Related
I posted in another thread, but it seems to have been overlooked as it was more of a question regarding another reply and didn't fit, so I'll make a new thread.
While my MDA seems to be functioning fine with very little lag and no hanging, I am still interested in trying to overclock it.
What I'm wondering is what is the best program to do it and where can I get it? I've seen several threads that refer to a program that allows you to tell the program which things to speed up, which things to run at native speed, and which things to even slow down (why you would want to slow it more than native I don't know).
Can anyone point me in the right direction? I knew of one program from a LONG time ago that overclocked the omap processor, but I'm guessing there must be better things out there by now.
Try this thread - http://forum.xda-developers.com/showthread.php?t=270751
research batterystatus, OMAP, SmartKey, and XCPU Scaler.
Crossbow reloaded has an overclocker included.
It also features variable clocking - increasing or decreasing the speed based on the percentage of current processor usage.
Mine is charging so I cannot say exactly - but it is something like....
Standard = 195
Max = 240
Boost = 265
Minimum = 69
I dropped the minimum speed as I figure it stretches the battery life - and why spin the processor to do almost nothing - ie wait for a call or text.
In reality the lowest it goes is 99 i think.
Thanks guys. I thought the battery status thing was just a more cool way to tell how much juice was left (as opposed to the original bar-based indicator)
Looks pretty interesting. I just hope I don't fry my phone now... LOL
I was almost 100% sold on the Polaris as my next device but battery life is very important to me. I currently have an HTC P3300 (well, O2 XDA Orbit) and with the increased battery capacity of the Polaris (1350mAh vs 1200mAh in my Orbit) plus the claimed better specs from HTC (GSM 7 hours talk/400 hours standby for the Touch Cruise vs 3.5-5 hours talk/150-200 hours standby for the P3300) I really hoped that HTC had dodged any battery life issues.
I just found this review however (http://www.mobile88.com/mobilegallery/phonereview.asp?phone=HTC_Touch_Cruise&pg=review&prodid=20785&cat=37) and it has me worried. The bit I'm worried about is right at the bottom of the page:
<Start of summary of review results>
The multimedia cycle tests in comparison to the results demonstrated by the original Touch and P3300 are given below:
Multimedia-cycle, video (AVI) Polaris=4:08 Artemis=5:20 Elf=5:38
Multimedia-cycle, audio (MP3) Polaris=13:49 Artemis=21:34 Elf=18:07
<End of summary>
You can see that for MP3 the Polaris is way worse than the Artemis (I'm assuming the numbers are <hours>:<minutes> of play time). With what I commented on in my first paragraph these results really suprise me.
Does anyone know the conditions/details of the video and audio multimedia-cycle tests above? I'm wondering if somehow the conditions for the Polaris test were less favourable than those for the Artemis. Maybe the MP3 decoder software was different between the Artemis and the Polaris and the latter was dramatically less efficient although I'm probably clutching at straws here. Any other thoughts, comments or real life results from owners?
- Julian
funny how the 2 wm6 devices have lower batt time then the wm5 device
would be intresting if a test with an aramis with wm6 was don
Rudegar said:
funny how the 2 wm6 devices have lower batt time then the wm5 device
would be intresting if a test with an aramis with wm6 was don
Click to expand...
Click to collapse
That's a really interesting point. I was considering upgrading my device to WM6 using one of the cooked ROMs from this site, or the official O2 Orbit one, but I decided not to after seeing quite a few posts here by people who had done it and reported that the battery life on their devices went down a lot after upgrading to WM6 so I do wonder if that is the issue.
- Julian
I already ordered a second battery. I always do regardless of the device. One less thing to worry about
I am impressed with the battery capacity of the Polaris compared to my MIO A201. Running Tomtom without charging on the MIO A201 about 2 hours, with the Polaris 4 hours.
---Alex--- said:
I am impressed with the battery capacity of the Polaris compared to my MIO A201. Running Tomtom without charging on the MIO A201 about 2 hours, with the Polaris 4 hours.
Click to expand...
Click to collapse
What screen brightness is that set at? I'm assuming that the MP3 test results I quoted are with the screen blanked because if not then almost 24 hours on an Artemis with the screen on is pretty amazing.
If the tests do involve the screen being on (and surely they must with the video tests?) then I do wonder if the brightness settings were the same between the Artemis and Polaris tests.
