Database Users

Xoom Benchmarks?

Can anyone report on some benchmark scores for the Xoom?
I found this one that Anandtech uses:
http://www.glbenchmark.com/download.jsp

That is a good benchmark for general openGL performance testing. Others are:
Linpack to measure floating-point rate of execution = generally useless score.
Quadrant measures overall system performance
Smartbench 2011 measures general system performance
The xoom scores very well. The scores in GLBenchmarks will seem low, because all tests are run at the native resolution on each device. So a phone with a 480x800 resolution and a tegra 2 will get a much higher score than the xoom with the same processor and a 1280x800 resolution.

Linpack is useless for dual-core devices, it only tests one core.

muyoso said:
The xoom scores very well. The scores in GLBenchmarks will seem low, because all tests are run at the native resolution on each device. So a phone with a 480x800 resolution and a tegra 2 will get a much higher score than the xoom with the same processor and a 1280x800 resolution.
Click to expand...
Click to collapse
Good to know! Hopefully, there will be some dual core benchmarks coming out soon.

Smartbench 2011 just added dual core support! I haven't tried it yet, only quadrant, but am anxious to see the differences.
Sent from the most 'Epic' phone in the world! ...Using XDA Premium app

js042507 said:
Smartbench 2011 just added dual core support! I haven't tried it yet, only quadrant, but am anxious to see the differences.
Sent from the most 'Epic' phone in the world! ...Using XDA Premium app
Click to expand...
Click to collapse
It runs in a tiny little window on the Xoom. Don't think this is a worthy benchmark for tablets.

Wow!
in SmartBench 2011 I got highest score there - 3895/2466. Some game scores were higher on other devices, but they were overclocked to hell.
I'm impressed.

My results are posted here (benchmark and quadrant) http://4pda.ru/forum/index.php?showtopic=173393&st=320#entry6752974

keitht said:
It runs in a tiny little window on the Xoom. Don't think this is a worthy benchmark for tablets.
Click to expand...
Click to collapse
Hi guys,
I am the developer of Smartbench 2010/2011.
Basically, I use device independent pixel system for rendering those bitmaps. For some reason, Honeycomb doesn't work so well with DIP. I need to look into why.
On the other hand, I don't believe there's any other benchmarks that spawn multiple symmetric threads for performing multi-core benchmark tests. All CPU tests are absolutely not affected by the screen size/resolution at all, so it shouldn't matter. IMO, I think its a bit premature to deduce that it is not a worthy benchmark from this observation.
And as usual, I am open to suggestions!
(Oh and Hi Arcadia, long time no see! )

Discovery(Can I call it so?) about Sensation's and Galaxy S 2's CPUs' performance

Hey guys,a long time ago in a galaxy far far away(well,it was only last week,but I gotta make it sound more dramatic ),I was watching video comparisons of the two phones on youtube and reading others like a maniac,trying to decide which is best.In every review that someone made a benchmark between the two and used Linpack to measure the CPU's power(believe me,there is nothing more reliable for the CPU),the end result was that the Sensation's CPU is better clock for clock marginally.
So there I was earlier today,trying to show myself how good my purchase was,and I was comparing my GS2 with my Desire HD.Before someone comes up and tries to defend the Sensation,bear in mind that:
1)The Sensation's dual-core CPU is actually two identical Scorpion cores,like the one found inside the Desire HD,just overclocked to 1.2GHz insted of 1GHz.
2)Linpack isn't able to properly benchmark dual-core CPUs,so the result you get is from one core.
Now,as most reviews showed,on 1.2GHz(yes,my DHD is rooted etc,check my sig) the scores of the Desire HD were slightly better than those of the GS2.Better...like 46MFLOPS against 45MFLOPS or so.
Then,the spirit of the overclocker took over my mind.So I grab both phones and overclock them all the way up to 1612MHz(usign Tegrak overclock ultimate on my GS2).The results are what I want to point out with this thread.
At that clock speed,the Desire HD struggled to get 60MFLOPS(In fact I don't think it ever reached that much,more like 58-59),while the GS2 was confidently between 63 and 65 MFLOPS.
Bottom line is that the GS2's performance increases in a more "linear" way than the Sensation's.I know most people won't overclock,but I am here to defend our phones!
In case you don't even care or find this thread pointless etc,you can pretty much press your back button and get out without flaming/trolling.Please?

