I just updated like 2 hours ago, installed a few apps got root and twrp on it.
Installed antutu just to check the score and wow a score i got with encrypted on.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
My score on 5.0.1 was 43k which i thought was ok but this new score with many apps installed and 10+ in recent was not bad at all.
What a snappy beast i got, much better then what i got when it was released!
thats a normal score on a nexus 6, 5.0, 5.01, 5.02, or 5.1. it hasnt changed. your score of 43000 was just a very low score to begin with.
simms22 said:
thats a normal score on a nexus 6, 5.0, 5.01, 5.02, or 5.1. it hasnt changed. your score of 43000 was just a very low score to begin with.
Click to expand...
Click to collapse
Well 43k was my score when i brought the device with 5.0 on it.
To me this is a huge step forward, also the performance of the device is much better now.
Android-Desire said:
Well 43k was my score when i brought the device with 5.0 on it.
To me this is a huge step forward, also the performance of the device is much better now.
Click to expand...
Click to collapse
i was seeing 53000+ on 5.0, stock. your 43000 score was a very very low nexus 6 score. 53000+ is the normal score for a nexus 6, regardless of what your score was before. i bet it was so low because youre benchmarking it wrong. if your cpu is set to max cpu speed and min cpu speed, then thats the wrong way to bench. as your phone then decides what speed itll use. youre supposed to set your cpu speed for max cpu/max cpu, so it only tests your max cpu speed, so it will only give you your highest scores. otherwise, you have no idea what cpu speed its actually testing. obviously 43000 it wasnt testing your max cpu speed.
simms22 said:
i was seeing 53000+ on 5.0, stock. your 43000 score was a very very low nexus 6 score. 53000+ is the normal score for a nexus 6, regardless of what your score was before. i bet it was so low because youre benchmarking it wrong. if your cpu is set to max cpu speed and min cpu speed, then thats the wrong way to bench. as your phone then decides what speed itll use. youre supposed to set your cpu speed for max cpu/max cpu, so it only tests your max cpu speed, so it will only give you your highest scores. otherwise, you have no idea what cpu speed its actually testing. obviously 43000 it wasnt testing your max cpu speed.
Click to expand...
Click to collapse
43k was stock without any modifications to cpu kernel or anything that should interfer with the benchmark, same as this screenie. stock settings all the way.
Why is it everytime i post something here that all people take it like this is my first time using an android phone? I have build roms before - 2.3.4 on HTC Desire - 2.3.4 on LG Optimus 2x and a cupcake 1.5 rom for a huawei and i have also build 5.0 aosp for this Nexus 6.
My experience with this device at 43k was when i was omw home with only antutu installed on phone and not even connected it to pc yet.
I think what he is saying is other people, like myself, got in the fifties with a standard stock 5. 0 device.
If it really was a per device issue then I would think I would get in the sixties on 5.1 , but I don't think that's going to happen. i. e. you going from 5 .0 to 5. 1 did not jump your stats up by 10k.
But I am glad to hear it resolved your issues.
Sent from my Nexus 6
Android-Desire said:
43k was stock without any modifications to cpu kernel or anything that should interfer with the benchmark, same as this screenie. stock settings all the way.
Why is it everytime i post something here that all people take it like this is my first time using an android phone? I have build roms before - 2.3.4 on HTC Desire - 2.3.4 on LG Optimus 2x and a cupcake 1.5 rom for a huawei and i have also build 5.0 aosp for this Nexus 6.
My experience with this device at 43k was when i was omw home with only antutu installed on phone and not even connected it to pc yet.
Click to expand...
Click to collapse
well, you were doing something wrong then, whatever it was. 43000 is a nexus 5 antutu score, lol. btw, me, and everyone i know with a nexus 6 hit 53000+ on stock, while still stock. btw, just because you can build a rom, doesnt mean that you know everything android, as simple noobs can build aosp from source, and add things to it.
43k is low. I normally get around 55k.
I've had 34k before that was because the battery was low and limiting the cpu speed.
Not sure how a discussion of Nationality started in a phone benchmark thread? Either way, it is unneeded and removed. Please stop.
Back to talking about meaningless numbers now....
Simm is correct - 53k is average for stock unrooted.
I can get this on my wife's phone all day without trying.