If you look at some pictures further up in the review that I linked to in the first post of this thread then you can pretty clearly see that, I assume with both set to maximum brightness, the Polaris screen is noticeably brighter than the Artemis, so it isn't really fair on the Polaris to run any <screen on maximum> tests with both having the backlight set to full; it would be much fairer to adjust the Polaris backlight to give as close to possible the same brightness as the Artemis screen on full. Just maybe this accounts for some of the difference.
Thanks a lot Alex for the info on the TomTom results but I'd love to hear some real-life results of people playing music in a loop with the screen blanked and also discharge rates with the device just left at idle but with the auto-off and backlight-off disabled so that the screen stays alive. How much does the battery drain after 2 hours of sitting idle like this?
The reason I ask my questions is because the things that burn the most "activity hours" on my device are playing music with the screen blanked (hence my first request) and reading ebooks, for which my second requested test is probably a fairly reasonable approximation.
- Julian
They're comparing it to two devices with OMAP processors. Power savings is one of the reasons the OMAP was used before. It's either power or battery, rarely both unless the device is large. You could probably underclock to increase battery life.
JwY said:
They're comparing it to two devices with OMAP processors. Power savings is one of the reasons the OMAP was used before. It's either power or battery, rarely both unless the device is large. You could probably underclock to increase battery life.
Click to expand...
Click to collapse
You hit the nail right on the head. I have not bothereed to check the source so I do not know the purpsoe of the test. In any case if you want 3 days of music you would be better of with an iP.... If you want processing power then you look for a device that will not make you fall asleep just waiting for a page to refresh
JwY said:
They're comparing it to two devices with OMAP processors. Power savings is one of the reasons the OMAP was used before. It's either power or battery, rarely both unless the device is large. You could probably underclock to increase battery life.
Click to expand...
Click to collapse
I was expecting someone to say this and I'm afraid that my hunch, which I am willing to admit could be wrong, makes me disagree with this. There is some reasoning to justify my hunch and it goes as follows.
Make the following assumptions (all numbers chosen for ease of arithmetic rather than accuracy since I am just trying to demonstrate a principle rather than derive results):
1) Yes, the CPU in the Polaris is more powerful than the one in the Artemis and lets assume that the Polaris CPU (PCPU for short) is exactly twice as powerful as the Artemis CPU (ACPU for short). By twice as powerful I mean that in any given second the PCPU will process twice as many instructions as the ACPU.
2) I am assuming that both CPUs have some sort of speed stepping technology such that, when they are not under load, the power consumption drops significantly to a fairly trivial value and I will assume that for both CPUs the idle power consumption is equivalent.
3) Assume that at full load the PCPU has twice the power consumption of the ACPU.
4) Assume that it takes a 100,000,000 instructions to decode and play 1 second of music (i.e. 100MIPS = 100 million instructions per second) and that the ACPU can only just manage this so when playing music the ACPU is at 100% load for 100% of the time.
With the above assumptions my point now is that a more powerful CPU won't create a serious decline in battery life when playing music because the 100 million instructions required to be executed for 1 second of music is constant so a 100MIPS processor will need to run flat out constantly to play music whereas a 200MIPS processor will only need to be at 100% load for 0.5 seconds in any given second and for the rest of the time it can be speed stepped right down. With the idealised assumptions above there would actually be no impact whatsoever on power consumption for any arbitrary processor power (for processors that have at least sufficient power to keep up with the music stream).
A further piece of real life evidence is, if it is solely or even predominantly down to the processor, then why is the Elf managing 18:07 on the MP3 test compared to the Artemis 13:49 (and that I believe this is with a smaller battery than the Artemis, 1100mAh vs 1250mAh; info taken from the specs on the HTC web site)?
Maybe assumption (2) is wrong which does hurt my argument somewhat, or maybe there are software differences, in the MP3 player and/or in WM6 itself, that stops the PCPU dropping its power consumption down as much when it's not actively decoding, but that Elf vs Artemis test result difference still makes me wonder what else is going on.
Honestly, I'm really hoping this is just due to a badly run test on the Polaris (not same conditions as Artemis test) and that the result is an Anomaly.
- Julian
I've had the xda Stellar (tytn II), which is very very similar to the polaris, for the last week, and the battery is, I'm sorry to say, the worse I've ever come across.