Hey you went to all that effort so ill say good work, and having a positive thread about the gs2 is a good thing so double thumbs up.
Sent from my GT-I9100 using XDA Premium App

NIK516 said:
Hey you went to all that effort so ill say good work, and having a positive thread about the gs2 is a good thing so double thumbs up.
Sent from my GT-I9100 using XDA Premium App
Click to expand...
Click to collapse
Haha!Yep,I thought so!
Well,it's a nice find for me.At least I've proven that the Sensation doesn't have a better CPU as its users would like to believe and boast about!
Sent from my GT-I9100 using XDA Premium App

erm... aren't sensation and desire hd two different phones? also, you first say that linpack is the best cpu benchmark out there, but then you continue by saying that it cannot test dual core cpu?
i'm a bit confused.

andrej.marinic said:
erm... aren't sensation and desire hd two different phones? also, you first say that linpack is the best cpu benchmark out there, but then you continue by saying that it cannot test dual core cpu?
i'm a bit confused.
Click to expand...
Click to collapse
Yep,one doesn't cancel the other.The only benchmarking tool that uses both cores is smartbench 2011 as of now.Others will probably follow.It's not a fault that Linpack doesn't benchmark both cores,the programmers haven't done the coding yet.Just that.
And yes,the Desire HD and the Sensation are different phones,albeit with the same CPU CORE,not the same CPU.
Sent from my GT-I9100 using XDA Premium App

tolis626 said:
Hey guys,a long time ago in a galaxy far far away(well,it was only last week,but I gotta make it sound more dramatic ),I was watching video comparisons of the two phones on youtube and reading others like a maniac,trying to decide which is best.In every review that someone made a benchmark between the two and used Linpack to measure the CPU's power(believe me,there is nothing more reliable for the CPU),the end result was that the Sensation's CPU is better clock for clock marginally.
So there I was earlier today,trying to show myself how good my purchase was,and I was comparing my GS2 with my Desire HD.Before someone comes up and tries to defend the Sensation,bear in mind that:
1)The Sensation's dual-core CPU is actually two identical Scorpion cores,like the one found inside the Desire HD,just overclocked to 1.2GHz insted of 1GHz.
2)Linpack isn't able to properly benchmark dual-core CPUs,so the result you get is from one core.
Now,as most reviews showed,on 1.2GHz(yes,my DHD is rooted etc,check my sig) the scores of the Desire HD were slightly better than those of the GS2.Better...like 46MFLOPS against 45MFLOPS or so.
Then,the spirit of the overclocker took over my mind.So I grab both phones and overclock them all the way up to 1612MHz(usign Tegrak overclock ultimate on my GS2).The results are what I want to point out with this thread.
At that clock speed,the Desire HD struggled to get 60MFLOPS(In fact I don't think it ever reached that much,more like 58-59),while the GS2 was confidently between 63 and 65 MFLOPS.
Bottom line is that the GS2's performance increases in a more "linear" way than the Sensation's.I know most people won't overclock,but I am here to defend our phones!
In case you don't even care or find this thread pointless etc,you can pretty much press your back button and get out without flaming/trolling.Please?
Click to expand...
Click to collapse
In regards to the linpack, i think i've seen the same comparison review. I mentioned it in the same thread that i ran the test and also got near 47MFLOPS when i ran the test. Same thing goes for quadrant, i've hit as high as 3450ish on stock rom and then gotten as low as 2900. I've posted in the same website that their testing method is flawed because they don't run multiple tests

I hate the fact that in comparision reviews between GS2 and Sensation, reviews usually just say both of them use 1.2GHz dual-cores and fail to mention that the CPUs are not the same.
As tolis626 pointed out that the CPU's cores in the sensation are the old Scorpion CPU used in the DHD but have one more core. They are partially out of order, while the Cortex-A9 in GS2 are fully out of order which translate to a win for GS2.
On the Linpack benchmark, it measures floating point performance and Scorpion has a better implementation of floating point unit hence Scorpion is better than Cortex-A9 clock for clock in Linpack. But smartphone application is not bound by floating point performance, according to AnandTech, so we won't see that extra performance all the time.