Mine did the reverse. On 5.01 I was at 53k. On 5.1 I am at 36k.
El Daddy said:
Not sure how a discussion of Nationality started in a phone benchmark thread? Either way, it is unneeded and removed. Please stop.
Back to talking about meaningless numbers now....
Click to expand...
Click to collapse
Cause that is how they are always.
You must be able to see it from his post, he is assuming this is my first time with technology, and talking down to me like i have lived in a cave somewhere and got this created account date by accident.
Just to add a bit of confirmation (on one side) to this thread, I never got into the 50s on my stock nexus 6 until I removed encryption from the device and switched kernels. My values were consistent with what has been posted. I'm not sure how you can run Antutu wrong on a stock, unrooted, and locked bootloader device, unless you're running it while it's charging or when the battery is dead. I did neither and was scoring in the mid 40s.
For some reason I'm getting scores of around 44,000 in stock 5.1 (rooted). Prior to flashing 5.1 I was getting around 52,000. Antutu's 3D tests aren't looking as smooth as they were before, either.
This is with stock 5.1 kernel.
Edit: OK, how silly of me. Now that I've actually tested it on a fuller battery I'm getting low 50s again.
Related
Just got everything set up after running the manual nooter and did some linpack and quadrant tests to see if there is any improvement over the old Eclair based rom. Here they are for those who care.
Used set CPU to run in performance mode for these tests.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
The linpack mflops looks great. I get 11 on a tweaked out 2.6.29 @800mhz
The quadrant isn't amazing but I think its at least an improvement over the stock 2.6.29 kernel. I get 1800-1900 running my latest test kernel @ 800mhz on Gingerbread(quadrant typically scores less in Gingerbread than Froyo unlike other benchmarks)
Looks like 2.6.32 has some potential.
Holy crap....
This is going to be great once we get kernel source and can undervolt/overlcock this sucka
Performance in general is pretty impressive. I dont know how they managed but at 800MHZ it feels just as smooth as CM7 at 1100. I am REALLY looking forward to getting this overclocked.
Flash is smoother for me in 1.2 than its ever been with any rom regardless of clock speed.
notinterested said:
Performance in general is pretty impressive. I dont know how they managed but at 800MHZ it feels just as smooth as CM7 at 1100. I am REALLY looking forward to getting this overclocked.
Flash is smoother for me in 1.2 than its ever been with any rom regardless of clock speed.
Click to expand...
Click to collapse
There is a post somewhere else that includes an APK for the 1.2 Flash so that CM7 users can replace their current install. Apparently it was optimized specifically for the Nook Color and runs better than the Falsh that's available on the market.
dsf3g said:
There is a post somewhere else that includes an APK for the 1.2 Flash so that CM7 users can replace their current install. Apparently it was optimized specifically for the Nook Color and runs better than the Falsh that's available on the market.
Click to expand...
Click to collapse
My understanding is that it's only true for youtube for some reason, but not for other stuff. I think dal explained it in some thread.
s52m3 said:
My understanding is that it's only true for youtube for some reason, but not for other stuff. I think dal explained it in some thread.
Click to expand...
Click to collapse
It is true for any site that is optimized for flash on mobile. Youtube is so it is easy to use as an example. However any site can be optimised, it is just a matter of if it is or not. Not sure where i read it either maybe the cm7 user thread?
Masterface7 said:
It is true for any site that is optimized for flash on mobile. Youtube is so it is easy to use as an example. However any site can be optimised, it is just a matter of if it is or not. Not sure where i read it either maybe the cm7 user thread?
Click to expand...
Click to collapse
This is what I've been saying. Unfortunately, the flash B&N uses is worse on videos that are not optimized for mobile. But it performs great on those videos that are for mobile. "Mobile" videos only have 40% CPU usage @ 600mhz. Thats pretty impressive for flash, just wish it applied to all videos.
dalingrin said:
This is what I've been saying. Unfortunately, the flash B&N uses is worse on videos that are not optimized for mobile. But it performs great on those videos that are for mobile. "Mobile" videos only have 40% CPU usage @ 600mhz. Thats pretty impressive for flash, just wish it applied to all videos.
Click to expand...