For the first few days, I was using it heavily and managed 1.5 days. Figured this would increase as I used it less. Took it off charge 5 hours ago, made a 20 minute phone call, sent 3 text messages, and used the word processor for 30 minutes. 75% battery left.
I'm sending it back and waiting for the Orbit 2 to be released. Fingers crossed the keyboard has some strange battery draining feature.
sonesh said:
I've had the xda Stellar (tytn II), which is very very similar to the polaris, for the last week, and the battery is, I'm sorry to say, the worse I've ever come across.
Click to expand...
Click to collapse
The Stellar only has an 1100mAh battery while the Polaris has 1350mAh. Sounds little but makkes a differece of 200 (yes, 200!) hours of StandBy time.
This means the Stellar specs say 250h whereas the Polais spec sheet says 450h.
sonesh said:
For the first few days, I was using it heavily and managed 1.5 days. Figured this would increase as I used it less. Took it off charge 5 hours ago, made a 20 minute phone call, sent 3 text messages, and used the word processor for 30 minutes. 75% battery left.
Click to expand...
Click to collapse
Yeah. That's very bad. I wonder if there's something wrong with your unit. According to the spec on the HTC web site you should be getting talk time of "Up to 264 minutes for UMTS - Up to 420 minutes for GSM". As far as I'm aware holding a call is by far the most battery-draining thing you can do on a device so if we count your 30 minutes of word processing as another 30 minutes of talk time we're probably grossly over-estimating but, on that assumption (and adding 5 minutes for the texts) that's only an equivalent of 55 minutes of talk time. Were you on 3G or GSM for all this? I suppose if it was 3G then maybe you could have expected the battery to be down to 79% but if it was GSM then the battery really shouldn't have been much below 87%.
Admittedly I think these talk time figures quoted by manufacturers are in very ideal conditions (i.e. very high signal strength from the mast) so what sort of reception area you were in when you made the call is an issue but I think that is probably counteracted by the fact that I probably vastly overestimated the power drain when word processing.
An interesting (free) utility is abcPowerMeter (http://www.acbpocketsoft.com/). It would be very interesting to compare the current consumption on the Artemis, the Polaris and the Tyan II.
If anyone is up for this then I suggest the state to test would be with the system sitting on the today screen, disable the auto-off but do allow the backlight to go off or manually disable it so that differing screen brightness isn't a factor. Also put the phone into flight mode so that all radios are off (to factor out any effects of different signal strengths) and then start abcPowerMeter and let it run for about 30 minutes during which time it should, in theory, settle to a pretty straight line showing the current drawn from the battery in this state.
I would most certainly do an Artemis (or rather O2 XDA Orbit) test but unfortunately the reason for my intense interest in all things Polaris is that my Orbit died on new year's eve (no, I didn't drop it or sit on it at a party!) so my device is totally dead.
- Julian
Could the Today plugins be making a difference? I'm thinking particularly the weather plugin. I've seen other discussions where people report battery life problems on various devices and one often-mentioned that people suggest to check is what Today plugins are running.
With the new devices now 3G, and 3G using a lot more power than 2/2.5G when active, I'm wondering if the weather plugin is causing 3G to be active and hurting power. How often does the weather plugin connect and can it be disabled? Not wanting to take the thread off topic but one thing I'd like to do if/when I get a Polaris is to disable the weather plugin on the Today screen, is this possible?
For my use, since I am concerned about battery life, I intend to keep my device in GSM mode and only switch to 3G when I want to connect to the internet and, when finished, I will immediately disable 3G again. I certainly don't want any plugins regularly polling the internet and turning stuff on without my permission.
- Julian
acbPowerMeter
I installed acbPowerMeter this afternoon, after noticing that my battery was draining extremely fast (~ 4 to 5 hours). The tool showed that the TC was using approx 320mA on average
I've been running the tool for some hours now, and after two hours the power consumption slowly lowered. I've been switching the screen off and on and have had the GSM radio on all the time. The avg value returned to ~20mA.
I'll keep an eye on power consumption, because now I don't trust the device anymore. I don't know what caused the huge change in power consumption in the first place.
Thanks for those test results Muyz. It's great to have some real data, but I'm sorry to hear that you're having battery problems too.