well I got 48.208MFLOPS without any overclock on my SGS2

Like the majority of users, I'm a noob when it comes to overclocking/rooting/benchmarking, I just haven't gone near it (yet )
I watch others' videos on it. And basically, on the stock browser, on the stock ROM, the GSII always get significantly better than the Sensation....so I'm happy with my purchase. Seems soo pointless going for a "HTC" because it has the "HTC" logo, even though it's browser is poor and it's signal fails.
ok, the above is a stupid troll comment. But you get the point.

nhat2991 said:
I hate the fact that in comparision reviews between GS2 and Sensation, reviews usually just say both of them use 1.2GHz dual-cores and fail to mention that the CPUs are not the same.
As tolis626 pointed out that the CPU's cores in the sensation are the old Scorpion CPU used in the DHD but have one more core. They are partially out of order, while the Cortex-A9 in GS2 are fully out of order which translate to a win for GS2.
On the Linpack benchmark, it measures floating point performance and Scorpion has a better implementation of floating point unit hence Scorpion is better than Cortex-A9 clock for clock in Linpack. But smartphone application is not bound by floating point performance, according to AnandTech, so we won't see that extra performance all the time.
Click to expand...
Click to collapse
What I was trying to get at is that the Sensation is NOT better clock for clock,it just gives better numbers at that frequency.It also kneels when underclocked and under terrible load,something the GS2 doesn't suffer from.
Don't be mistaken,the Scorpion CPU core is a really good one.And as far as I'm concerned it's by far better than the CPU of the Hummingbird inside the original GS.What plagues the Sensation is a terrible implementanion of the Scorpion core in a dual-core chipset.We should wait till some devs do their magic on that thing before we can call it worse.
Just my...5 cents,I wrote too much!
Sent from my GT-I9100 using XDA Premium App

tolis626 said:
And as far as I'm concerned it's [scorpion] by far better than the CPU of the Hummingbird inside the original GS.
Click to expand...
Click to collapse
is this a subjective or an objective observation?
http://smartphonebenchmarks.com/for...-gen-snapdragon-processors-how-fast-are-they/
they both seem quite competitive

vizir said:
is this a subjective or an objective observation?
http://smartphonebenchmarks.com/for...-gen-snapdragon-processors-how-fast-are-they/
they both seem quite competitive
Click to expand...
Click to collapse
Woah,I did state that ONLY the CPU is better on the QSD8255,not that the whole SoC is better than the Hummingbird.The latter was a beast in the graphics department(and still is),so it's more than competitive actually.
Sent from my GT-I9100 using XDA Premium App

Benchmarks

Tegra got smoked
http://androidcommunity.com/galaxy-s-iii-quad-core-benchmarks-blow-us-away-20120503/

Hm.. Tegra 3 seems to win the GPU bench?
Quadrant sucks though. Will be waiting for more benchs.

Here are some other benchmarks (next to quandrant also SunSpider and browsermark)

Quadrant scores -
sgs3
cpu - 12781
memory - 4652
io - 7606
2d - 1000
3d - 2171
total - 5642
S4
cpu - 8505
memory - 7547
io - 6394
2d - 990
3d - 2204
total - 5128
tegra3
cpu - 12493
memory - 3472
io - 4769
2d - 962
3d - 2346
total - 4804
they seem pretty even in cpu/gpu capability. the s4 gets smoked in cpu performance according to those quadrant scores. interesting.. i thought it was faster.

Wow, the S3 doesn't seem to smoke the competition.

im interested about the antutu...can someone bring some more benchmarks?

Even if Tegra got smoked games will look better, so i'm really thinking how the S3 will be better in general use except for benchmarks...
Sent from my Quad Core Monster the HTC One X using Tapatalk v 2

GPU performance gap between Tegra3 and Exynos4 Quad is huge
http://www.anandtech.com/show/5810/samsung-galaxy-s-iii-performance-preview
Faster than iPhone4S' powervr sgx543mp2

LOL quadrant sucks so hard. the new mali 400 is killing tegra 3. according to http://www.anandtech.com/show/5810/samsung-galaxy-s-iii-performance-preview

The benchmarked result http://www.icsforums.com/news/samsung-galaxy-s-iii-gets-benchmarked-shows-plenty-promise.html

bocautrang.pt said:
The benchmarked result http://www.icsforums.com/news/samsung-galaxy-s-iii-gets-benchmarked-shows-plenty-promise.html
Click to expand...
Click to collapse
um no
That's stolen straight out of engadget and they have recorded the lowest performance out of any of these tech blog/news sites

Antutu Benchmark available here:

ph00ny said:
um no
That's stolen straight out of engadget and they have recorded the lowest performance out of any of these tech blog/news sites
Click to expand...
Click to collapse
In the video above it's got 5324 Quadrant and 11492 Antutu. Not much different from HOX.