Click to collapse
I have found the exact opposite to be true. I go to a site that has videos not optimized for mobile and on CM7 they are simply not watchable. With the 1.2 flash its not perfect but certainly better than it is in CM7. Which is surprising because when im on CM7 i run 1100MHZ compared to the 800 with the 1.2 flash.
Anyone else have linpack scores on 1.2 they can post?
Linpack score sux
davidukfl said:
Just got everything set up after running the manual nooter and did some linpack and quadrant tests to see if there is any improvement over the old Eclair based rom. Here they are for those who care.
Used set CPU to run in performance mode for these tests.
LINKPACK
score is 6.056 mflops, in 13.84 seconds, normal resolution 5.68, set cpu 925 mhz max/min running cyanogen 7.0.3 will all processes closed.
QUADRANT STANDARD
score is avg roughly 900 and i'm getting anywhere from 3 to 38 fps.
i've owned n1, m.droid, evo and now dinc that all ran cyanogen and never saw such low scores. i don't know if this nook color is just a little premature or that i'm putting too much emphasis on these scores but i believe there's a 2 ghz tablet around the corner with more horsepower. what do you think?
Click to expand...
Click to collapse
Why did you bump this just to scatter crap?
Running CM7 and OC, I'm hitting quad scores of 2400-3000
Try messing with the NC more before you write it off.
I also agree...
If you are getting scores that low, then something is wrong, wrong, wrong.
I am operating on the CM 7.0.3 and have the thing OC'd at 1300.
I get consistent scores in the 2500 - 2900 range...
Lego
2.55 GHz overclock for SHIELD
Please help test this - while there is definitely a measurable improvement in performance, it is kind of inconsistent with the sort of increases I'd expect this clock speed to bring. I have been testing this for a few days, and I'm not quite sure whether this is due to the hardware reporting the wrong clock speeds on the relevant PLLs, an overgressive power management/dvfs quirk, or a bottleneck somewhere. I did have to track down a kernel bug preventing higher clocks that turned out to be an issue with NVIDIA using a signed long where they should have used an unsigned long.
Flashing:
0. Make sure you are on the current OTA.
1. Unlock your bootloader.
2. Test the kernel: fastboot boot boot-roth-oc.img
3. See if it crashes while booting. If it does, don't proceed to the next step, and post a reply below. Before you flash, verify that you're running my kernel by checking for "[email protected]" either in the Kernel Version field in About SHIELD, doing a cat /proc/version (in case you didn't see it crashing), or verifying the maximum CPU speed is 2550 MHz.
4. If all goes well, flash the kernel: fastboot flash boot boot-roth-oc.img
Note: This boot.img is insecure (ro.secure=0) for easier debugging. Userspace apps shouldn't work yet, so this boots at 2550 MHz.
As always, boot/flash this at your own risk, though this kernel doesn't raise voltages much.
Download: http://setcpu.com/boot-roth-oc.img
Relevant patches: https://github.com/coolbho3k/kernel-roth
Picture:
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Mike
Mine just came in the mail. Ill try this out tonight and let you know of it goes
AndroidFileHost said:
Mine just came in the mail. Ill try this out tonight and let you know of it goes
Click to expand...
Click to collapse
DUDE, this is CRAZY! I was getting almost 20 Thousand in Quadrant and 40 Thousand in AnTuTu. I want to see what this can do!
https://twitter.com/wwjoshdew/status/363142322529116160/photo/1
I haven't unlocked mine, but I will soon. Then I'd be more than happy to try out a custom kernel.
OH GOD. What are heat levels? This seems like a good way to melt hands off haha. Has anyone gotten access to the internal fans yet? If I attempt an overclock like this I'd want to have control over them like I do my desktop GPU.
Confirmed working here. Neat! Thanks very much.
Tried your Kernel. So far so good.
And the reason why you see such a inconsistent performance level is because nVidia has different CPU clocks for the Core usage:
1 Thread - Max Clock (1912MHz or with your Kernel)
2 Thread - Max Clock (1810 - 1912MHz)
4 Thread - Max Clock (1420 MHz)
/edit: Forcing a clock helps to prevent from downclocking. Antutu was 42xxx points. PPSSPP runs much better. Great work.
/edit 2: I think Tegra 4 can't statisfy a quadcore A15 with enough bandwidth. I have a CPU benchmark with scales up from 1400MHz to 2520MHz and giving 38% more performance. I see no throttling.
