and beyond, just as a reference.
73. kabel (Posts: 58; Member since: 06 Nov 2011)
I think you are mixing up the the things. First thing "I/O" is related to hdd, ssd, flash memory ... So CPU I/O is not the same with CPU Integer performance. CPU's have separate integer and floating point units, and with that how in the he** CPU Integer performance can be expressed with FLOPS (floating-point operations per second). Read-up little bit, because you are confusing the people. "I/O" != CPU INTEGER != CPU FLOPS
74. remixfa (Posts: 14255; Member since: 19 Dec 2008)
I/O simply means input/output. input the command, output the answer. As in how fast the CPU can take the equation and spit out the answer. That is all.
it is related to anything that can take in info and spit it out, which most definitely includes a CPU.. on a PC or smartphone. The integer test IS an I/O test. throw integers at the CPU, and measure how fast it spits out the answer.
80. kabel (Posts: 58; Member since: 06 Nov 2011)
"I/O" != CPU INTEGER != CPU FLOPS
CPU Integer IS NOT measured in FLOPS as you stated above ... I'm out ...
72. Jimstar (Posts: 259; Member since: 24 Oct 2011)
I've suspected it for weeks. Samsung's the only manufacturer that fiddles with the file system, cept it was giving low scores on the galaxy S devices, they must've switched tactics this time. And they do the same thing with the Tegra II in the Galaxy R I imagine.
3. ArmageddonX (Posts: 96; Member since: 11 May 2011)
The benchmarking community calls the SGS2 an "anomaly". Nobody knows why or how it does this & there is suspicion that somehow the numbers are simply not true. Nobody knows "for sure" what is going on but it certainly doesn't make sense; even to "experts" in bench-marking it's strange.
7. ZEUS.the.thunder.god (unregistered)
dont mean to disrespect you sir but samsung is way better than anybody on the market right now.. lets accept it. nothing is being faked. lets see what they achieve with sgs III
9. Synack (Posts: 676; Member since: 05 Jul 2011)
I Agree, nothing compares. I find myself comparing Samsung to everything else all the time, and Samsung always wins.
12. remixfa (Posts: 14255; Member since: 19 Dec 2008)
You are not going to find a bigger champion of Samsung's tech on this site than me. I would also be part of that elite group of technophiles that has in depth knowledge of the chips in question. The other 2 guys would be King Korugi (AKA Thump3r) and Protoloz. This chipset is not capable of this performance.
However I refuse to close my eyes to the competition or to the company who's products I buy. Blind fanaticism only gets you slapped by the direction that you close your eyes in.
A spade is a spade is a spade.. no matter who carries the spade. And in this particular stance, Samsung is playing dirty pool. The proof is right in front of you for all to see.
16. buggerrer (Posts: 306; Member since: 21 Sep 2011)
have you heard of resolution? and how they affect benchmarks ?
or the truth that galaxy software is smoother than htc ?
sorry but your post sound retarded! why a big company would risk their reputation for a benchmark test!
you have to prove it here or in xda-developers or stfu :)
btw there is a way to cheat in benchmarks you can look for it in xda-developers ..
18. remixfa (Posts: 14255; Member since: 19 Dec 2008)
if u actually read the comments, ive already answered EVERY question you bring up.
32. buggerrer (Posts: 306; Member since: 21 Sep 2011)
so does that mean there is also something strange about nexus benchmark ... since its the closest score to sgs2 ?
36. remixfa (Posts: 14255; Member since: 19 Dec 2008)
The nexus uses the OMAP chip which is the 2nd strongest chip next to the exynos. The nexus is also running ICS which is the first android OS optimized to actually USE dual core chips. Until now, android has not taken one bit of advantage of all that processor power. its only been used in games and such like that. As the other phones make their transition to ICS like the nexus has, all their scores will go UP because of this. Thats when we will REALLY see the difference in the chips even more so than we see now... because they will be fully optimized. They are barely being used at the moment.
