Nexus 4 annihilated by iPhone 5 in a variety of benchmarks

[DeathChill]

I doubt those initial Anandtech benchmarks were even close to correct. The benchmarks I have seen on other devices with the same Snapdragon S4 Pro SOC are way better.

It's not the first time Ive seen that site give great marks to an iPhone and poor marks for an Android device only latter to see completley different benchmarks for the same device on other tech sites. It makes me very suspicious.

So what you're saying is that it's Anandtech's fault? I find that hard to believe.

Anandtech does full, fair testing. I haven't seen any improved benchmarks pop-up, besides Anandtech's own improved ones due to putting the phone in the freezer.

 

[SlCKB0Y]

So what you're saying is that it's Anandtech's fault? I find that hard to believe.

It doesn't need to be their fault for their benchmarks to be invalid. What if the benchmark software wasn't functioning correctly under the brand new OS (4.2), what if the review unit they were using wasn't running the final software.... or hardware??

All of these things could make the test invalid through no fault of Anand nor any bias on their part.

 

[cynics]

----------



On device sales - 4 dollars is optimistic. They make money from knowing where you are at every second of every day, reading your email, reading your texts, etc to target you better for ads.

It's not a bad thing, of course, as the phone costs only 300... Sell your soul much? lol

----------



.

Lol, so Apple tracking you for iAds is ok because you are paying 600-1000 unsubsidized? LOL

 

[DeathChill]

It doesn't need to be their fault for their benchmarks to be invalid. What if the benchmark software wasn't functioning correctly under the brand new OS (4.2), what if the review unit they were using wasn't running the final software.... or hardware??

All of these things could make the test invalid through no fault of Anand nor any bias on their part.

The poster said it was suspicious on Anandtech's part.

Regardless, it was the final hardware and the final software was released to the phones before the review went up.

Any benchmarks that are improved over Anandtech's?

 

[jsw]

Also, the N4 cheats, horribly. It overclocks itself for burst periods to look good in benchmarks. If you look at the longer benchmarks like sunspider10x (10x the length of sunspider) it gets KILLED.

Overclocking for burst periods seems like exactly the behavior I want in a device, since I don't tend to be staring at it for long periods of time as it does benchmarks. I want it to be responsive when it needs to be.

So... I don't see your point about the overclocking. Phone app use is mostly sprints, not marathons.

 

[Oletros]

Also, the N4 cheats, horribly. It overclocks itself for burst periods to look good in benchmarks. I

Any proof that it overclocks?

 

[Dr McKay]

overclocks itself for burst periods to look good in benchmarks

Are you positive it Overclocks? Or does it just run at max clock speed? You are aware processors do downclock when idle? Infact the Tegra 3 "quad-core" actually has 5 cores, a lower clocked "companion core" which takes over and keeps the phone running when you lock it.

 

[cynics]

How is that cheating even if does overclock (which it doesn't)? The user will experience that when using the phone. However it runs at max clock speed when required, all modern devices do this even the iPhone.

 

[Tinmania]

Apple did it with the 4S' battery problems, and it took more than one update to fix for everyone.

Well that everyone didn't include me. My 4S's battery life is horrible--even now.

I can't see how my due-to-arrive-today N4 can be any worse. I am just starting work now and my 4S is at 50%. It was at 100% when I went to sleep last night. Basically just used iMessage, checked email, checked ups.com (about 30 times lol), and took a few turns on Words With Friends. Oh and two short phone calls. If I didn't have a charger everywhere I go, or spare external batt pack when I don't, I'd be in trouble.

This is true.

But, shouldn't a quad-core phone be performing better than some of its older dual-core counterparts? Is the software just not fully optimized yet for quad-core?

Personally, I don't care much about benchmarks. The reviews have said the device is smooth and fast and that's enough for me, but it is a bit of a mystery why it's not performing well in these tests. What's going on?

I don't know about you but my normal day-to-day usage pattern does not include running benchmarks. The Galaxy Nexus feels fine to me speed-wise. I have little doubt that the N4 will be better.