Here are some results from the Swedish androidsite "Swedroid". It's in swedish (doh!) but you should be able to look thru the scores anyway
http://www.swedroid.se/hands-on-med-samsung-galaxy-s-iii-forhandsitt/#disqus_thread
In my humble opinion. To say that the exynos "kills" the Tegra 3 is just...plain...wrong.
They seem to be very capable CPUs in both of these beasts!

umd said:
In the video above it's got 5324 Quadrant and 11492 Antutu. Not much different from HOX.
Click to expand...
Click to collapse
Did you see the gpu scores. Kills tegra 3.
Sent from my GT-I9100 using Tapatalk 2

Imo they are very close to eachother, i would say exynos wins by a little margin could be just the software dragging it down, would love to see a pure ICS build doing the test.

Benchmarks mean squat.. it comes down to user experience.. if u can really feel the difference.. also, apps need to be fine tuned for quad core too.. so I guess u won't feel an actual difference..
Sent from my HTC One X using XDA

jits1988 said:
Benchmarks mean squat.. it comes down to user experience.. if u can really feel the difference.. also, apps need to be fine tuned for quad core too.. so I guess u won't feel an actual difference..
Sent from my HTC One X using XDA
Click to expand...
Click to collapse
good benchmark do shows the raw power of the mobile, its never meant to test the user experience.

jits1988 said:
Benchmarks mean squat.. it comes down to user experience.. if u can really feel the difference.. also, apps need to be fine tuned for quad core too.. so I guess u won't feel an actual difference..
Sent from my HTC One X using XDA
Click to expand...
Click to collapse
All the hands-on video clip made it seemed as if s3 is doing great in general ui transitions. It was noticeably quicker when used side by side.

Always astounds me how people are judging by benchmarks. I mean, its a phone, not a server station.
I wont be running virtual machines for rendering movies on it. I wont be playing Crysis 3 on it.
I`ll phone, message, watch clips, pictures, surf and use navigation. Even the single core phones can do that perfectly.

North American Samsung Galaxy S III LTE Hands-on

http://www.youtube.com/watch?v=_bq3E_RfzxE&feature=g-u-u
You're welcome!

Benchmarks?
Sent from my GT-I9300 using XDA

SlimJ87D said:
Benchmarks?
Sent from my GT-I9300 using XDA
Click to expand...
Click to collapse
End of the video?

madhad said:
End of the video?
Click to expand...
Click to collapse
Awesome. Gf is sleeping can't watch. How much ram is actually available after adreno and the 4g lte takes up some ram?
Sent from my GT-I9300 using XDA

youtube blocked at work, can someone please summarize the video?

tsukurimashou said:
youtube blocked at work, can someone please summarize the video?
Click to expand...
Click to collapse
Good 14min summary of the phone. Much like every other S3 video. Does two benchmarks at the end.
Speed test gives ~21mb down and ~11 up.
quadrant was blazing fast during the benchmark. Scored 5099, beating everything in the market.
Sent from my Samsung Infuse running AOKP.

Thanks! What is the score for the int edition? Is he talking about battery life in the video?

DangKid said:
Good 14min summary of the phone. Much like every other S3 video. Does two benchmarks at the end.
Speed test gives ~21mb down and ~11 up.
quadrant was blazing fast during the benchmark. Scored 5099, beating everything in the market.
Sent from my Samsung Infuse running AOKP.
Click to expand...
Click to collapse
The international version gets around 5665. Just saying.
Sent from my GT-I9300 using XDA

I have the international quad core version. Benchmark around 5200 - 5400 everyttime. The advantage comes from mostly the cpu score where all cores are used. However apart from benchmarks, I never saw more than 2 cores being used.

nativestranger said:
I have the international quad core version. Benchmark around 5200 - 5400 everyttime. The advantage comes from mostly the cpu score where all cores are used. However apart from benchmarks, I never saw more than 2 cores being used.
Click to expand...
Click to collapse
Quadrant is the worst benchmark to compare anything to now a days.