62. buggerrer (Posts: 306; Member since: 21 Sep 2011)
where do you read that ICS supports dual core ? I didn't hear them saying anything about it at the conference.
and SGSII is optimized by Samsung software engineers to use the dual cores .. maybe not as it should be .. but the dual core in GSII are in use
68. remixfa (Posts: 14255; Member since: 19 Dec 2008)
Dual Core optimization was the first thing we ever learned about ICS. Its built off of Honeycomb, which itself is dual core optimized (just not done well..lol). At the conference they hit many of the OS highlights and didnt get into a lot of the geekery. There is an article from Michael from a week or 2 ago talking about all the ICS stuff that they didnt mention during the conference.
SGS2 can not be optimized by samsung to use dual core in the android OS unless the andriod OS allows it by having system drivers to run it. Which as of right now, no android OS other than honeycomb and ICS has that feature. Its optimized for 3D performance, battery, and a billion other things, but they can not force Android itself to be dual core optimized until Google puts that feature into android... which they are with ICS.
70. buggerrer (Posts: 306; Member since: 21 Sep 2011)
i'll stop here .. just wanna anything prove that ICS support multi-cores .. something official or any one report that ICS SDK or kernel do support cores
71. remixfa (Posts: 14255; Member since: 19 Dec 2008)
at work right now so internet is limited, but you can just google "ICS dual core optimized" and it will probably pull right up.
23. cheetah2k (Posts: 1373; Member since: 16 Jan 2011)
Man, there's always one crying foul in the crowd..
Its called being able to optimise the hardware for the software, and vice versa. Samsung's collaboration with Google for a few years now producing Nexus phones has helped them to a large degree. Either way, Samsung know how to produce quick phones.
Comparing Samsung to HTC, where HTC don't give a fcuk about optimising anything is just rediculious!
24. remixfa (Posts: 14255; Member since: 19 Dec 2008)
lololol. you know, if it were just that easy and simple.
I'll throw away my entire knowledge base, and just say "because its samsung, it must be so".
Even though i am a HUGE champion of samsung tech, I can look with open eyes and compare known reality to what's presented infront of me, and make a comparison.
There is always a dissenter, and in this case, it happens to be one of the biggest geeks on the site that is also one of the biggest fans of samsung tech.. rallying AGAINST a samsung product. Maybe u should take that with a grain of salt and think about it for a second?
27. remixfa (Posts: 14255; Member since: 19 Dec 2008)
Just as a quick thought.
closing your eyes to samsung and just pretending everything they do is beautiful flowers and flawless makes you just as bad as taco and gallito and their blind fantacism to Apple.
You dont want to be like them do you??
30. gemmabba (Posts: 51; Member since: 09 May 2011)
i agree with you remixfa..there's something fishy about Samsung benchmark score...i dont know what it is, but i hope someday PA brought this to attention and prove that the score is legit (or not)...so no more false accusation to Samsung for manipulating the score
34. KingKurogiii (Posts: 5704; Member since: 23 Oct 2011)
you know i'm like right there with you on that one guy. xD
i had my suspicions since like the T-Mobile Galaxy S II.
38. remixfa (Posts: 14255; Member since: 19 Dec 2008)
I was wondering when you would chime in. :)
Bench and spec comparison proves it. Samsung is playing with the numbers. Oh well.. just proves one thing.. never buy a snapdragon.. lol :)
39. KingKurogiii (Posts: 5704; Member since: 23 Oct 2011)
yeah, i put that out of my mind long, long ago. xD
i'd still get any of the S3 using Galaxy S IIs over nearly any HTC...MAYBE not the Rezound. i kind of like it.
40. remixfa (Posts: 14255; Member since: 19 Dec 2008)
I think ill just hold on to my Vibrant until something more appealing comes along. Dont want an incrimental upgrade.. want a big fat major upgrade..
Maybe that Tegra3. Its soooo close to a phone launch.. lol. Then again, by the time those launch in a phone, the Exynos quadcores will probably be out, if not at least the 4212 dual core which will be semi-close to performance.
ooooooooh the choices!!! lol.
45. KingKurogiii (Posts: 5704; Member since: 23 Oct 2011)
yeah, i might have to get creative and get myself another phone. xD
44. Dr.Phil (Posts: 1064; Member since: 14 Feb 2011)
So then, using that logic, wouldn't Samsung then be fidgeting the numbers when it comes to the Exynos? I mean think about it, people have hyped up that the Exynos is the best processor out there. Or is it just what Samsung wants us to believe? (twilight zone music starts playing).