And no benchmark is going to make the iPhone 5 have a bigger screen.



Michael

 

[jsw]

Well that everyone didn't include me. My 4S's battery life is horrible--even now.

Ditto for me. I have a slide-on battery (which I don't activate until the 4S gets well below 50%), and I need to turn it on every day before leaving work, and that's with a phone that starts the day at 100%.

I'm hoping the N4 makes it through the day. If so, that's all I need. If not, it's no worse than the 4S.

 

[onthecouchagain]

And no benchmark is going to make the iPhone 5 have a bigger screen.

Touche, Michael.

----------

I can't imagine an HSPA+ phone not lasting the whole day. My Galaxy Nexus lasts easily even with heavy use.

I think y'all be fine.

 

[jsw]

I can't imagine an HSPA+ phone not lasting the whole day. My Galaxy Nexus lasts easily even with heavy use.

I think y'all be fine.

I dunno. I'm sort of terrified. From what I've heard of the unboxings, it's showing around 50% battery life as soon as they turn it on out of the box!

I assume that means just turning it on uses half the battery. Even my 4S takes at least 5 minutes to get to the 50% mark.

 

[onthecouchagain]

I dunno. I'm sort of terrified. From what I've heard of the unboxings, it's showing around 50% battery life as soon as they turn it on out of the box!

I assume that means just turning it on uses half the battery. Even my 4S takes at least 5 minutes to get to the 50% mark.

And turning the screen off drains the other 50%. It's that damn analog TV animation! The Nexus 4 is a one-charge-one-time use.

 

[jsw]

And turning the screen off drains the other 50%. It's that damn analog TV animation! The Nexus 4 is a one-charge-one-time use.

Well, in the end, I'll still have that pretty back-side reflection. As a work of art, it's easily worth twice what I paid, and the frame will be cheap, since it's so small.

 

[cynics]

I dunno. I'm sort of terrified. From what I've heard of the unboxings, it's showing around 50% battery life as soon as they turn it on out of the box!

I assume that means just turning it on uses half the battery. Even my 4S takes at least 5 minutes to get to the 50% mark.

That doesn't mean much though. Battery should still be in its standby state when it leaves the factory. It will hold that charge for months, if a lithium ion goes completely flat (0 vdc) then it usually isn't salvageable.

Thats why you should keep old devices topped off and not sitting on a shelf collecting dust.

 

[dalbir4444]

That doesn't mean much though. Battery should still be in its standby state when it leaves the factory. It will hold that charge for months, if a lithium ion goes completely flat (0 vdc) then it usually isn't salvageable.

Thats why you should keep old devices topped off and not sitting on a shelf collecting dust.

I think he was being sarcastic or maybe I didn't catch your sarcasm.

 

[jsw]

I think he was being sarcastic...

Yup.

 

[cynics]

Yup.

Well uh, well uh, I was being sarcastic too!!

 

[flopticalcube]

And turning the screen off drains the other 50%. It's that damn analog TV animation! The Nexus 4 is a one-charge-one-time use.

That will still work out cheaper than the iPhone 5.

 

[vcoleiro1]

The poster said it was suspicious on Anandtech's part.

Regardless, it was the final hardware and the final software was released to the phones before the review went up.

Any benchmarks that are improved over Anandtech's?

Ive seen other benchmarks for devices with the same SOC with totally different results. Don't believe me ,look them up and see for yourself. For example other benchmarks gave the same SOC a score of 1200(using default Chrome browser) for the sunspider test - big difference to 1800. All the other tests were way higher for graphics etc .

As others have said, it usually takes a while to tune the benchmarking softare to new SOC's - particularly Quad core ones. There's a number of factors they need to get right. Expect to see a whole different set of benchmarks for the SOC in 2 months time. Also, the sunspider test is run on stock browsers , the default browser under Android isn't the snappiest, so that result isn't a true indication of the HW capabilities - other snappier browsers could have been used - it's a bit of a silly test really - ie it doesn't reflect the so much the HW as it does the software (ie browser).