Regarding your results, did you start abcPowerMeter (APM for short) while your device was still plugged into the USB port? I noticed that when I had APM running when my device was charging then I got a very high mA reading (about 250mA on my 1200mAh O2 XDA Orbit) so I think the fact that there is charge current coming in confuses APM somewhat and, for the average, it could be that after you unplug it will take a while for the minutes or hours of false high readings to creep out of the statistics. I always made very sure that my device was fully unplugged before starting an APM session.
Regarding the 20mAh reading, that actually sounds very good to me. Unfortunately I'm recalling all this from memory because, as I said in an earlier post, my device is now totally dead, but I seem to remember about 29mA as as low as I saw. With the screen on the figure of 79mA sticks in my mind. As with you, these figures were all with GSM on. I had Bluetooth and WiFi disabled.
One thing that really suprised me with my tests was that, when I had a good GSM signal, I couldn't detect any difference in current drain between having GSM switched on or off (just registered to the network of course, not actually with a call active). The additional current drawn from having GSM on didn't even seem to be 1mA. The story is different in a no-signal area where the GSM keeps searching for a signal, that drained the battery really quickly.
- Julian
3G CONNECTION need more battery consumation(if you dont need disable), configure for normale use GSM, and the duration is guarantee...
New battery monitor tool
I've discovered a serious bug in the acbPowerMeter tool. The implementation of the tool does not use the proper types
This is why I've done a little programming myself last evening. The attachment to this tool contains a preliminary version of my own battery monitoring tool. It provides the correct battery readings for remaining power and current power consumption. One can adjust the polling frequency. I will complete the tool somewhere this or next week and put it on my website for download (freeware).
Muyz said:
I've discovered a serious bug in the acbPowerMeter tool. The implementation of the tool does not use the proper types
This is why I've done a little programming myself last evening. The attachment to this tool contains a preliminary version of my own battery monitoring tool. It provides the correct battery readings for remaining power and current power consumption. One can adjust the polling frequency. I will complete the tool somewhere this or next week and put it on my website for download (freeware).
Click to expand...
Click to collapse
Hi Muyz,
Not wanting to jinx this, but I'm cautiously optimistic that I just un-bricked my O2 XDA Orbit (Artemis) so, as long as it doesn't go bad on me again I might now be in a position to contribute Artemis (or re-branded Artemis) data for comparison.
In preperation for the tests I'm really interested in what are the problems you found with abcPowerMeter. You say above that it "does not use the proper types". What do you mean by this? As a (fairly old) computer scientist this immediately makes me think of types as in a typed computer language (e.g. it's declaring something as int instead of long or short instead of bool). Is this what you mean? Please excuse my ignorance, I know theory but have never been anywhere near any Windows, let alone Windows Mobile, programming environments; if you do mean language types then how did you detect this without the source code (or did you find that somehow or disassemble)?
If you don't mean variable typing then what do you mean?
In any event, many thanks for the work Muyz.
I un-bricked my device by re-flashing (twice) the official O2 WM6 ROM so when (assuming my device stays stable) I start running current consumption tests then I think it's probably quite good that I upgraded my device to WM6 because that removes one variable (is the current drain a WM5 -> WM6 issue?) so we can at least compare Artemis to Polaris on a fairly like-for-like basis regarding the OS version.
I'm really glad I started this thread, I think there's been some really interesting and useful discussion here and I certainly wasn't expecting to have people jumping in and posting fixed versions of the current-monitoring software, that was a really nice surprise (but I'm still really looking forward to hearing exactly what the issue was).
- Julian
Battery life in cooked ROMs - weather plugin
Hi guys,
in some cooked ROMs there seems to be a bug in the radio rom or simply in general. I've read several posts where people had the gsm units in their Artemis on full power most of the time causing the device to lose power very fast. This happened only with cooked ROMs, not with official versions. They noticed that when they had their phone close to speakers which caused the buzzing noise that you usually get when there is an incoming call or short message.
The other thing that consumes power is the weather plugin, but only when you have the auto-update activated! I disabled it and never had problems.
I think I'll buy this device anyways, haven't read any serious reasons not to buy it.
Thanks for your response, and I hope your device works properly now.