46. GoodFella (Posts: 112; Member since: 18 Oct 2011)
My Samsung Galaxy S II is overclocked to 1.6 Ghz and it just ran a 5039 Quadrant Benchmark. The Exynos Processor puts Snapdragons to absolute shame. Absolutely amazing!
47. ledbetterp3 (Posts: 467; Member since: 31 Aug 2011)
No, dude, I totally figured that since the sgsII's benchmarks came out, I just never wanted to argue with anyone lol...
66. shayan (Posts: 159; Member since: 09 Sep 2010)
dont be silly. samsung can not do that! I am not saying they wont do it because its moraly wrong, Im saying they can NOT do that.
the benchmark is higher because same SoC dosent mean same performance. even the same exact hardware doesent mean same performance. and benchmark scores depend on many things.
a dual chanel ram does make a huge diffrence in cpu performance.
in computers, even a bump in gpu clock can make cpu more functional.
8. downphoenix (Posts: 3050; Member since: 19 Jun 2010)
Or it could be that HTC's sense is crazy bloatware and is a big drain on performance. I vote that.
17. remixfa (Posts: 14255; Member since: 19 Dec 2008)
could be true, exept XDA folks have been removing sense from the sensation and the resulting performance benchmarks didnt change all that much. Sense (or any other UI) and all of its goodies are marginalized during a benchmark test, so most of it has no bearing on the final result.
19. Synack (Posts: 676; Member since: 05 Jul 2011)
They might've removed the visual part of Sense but Sense is still the underlying UI and skin running it all. Just like when people run AOSP in place of Touchwiz. It just makes the looks of it like AOSP but it's still Touchwiz in the end.
21. remixfa (Posts: 14255; Member since: 19 Dec 2008)
no, my point was they have roms that run without sense completely.. not just visually removed. Its not impossible. There are many of them, including the world famous Cyanogen. Your thinking of a App drawer replacement. Im thinking of completely different programming.
10. IEatApples (Posts: 66; Member since: 06 Jul 2011)
i agree with you remix, but wouldnt this evidence cause some questions about the exynos benchmark's too? belive me being a gs2(exynos) i dont want to believe it but it does make me wonder. gs2(exynos) runs so smooth i doubt i would care though.
14. remixfa (Posts: 14255; Member since: 19 Dec 2008)
no, as u can tell by the Exynos benchmarks, its I/O scores are not inflated compared to the rest (the skyrockets I/O scores border on rediculous comparitively). The important parts to look at are the categorys i mentioned in the previous post. With just looking at those, the exynos is still the clear winner. It IS the most powerful chip on the android market right now. there is no doubt about that.
I think samsung is just pulling this little slight of hand to reduce fears that the Crapdragon induced SGS2's are still as powerful as the Exynos SGS2s. If it was an obvious difference then they wouldnt sell as well. Thats also part of the reason the crapdragon is clocked up to 1.5ghz to compensate for the lack of actual power
. I included the 1.2ghz HTC Sensation score to give a relative comparison to the Exynos at 1.2ghz, as well as the 1.5ghz Amaze which runs the exact same hardware set up as the Hercules/skyrocket except the screen.
And before anyone points to the resolutions as the difference in power, test after test has shown that the same chipset running different resolutions has only a 1-3% affect in benchmark performance.. AKA nearly none.
65. bucky (Posts: 2279; Member since: 30 Sep 2009)
THIS! The are playing with the numbers simply to make it seem as though the Snapdragons are as fast. They are not.
13. networkdood (Posts: 6330; Member since: 31 Mar 2010)
I hear this phone has more than 1GIG RAM....is that true?
15. remixfa (Posts: 14255; Member since: 19 Dec 2008)
nope. same as the Hercules through and through.. 1 gig ram, 1.5ghz snapdragon, 4.52inch screen.. only difference is the trade from HSPA+ to LTE.
22. barondebxl (Posts: 121; Member since: 04 Nov 2011)
I guess the snapdragon is a very good chip unlike what people think. Hell it's capable of LTE and 42 Mbps.
25. remixfa (Posts: 14255; Member since: 19 Dec 2008)
neither of which have anything to do with the actual chip performance, and are just add on options available from qualcomm.