SlimJ87D said:
Quadrant is the worst benchmark to compare anything to now a days.
Click to expand...
Click to collapse
It has suddenly become the worse benchmark because somehow the dual core s4 can keep up with the tegra 3 and not far off exynos 4. People like to see benchmarks that's plays to their devices advantages. Antutu has become the new favorite as it scales well to four cores and not that critical of tegra 3 and exynos lack of memory bandwidth/ IO performance against krait. However in realworld usage I would say even quadrant paints an over optimistic idea of actual performance for quad core devices. Despite similar quadrant scores, multiple users who have both commented on faster and more fluid experience on the krait chip.
http://forum.xda-developers.com/showthread.php?t=1685030

Why do people argue over benchmarks? Both the s4 and exynos are fast and fluid and won't go wrong with either.
Sent from my GT-I9300 using Tapatalk 2

Wonder if the 2GB of RAM will help the S4 not lag. Especially while doing the popup video.

SiNJiN76 said:
Wonder if the 2GB of RAM will help the S4 not lag. Especially while doing the popup video.
Click to expand...
Click to collapse
Pop up video doesnt lag on my quad core version. However i am experiencing out of memory home screen restarts and browser crashed when opening too many tabs with flash enabled. My music player also auto quits sometimes when open up a new tab in browser or using any other memory hungry apps. Very annoying.

I have noticed using TouchWiz/Apex was causing the phone to lag a lot. I switched to Nova and the experience is smooth as silk. No freezing or ram issues yet, and i normally have 20 tabs open on Opera Mini + Music playing and whatsapp + email + gtalk+ wordfeud + a calendar widget + Gtasks widget
Opera Mini does its own memory management tho, clears pages once in a while, but not often.
edit - using the international quad core version -

tsukurimashou said:
Thanks! What is the score for the int edition? Is he talking about battery life in the video?
Click to expand...
Click to collapse
He had used the phone for over 8 hours before filming and the battery was at 40%. Looks like a strong battery, should definitely last as long as other phones.
Sent from my Samsung Infuse running AOKP.

So i guess, both the International version owners and the N American version owners have something to look forward to when Samsung actually releases the next generation Exynos. Maybe then, the phone will truly be Universal (same internals/user experience).

where is the branding on the back?

nativestranger said:
It has suddenly become the worse benchmark because somehow the dual core s4 can keep up with the tegra 3 and not far off exynos 4. People like to see benchmarks that's plays to their devices advantages. Antutu has become the new favorite as it scales well to four cores and not that critical of tegra 3 and exynos lack of memory bandwidth/ IO performance against krait. However in realworld usage I would say even quadrant paints an over optimistic idea of actual performance for quad core devices. Despite similar quadrant scores, multiple users who have both commented on faster and more fluid experience on the krait chip.
http://forum.xda-developers.com/showthread.php?t=1685030
Click to expand...
Click to collapse
No, it's not programmed to bench current day processors that's why.

Send from the Noteorious BIG 5.3" Bell Canada

SD865 performance throttled to match Exynos990

It seems that Samsung might be throttling the SD865 performance to decrease the gap between it and the Exynos version. Check this video to see how the SD S20ultra compares to other SD865 phones in the market right now, such a gap indeed.

Deleted.

Also check out this Antutu benchmark score from my Exynos S20U.

mohammed510 said:
Also check out this Antutu benchmark score from my Exynos S20U.
Click to expand...
Click to collapse
Is this a freezer job?

mohammed510 said:
Also check out this Antutu benchmark score from my Exynos S20U.
Click to expand...
Click to collapse
how did you get such a high score? mine doesn't go higher than 527k

This is pretty ordinary if Samsung is throttling the s20 865... Definitely a dog act
I just ordered a G9880 from hong kong... Maybe I should cancel my order..
---------- Post added at 06:25 PM ---------- Previous post was at 06:23 PM ----------
Any throttling will be via the kernel wont it? Can any developers tell us if there is anything that we should know about???

cheetah2k said:
This is pretty ordinary if Samsung is throttling the s20 865... Definitely a dog act
I just ordered a G9880 from hong kong... Maybe I should cancel my order..
---------- Post added at 06:25 PM ---------- Previous post was at 06:23 PM ----------
Any throttling will be via the kernel wont it? Can any developers tell us if there is anything that we should know about???
Click to expand...
Click to collapse
Most likely a kernel thing. Some, if not all bootloaders in the SD version can't be unlocked, so nothing can be done about it (like flashing a custom kernel that won't throttle the chip)..