/edit3: Okay, i think you are right. It's not working. There is no scaling going up from 1912Mhz to 2550.
Is there a stock kernel image readily available?
Thanks!
Jesus, I go away from the forums for a week, come back, see overclocked Shield..... WHY????
Damn sweet though
DroidDont said:
Confirmed working here. Neat! Thanks very much.
Click to expand...
Click to collapse
The shield is already a fast device but am still willing to test but first how is the battery life with this kernel because the battery life with stock kernel is amazing but if this kernel shortens the battery life too much then am better off with stock kernel, I will test when there is a stock kernel available just in case :good:
eman2001 said:
The shield is already a fast device but am still willing to test but first how is the battery life with this kernel because the battery life with stock kernel is amazing but if this kernel shortens the battery life too much then am better off with stock kernel, I will test when there is a stock kernel available just in case :good:
Click to expand...
Click to collapse
There is a stock image available; just use the factory boot.img with the above instructions to reflash to it.
coolbho3000 said:
There is a stock image available; just use the factory boot.img with the above instructions to reflash to it.
Click to expand...
Click to collapse
Have you found a fix for the slow performance at 2,55GHz?
2.55ghz...oh my! Galaxy s4 aint got a chance!:highfive:
I've been watching this thread for a couple days, nooo further responses? I take it everyone fried their Shields? Just waiting in anticipation for good news
Sent from my SHIELD using xda app-developers app
How about GPU overclocking?That will be interesting also.
Going to give this a go tonight, would it be possible to provide the stock kernel in case this one is not as stable as it might appear after initial stress testing?
Edit: nvm on stock kernel, just saw that you flash the stock boot.img
Update: booted the kernel and ran a series of benchmarks to stress test. All is stable, will be flashing for daily use.
sontin said:
Have you found a fix for the slow performance at 2,55GHz?
Click to expand...
Click to collapse
Not sure if this has been brought up but when the Shield goes to sleep clock rate drops by almost half on the factory kernel both on the CPU and the GPU.
Could this issue be due to that as well?
Unity engine games and live wallpapers seem to restore it to full speed.
I'll be glad to test this. as soon as i get my Shield, that is.
hopefully by the end of next month.
antutu scores have not noticably gone up with the kernel update. fyi
Flashed this kernel after the new build 59 update today and my wifi is broken.
Could be the update itself since I didn't check before I flashed the kernel. Just throwing out to see if others have the same issue.
I just got my Mi Note Pro last night. I am not familiar with MIUI nor have any experience with flashing / rooting, etc, but I have managed to Flash (I think) the Developer Rom from en.miui. Fast forward to now, I have added the Play Store and the apps that I usually use. I have also deleted most of the Chinese Apps which were pre-installed into the device.
Moving forward, I have tried to benchmark using Antutu. I get around 41-44k~ on Performance mode using the 64-bit version, and around 39k~ on the 32-bit. While the score is way better than any of my previous phones, it falls short of what other users share (around 50k out of the box, some optimized out to 60k). Did I do something wrong, or do I have to do anything, to improve it?
It still doesn't have a sim card right now, if that matters.
P.S. I have already verified the product through Mi Verification, it is authentic.
renrenren34 said:
I just got my Mi Note Pro last night. I am not familiar with MIUI nor have any experience with flashing / rooting, etc, but I have managed to Flash (I think) the Developer Rom from en.miui. Fast forward to now, I have added the Play Store and the apps that I usually use. I have also deleted most of the Chinese Apps which were pre-installed into the device.
Moving forward, I have tried to benchmark using Antutu. I get around 41-44k~ on Performance mode using the 64-bit version, and around 39k~ on the 32-bit. While the score is way better than any of my previous phones, it falls short of what other users share (around 50k out of the box, some optimized out to 60k). Did I do something wrong, or do I have to do anything, to improve it?
It still doesn't have a sim card right now, if that matters.
P.S. I have already verified the product through Mi Verification, it is authentic.
Click to expand...