About the type issue: acdPowerMeter (obviously) uses the Windows API to retrieve battery information. It receives the information through a structure that contains signed integers. It seems as if acbPowerSoft managed to introduce a typing error by using unsigned integers instead of signed integers. The effect is that for negative values (e.g. when current is drawn from the battery), acbPowerMeter shows extremely large values. I discovered what is the most likely reason for this mistake: Microsoft shows an example on their website on how to retrieve battery status information. Their example shows the error clearly: the C# class use unsigned integers, whereas the native structure contains signed integers as you can see here. So I guess acbPocketSoft copied some code without checking the result
I do not have a clue on what caused the extreme battery drain I encountered a few days ago. I have not seen it since. Soon, my tool will include a warning mechanism. I first added a few other small things, such as battery temperature, as you can see in the new attachment
(Yes, I know, it contains a small glitch on exiting the application, but that will be fixed asap)
We have so many wonderful roms and cooks in this forum. I would like to start testing with some help from others in this group all the current roms and their respective radios for battery life and performance. Since there are so many roms and many radios, I request the help of other users. Some testing has already been done thanks to Alltheway, and I would like to continue with his great work. I would like to use BattBench .Net for battery testing, and Spb Benchmark for performance testing. The links for each are available here-
BattBench .Net
http://classic.pocketgear.com/download.asp?product_id=14741
Spb Benchmark - free personal use (must submit email for link from Spb)
http://www.spbsoftwarehouse.com/products/benchmark/download.php?en#personal
For testing with .BattBench this guideline should be followed by all who help testing please
Airplane Mode - On(thanks for advice guys)
Bluetooth - Off
Wifi - Off
Exchange or Windows Live Syncing - Off
Brightness - Default Setting
Power - Default Setting
BattBench .Net requires you uncheck the Turn Off Backlight in Start, Settings, Backlight
There will be one test for each set for now with each compatible radio:
I am using a 1600mah battery for testing (I will calculate results for 1350 also)
Each test result will be an average of three tests of 3% battery consumption.
For Testing with Spb Benchmark this guideline should be followed for accurate testing please
All tests performed with default rom settings
Benchmark test using MAIN test group with default test settings(Follow SPB Caution Instructions)
Testing Begins TONIGHT!
You must set the phone to airplane mode to get numbers w/ battbench that will be the same numbers other users will get. W/o setting a standard, all numbers will fluctuate. By setting the phone to airplane mode, this does not disable the radio completely. It just disables reception. By doing so, you make it standardized because some users have 3G, some HSPDA, some Edge and others just standard signal. If you dont standardized the test will be location dependant which will tell noone anything.
Battery Test for Dutty's Official WM6.1 5.2.19199 UC RTM (3-19-08)
Seidio Battery(1600mah) Test with Radio 1.64.08.21:
***CURRENTLY TESTING***
WILL POST RESULTS AS THEY BECOME AVAILABLE
Performance Test for Dutty's Official WM6.1 5.2.19199 UC RTM
Saved for future Spb Benchmark test results
I personally think that changes the variables to ungodly proportions.
Testing variables
Well i guess youre right but it still sucks. So here are my testing variable possibilities:
Flight mode - Always on for radio reception reasons above
Bluetooth - Always off for same reasons as radio
Wifi - Always off for same reasons
What about these?
Beam - ?
Volume - ?
Brightness/Power settings - ?
Am I missing anything else anyone? I want to start first test A.S.A.P
myteematt said:
Well i guess youre right but it still sucks. So here are my testing variable possibilities:
Flight mode - Always on for radio reception reasons above
Bluetooth - Always off for same reasons as radio
Wifi - Always off for same reasons
What about these?
Beam - ?
Volume - ?
Brightness/Power settings - ?
Am I missing anything else anyone? I want to start first test A.S.A.P
Click to expand...
Click to collapse
Beam is usually disabled from start, so I usually leave it.
VOlume I leave at the default settings as well.
Brightness/Power, same as above.
Gettin somewhere
Unless stated otherwise I am just going to use the default performance test with Spb Benchmark unless someone states otherwise. Should same standards set as above be used for this performance test also? Also what perecentage tests should be used for .battbench? 100%-75%, 100%-50% 100% - dead battery, any ideas for standardizing?
myteematt said:
Unless stated otherwise I am just going to use the default performance test with Spb Benchmark unless someone states otherwise. Should same standards set as above be used for this performance test also? Also what perecentage tests should be used for .battbench? 100%-75%, 100%-50% 100% - dead battery, any ideas for standardizing?