Example, here are some benchmarks for the Nexus 10 sunspider test

N10 using Chrome (default stock browesr) 1365ms

N10 using the QQ Browser 881ms

Reference - Iphone 5 (safari) 905ms

Quite the big difference isn't it!


Anandtech seem to race out with the benchmarks before they have gotten the benchmark software sorted for the SOC they are testing. It looks like they just want to be first to publish

To say the behemouth Snapdragon Pro S4 has mediocre-poor performance is ridiculous.

 

[DeathChill]

Ive seen other benchmarks for devices with the same SOC with totally different results. Don't believe me ,look them up and see for yourself. For example other benchmarks gave the same SOC a score of 1200(using default Chrome browser) for the sunspider test - big difference to 1800. All the other tests were way higher for graphics etc .

As others have said, it usually takes a while to tune the benchmarking softare to new SOC's - particularly Quad core ones. There's a number of factors they need to get right. Expect to see a whole different set of benchmarks for the SOC in 2 months time. Also, the sunspider test is run on stock browsers , the default browser under Android isn't the snappiest, so that result isn't a true indication of the HW capabilities - other snappier browsers could have been used - it's a bit of a silly test really - ie it doesn't reflect the so much the HW as it does the software (ie browser).

Example, here are some benchmarks for the Nexus 10 sunspider test

N10 using Chrome (default stock browesr) 1365ms

N10 using the QQ Browser 881ms

Reference - Iphone 5 (safari) 905ms

Quite the big difference isn't it!


Anandtech seem to race out with the benchmarks before they have gotten the benchmark software sorted for the SOC they are testing. It looks like they just want to be first to publish

To say the behemouth Snapdragon Pro S4 has mediocre-poor performance is ridiculous.

It's not Anandtech's fault that the software (such as the browser that comes with the phone) isn't optimized as of yet. The iPhone gets subjected to the same treatment upon release.

I don't think anyone is saying that its performance is poor. In fact, all we've really seen is browser based benchmarks which are more about the optimization in the browser software.

The graphics benchmarks all seem to be in line as far as I can tell. They are clocking in at number similar to the Optimus G (when thermal throttling is not a factor, of course).

 

[Vegastouch]

The phone is fast and the battery lasts the whole day under normal use and its unlocked. The only benchmark i care about is $349. The others,...who gives a s***.

Oh wait, there is one benchmark that is a near fail for me but could be do-able...........16GB.

 

[vcoleiro1]

It's not Anandtech's fault that the software (such as the browser that comes with the phone) isn't optimized as of yet. The iPhone gets subjected to the same treatment upon release.

I don't think anyone is saying that its performance is poor. In fact, all we've really seen is browser based benchmarks which are more about the optimization in the browser software.

The graphics benchmarks all seem to be in line as far as I can tell. They are clocking in at number similar to the Optimus G (when thermal throttling is not a factor, of course).

The results posted by Anandtech are very missleading, the N4 sunspider figure is only meant to be compared to other android devices for Chrome only, yet they put it in a list which included the iphone 5 running Safari.

It's like running a test to see how quickly MS Word runs on a PC and comparing it with how long Excel takes to run on a MAC. It's completely wrong to do and misleading.

I don't have much faith in anyone that will compare two completely different things like that and comment that that one is then faster. All credibility is lost in my eyes, right there.

 

[DeathChill]

The results posted by Anandtech are still wrong, it should be 1200 for sunspider for chrome as opposed to the 1800 they said. Also, that figure is only meant to be compared to other android devices for Chrome, yet they put it in a list which included the iphone 5 running Safari.

It's like running a test to see how quickly MS Word runs on a PC and comparing it with how long Excel takes to run on a MAC. It's completely wrong to do and misleading.

How are his results wrong? They are what the Nexus 4 scored. You can say it should have scored something different, but it didn't. That's on Google to update and optimize the software.

The idea of the SunSpider benchmark shows how each OS and browser performs in an identical test. It isn't a pure CPU benchmark meant to say, "This device is faster and always will be."

 

[Grolubao]

Who cares about benchmarks, it's frickin half the price of an iPhone 5