Then the gap of battery life will be even bigger. 865 will be even more efficient after throttling

I think you guys being paranoid on the throttling thing. But me and another guy ran 3dmark benchmark on my Snapdragon vs his Exynos and my scores were better. So if they are throttling the SD it still kicking Exynos butt! Get the 3dmark app off playstore and run slingshot extreme and post your scores.

Paul_Deemer said:
I think you guys being paranoid on the throttling thing. But me and another guy ran 3dmark benchmark on my Snapdragon vs his Exynos and my scores were better. So if they are throttling the SD it still kicking Exynos butt! Get the 3dmark app off playstore and run slingshot extreme and post your scores.
Click to expand...
Click to collapse
Have no clue about the results compared to the 865.....but here are mine.....

lch920619x said:
Then the gap of battery life will be even bigger. 865 will be even more efficient after throttling
Click to expand...
Click to collapse
If it's true. Then it might be hard to tell when exactly they started doing it. Might be from the very beginning. Or right after people started signing the petition to stop using Exynos.

Aezhyr said:
how did you get such a high score? mine doesn't go higher than 527k
Click to expand...
Click to collapse
How did you get such a high score? Mine doesn't go higher than 474k

[/COLOR]
Kjam33 said:
How did you get such a high score? Mine doesn't go higher than 474k
Click to expand...
Click to collapse
photoshop

Duncan1982 said:
Have no clue about the results compared to the 865.....but here are mine.....
Click to expand...
Click to collapse
Hate how you can't just upload a photo directly. Anyway I don't have a photo website where I can link to my screenshot so will just type this out manually.
SM-G988U1
April 4, 2020 23:19
Overall score 9251
Graphics score 12477
Graphics test 1 (FPS) 72.60
Graphics test 2 (FPS) 43.30
Physics score 4856
Physics test 1 (FPS) 84.90
Physics test 2 (FPS) 53.20
Physics test 3 (FPS) 28.90

I don't think Samsung is intentionally throttling performance of Snapdragon 865 units. Samsung has been traditionally conservative (aka "non-aggressive") in performance tweaking. That means they uses the reference best Q (energy efficiency) tweaks from Qualcomm. However some OEMs (like Xiaomi, Realme and Oneplus) are known to be more aggressive in performance tweaking. They are more willing to sacrifice thermal/energy efficiency to get better perf/scores. It is unfair to compare Samsung devices with them and say "Samsung throttle Snapdragon to make Exynos not to look so bad". (You will probably get the same results when comparing Samsung's older devices with other OEM's older devices.)
Plus, Samsung devices get some crappy special security additions which impair performance and their software quality is kinda mediocre when it comes to performance optimization.
BTW, for S20 series, "Optimized" power mode is actually a power saving mode. Only "High Performance" mode enables full power of the phone.

Some say, if you get the Exynos S20U then what you're really getting is just a S20U "Lite". But we can now see the same on the SD865 version. What they are getting is a SD865 "Lite". Regardless if Samsung is intentionally doing it or not. Seems we are all on the same boat.

Thank you for this post. Haven't laughed so hard in ages. Enjoy your phone, speed tests are absolutely [emoji817] percent worthless!!!
Sent from my SM-G988U1 using Tapatalk

Another video. But this time the SD S20ultra score is even lower at 534k! Wow!

If you like score buy a red magic!

mohammed510 said:
Some say, if you get the Exynos S20U then what you're really getting is just a S20U "Lite". But we can now see the same on the SD865 version. What they are getting is a SD865 "Lite". Regardless if Samsung is intentionally doing it or not. Seems we are all on the same boat.
Click to expand...
Click to collapse
My SD Ultra Lite still faster than a EXYNOS Ultra Lite. Plus it runs cooler and battery life is better. So no I'm not in the same boat. Waving to Mohammed as I leave his boat in my wake.