Click to expand...
Click to collapse
Everything I have stated is purely for battery testing. I also like to use 100%-75%.
I don't know .battbench that well, is 100%-75% enough to determine the whole batteries lifespan to 0 or dead?
myteematt said:
I don't know .battbench that well, is 100%-75% enough to determine the whole batteries lifespan to 0 or dead?
Click to expand...
Click to collapse
You never want to take these ion batteries to 0%. Its not good for them. This is per GSLEON3.
Anyone feel free to test using GSM(Edge network). I will start another thread for that once results start coming in since I have around 20-30 tests to do on 3G first and Im doing this solo so far so each test takes 5-8 hours depending on results. I think there are other network types for this phone that I don't have access to in North America so thats as far as my testing can go so anyone can participate on that part.
P1Tater said:
You never want to take these ion batteries to 0%. Its not good for them. This is per GSLEON3.
Click to expand...
Click to collapse
actually, for the batteries 0 is just as bad as 100.
Optimum range for them is 40-60
Humor
It may be true but its still funny. So its bad to use your battery to its maximum potential meaning all the way, and 40-60% is the best testing range you think? How is the average joe supposed to determine how long his battery is gonna last from full to dead than? Im still a little confused about how testing from 100%-75%, and from 75%-50% is gonna tell me the whole battery life. I guess we will have to pick the one with the highest overall score in each testing class. Any further ideas/suggestions woud be great
myteematt said:
It may be true but its still funny. So its bad to use your battery to its maximum potential meaning all the way, and 40-60% is the best testing range you think? How is the average joe supposed to determine how long his battery is gonna last from full to dead than? Im still a little confused about how testing from 100%-75%, and from 75%-50% is gonna tell me the whole battery life. I guess we will have to pick the one with the highest overall score in each testing class. Any further ideas/suggestions woud be great
Click to expand...
Click to collapse
When I did my battery life testing, I tested it at the default 3% five times in a row. This way I got the average instead of just performing one test. I felt this would give me better results instead of testing it once at 25%.
Sweet
I was waitin for you to jump in here good to see you. You are the reason I have decided to do this and any further advice is greatly appreciated. I will do the 3% tests as you recommend. Since this will be alot quicker I can do both 3G and GSM testing back to back and get more results that everyone will be happy with
myteematt said:
I was waitin for you to jump in here good to see you. You are the reason I have decided to do this and any further advice is greatly appreciated. I will do the 3% tests as you recommend. Since this will be alot quicker I can do both 3G and GSM testing back to back and get more results that everyone will be happy with
Click to expand...
Click to collapse
I also agree with P1Tater that you should actually do the testing w/ Flight Mode on. Each radio will respond different to various Wireless carriers around the world. I also think that to do some serious battery testing you would have to flash to the semi-official WM6.1 or WM6 ROM's. This way no bias is going to made on using a custom ROM. I wouldn't change the PagePool or anything to that nature.
Just some ideas that might make the testing better.
So far just from using various ROM's around the forum the ones based on the 5.2.19199 builds have the best battery life hands down.
Testing
Well now that I am taking your average of the 3% test run five times advice it will be alot quicker than before. Now that testing will be shorter I can go through all the compatible radios (7 so far) for the 6.1 series. Im starting with the newest rom obviously dutty's and working my way back to the HTC factory rom within the next few weeks. In regards to the Airplane mode can't I just run one set of tests with for example with only 3G data, and then run a second set for GSM(Edge where I am). I know it won't be as accurate as everyone wants but I am encouraging others around the world to test with me. I would gladly do all the North America testing, and I could get help from the European community to test in their own locations to get better accuracy. To me it means better results for each type of user in their own geographical area. For example I always use 3G while I travel up and down the East Coast of U.S and try to avoid GSM at all cost because of the speeds since I stream A/V alot. If I am still way off course with the whole Airplane Mode thing I apologize, I guess I need another explanation, I'm kinda slow today . I just want the tests to be as accurate as possible given everyones different situations. Thanks in advance for the great advice
signal strength greatly affects battery as does carriers and data coverage. this is why they want airplane mode. Many things can affect signal strength, and thus affect your test.
Airplane mode cuts out all of those variables.
Understood
I guess I finally get the whole airplane mode thing now, thanks. First set of results should be in by breakfast. Ill correct first post now
I've recently bought a honor 8, it's beautiful blazing fast as compared to my previously owned Xperia z1 but today I ran Antutu benchmark on this device and found that my score is really very unstable... I've seen couple of videos on YouTube doing their tests of Antutu on their honor 8 devices and their score comes out more than 88k but man my phone first got 71k followed by 79k followed by 65k....
I mean like is anyone else who knows what's going on? Please guys tell me what I am doing wrong is there some kind of performance mode or something? I am running nougat (7.0) and emui (5.0).
Check out this screenshot.
Do you have programs running in the background? Antutu scores seem to vary a good bit in my experience, but ~70K seems pretty low.
Pretty consistent over here on B380. Did you make sure no apps run in background etc. during testing? It does seem a little odd
Sent from my FRD-L09 using XDA Labs
On B380 without restart the phone. I think result is good!
66k with no apps running in the background :|
victor1g said:
66k with no apps running in the background :|
Click to expand...
Click to collapse
Fresh boot...
are you sure you have a Honor 8 ? why does your model not show ? why is your battery icon look like that ?
Sounds like you got power plan enabled or something , i'd do a factory reset or figure out the issue
b360 - l04 (USA)
Serpentdrago said:
are you sure you have a Honor 8 ? why does your model not show ? why is your battery icon look like that ?
Sounds like you got power plan enabled or something , i'd do a factory reset or figure out the issue
b360 - l04 (USA)
Click to expand...
Click to collapse
Themes are a thing.
victor1g said:
Themes are a thing.
Click to expand...
Click to collapse
Can you read ?
that doesn't explain the lack of model number in the screenshots
Good job pointing out the one thing i said that can be explained ! congrats !
Just downloaded the app and did the test with no reboot or anything.
Got a better score than before, but still guys it 10k less than the average of all yours. Everyone is above 90k.
Holy crap now look at this... What should i make of this.. just few minutes after it. I won't say the hardware has issues but it's just too unstable..
charge the phone. let it cool down. test it. last 2 results were due to low battery throttle. u had 10-15% left. obv, performance will drop. it also depends in which mode u're running ur phone and whats running in the back. in the end, antutu is usually unreliable. check if you get consistent results with geekbench, when the phone is charged. honor should be around 92k on emui4.1 and 94k on 5.
you should NEVER bench (if you actually want to get good scores) while under 20% or on charger
sikica133 said:
charge the phone. let it cool down. test it. last 2 results were due to low battery throttle. u had 10-15% left. obv, performance will drop. it also depends in which mode u're running ur phone and whats running in the back. in the end, antutu is usually unreliable. check if you get consistent results with geekbench, when the phone is charged. honor should be around 92k on emui4.1 and 94k on 5.
you should NEVER bench (if you actually want to get good scores) while under 20% or on charger
Click to expand...
Click to collapse
I do not think this is true. With 17% of the battery I got this score. Honor 8 does not gradually reduce by 20%?
Sent from my FRD-L09 using XDA Labs
its a general statement. phones tend to behave weirdly when battery is low. yeah in my case it performs pretty much the same despite the %, however this might not be the case for everyone because current state of the os/apps is different from one individual to another, and it could affect the end result greatly. if he thinks his phone is faulty, he should deffo service it, however sudden drop in performance, doesnt necessarily mean the phone is damaged. these drops are common even on nexus devices +bench methodology is different from one individual to another so scores should obviously differ. yeah, 80k is not 92, BUT it can reach 80, meaning it has potential. that sudden drop from 80 to 60 means either the phone is really and i mean really rekt, or there is something else going on, which is definitely a result of stuff installed on his phone. simple factory reset and pure antutu score in that scenario will show the real picture.
sikica133 said:
its a general statement. phones tend to behave weirdly when battery is low. yeah in my case it performs pretty much the same despite the %, however this might not be the case for everyone because current state of the os/apps is different from one individual to another, and it could affect the end result greatly. if he thinks his phone is faulty, he should deffo service it, however sudden drop in performance, doesnt necessarily mean the phone is damaged. these drops are common even on nexus devices +bench methodology is different from one individual to another so scores should obviously differ. yeah, 80k is not 92, BUT it can reach 80, meaning it has potential. that sudden drop from 80 to 60 means either the phone is really and i mean really rekt, or there is something else going on, which is definitely a result of stuff installed on his phone. simple factory reset and pure antutu score in that scenario will show the real picture.
Click to expand...
Click to collapse
Does this feel better ? wow man it worked a reset did the thing..
I was surprised so i did it again just to be sure, i think it's okay now rite? Emui5.0 ...
Serpentdrago said:
Can you read ?
that doesn't explain the lack of model number in the screenshots
Good job pointing out the one thing i said that can be explained ! congrats !
Click to expand...
Click to collapse
Who are you even talking to?
The model number in everybody's screenshot is shown. You never once mentioned themes.
Drugs.
victor1g said:
Fresh boot...
Click to expand...
Click to collapse
YOUR second post in this thread , first picture .....
Doesn't look like it has a model number to me ...
drugs ?
I have two U11 (I'm going to return one, or both) and I noticed significant differences in terms of battery life between them.
Running PCMark battery life test, auto-brightness OFF, brightness MAX, one phone consistently gets 4h53 of screen on time, and the other one 5h08 (from 100% to 20% battery). That's a 5% difference and it's not negligible. Phones are the same otherwise in terms of apps installed. Problem is, the phone that has the worst battery life starts apps a TINY bit faster...
So the bottom line is, phones have significant manufacturing differences.... and I don't know which one to return.
They have same firmware and all identical? are you sure? maybe some app is configurated different or something.. or one battery its more degraded
5% is close negligible. I've got two U11 and they don't perform any differently. You aren't going to get identical numbers on two different handsets running benchmarks.
Hardware parts may be more or less efficient than each other. Radios on one device may be a few percent better. Honestly just send one back and get it done with. You also forget about coy CPU variances (no two chips are the same).
Different bin could be the reason there is a slight difference.
Ivancp said:
They have same firmware and all identical? are you sure? maybe some app is configurated different or something.. or one battery its more degraded
Click to expand...
Click to collapse
Hardware/ firmware / software is identical.
Galactus said:
5% is close negligible. I've got two U11 and they don't perform any differently. You aren't going to get identical numbers on two different handsets running benchmarks.
Hardware parts may be more or less efficient than each other. Radios on one device may be a few percent better. Honestly just send one back and get it done with. You also forget about coy CPU variances (no two chips are the same).
Different bin could be the reason there is a slight difference.
Click to expand...
Click to collapse
Well the geekbench and 3D mark are exactly the same (like give or take 0.1%).
ppaasseeii said:
Well the geekbench and 3D mark are exactly the same (like give or take 0.1%).
Click to expand...
Click to collapse
Doesn't mean a battery benchmark is going to be within 0.1% the same as when you look at Youtubers who benchmark multiple of the same device and get somewhat different Antutu benchmark results.
5% is nothing to fuss over
Galactus said:
Doesn't mean a battery benchmark is going to be within 0.1% the same as when you look at Youtubers who benchmark multiple of the same device and get somewhat different Antutu benchmark results.
5% is nothing to fuss over
Click to expand...
Click to collapse
It's like starting with a battery at 100% vs starting at a battery at 95%. For the rest of the life of the phone. Kind of a big deal.
Anyways, I'm running another test now with a low screen brightness, in case one screen was brighter than the other one (at max brightness) and drained more battery, I'll keep y'all posted.
Try with all radios (wi-fi, gsm, bluetooth, nfc, gps) off, even in location settings (disable scanning nearby devices for improved accuracy), and force stop all installed apps before the test. That gives me about 10 to 15 % increase on AnTuTu score. Maybe log cpu load and speed while testing as well.
I agree 5% is not a big deal, but would definitely keep the longer standing
Edit: force stopping because you may have installed the same apps, but most background processes won't load until you first time run their app, so one of the two "identical" devices may actually be running less apps in background. Best for testing would be to factory reset both devices and not load anything into them nor change any settings except the backlight for test.
How do you use both phones in the same manner to expect same result? You text same person with both phones making sure time taken to perform the task is same for both? And different phones have different usage pattern hence the 5% difference..??
Besides....that's 15mins. I definitely consider that as negligible. A 20% difference would be something to concern yourself about.