Click to collapse
More apps you install the worse score you will get but with performance mode u should get 50k+ easy
renrenren34 said:
I just got my Mi Note Pro last night. I am not familiar with MIUI nor have any experience with flashing / rooting, etc, but I have managed to Flash (I think) the Developer Rom from en.miui. Fast forward to now, I have added the Play Store and the apps that I usually use. I have also deleted most of the Chinese Apps which were pre-installed into the device.
Moving forward, I have tried to benchmark using Antutu. I get around 41-44k~ on Performance mode using the 64-bit version, and around 39k~ on the 32-bit. While the score is way better than any of my previous phones, it falls short of what other users share (around 50k out of the box, some optimized out to 60k). Did I do something wrong, or do I have to do anything, to improve it?
It still doesn't have a sim card right now, if that matters.
P.S. I have already verified the product through Mi Verification, it is authentic.
Click to expand...
Click to collapse
My Highest Benchmark score so far,
-using developer rom, rooted, cool temperature, tweak using performance control =D
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
FxJohnnyEng said:
My Highest Benchmark score so far,
-using developer rom, rooted, cool temperature, tweak using performance control =D
Click to expand...
Click to collapse
Hi. Can you please tell how/what to root, as well as what are the things you tweaked? Thank you.
renrenren34 said:
Hi. Can you please tell how/what to root, as well as what are the things you tweaked? Thank you.
Click to expand...
Click to collapse
In order to root, you need to download and update to developer rom. (http://en.miui.com/download-273.html)
Use Mi Verification apps to verify your mi note is genuine or not, once done verify, turn on performance mode
Install performance control apps,
Make sure your mi note is in 16-18 celcius of temperature to cheat the thermal sensor and run antutu, u good to go
My Note pro, just out of the box, scored 61.000 points on antutu 64 bits (more than one moth ago).
But now, after install developer rom, and several updates later, I can not pass 45.000 on the same antutu. Normal is 41.000 points.
But no problem for me. The mobile is usally very fluid. And temperature problems seems that have gone away.... Also battery friendly....
MI Note Pro Custom ROM Low bench Score
I received my MI NOte Pro the other day, with a custom ROM, Scored 4800 Benchmark score An TuTu, which is considerably less, So Custom ROMs gives low bench scores, meaning performance is affected . What do you guys think from your experiences ??
nibinrehman said:
I received my MI NOte Pro the other day, with a custom ROM, Scored 4800 Benchmark score An TuTu, which is considerably less, So Custom ROMs gives low bench scores, meaning performance is affected . What do you guys think from your experiences ??
Click to expand...
Click to collapse
that is not true
i have a custom from xiaomi.eu and around 80000 antutu Benchmark
that is just fabulous , will try flash in that ROM to which you have provided the link, cheers
There are different versions of Antutu and they don't give the same results.
Sent from my MI NOTE Pro using XDA-Developers mobile app
in chinese rom I reachaed 75k but on the developer version only 55k.
Hi all!
See this news article: https://www.xda-developers.com/oneplus-5-benchmark-cheating-reviews/
Long story short - all cores on a test device were bumped to max performance (and stayed there) whenever a benchmark app was run.
This was done by identifying the app name (package name). Very simple and crude implementation.
Graph of performance when running a custom build app with no identifiers:
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Same when running standard apps with regular identifiers present (the phone is now cheating):
**** move OnePlus! Still ordered the phone though :fingers-crossed:
Are you sure this is called cheating? It's really the max performance of the phone, and it doesn't overclock itself. And yeah, it's a bad move to identify those apps, but I wouldn't call it cheating.
The article suggests its cheating, thats there wording Personally i partly agree with you though. But locking at max is not every day behaviour
Have OP learned nothing?!
I can care less about the benchmarks. I care about smoothness and ram management. I hope it is better than my Pixel XL. Time will tell.
Sent from my Pixel XL using XDA-Developers Legacy app
The 3T was one of the best android experiences I have ever had. I'm looking forward to receiving my Oneplus 5
Agree benchmarks are not important in everyday life, or actually not even when choosing phone.
Sent from my NEM-L21 using XDA Labs
i ordered OP5 right after the launch event.. and i couldnt care less for the benchmark scores..
yet, this is cheating..
i dont think most people want to know how the device performs at its peak before burning itself.. instead actual real-life performance is the focus..
it (the "cheat") gives false impression/expectation on performance to those non-techie people with non-normal/in-lab-only usage pattern..
and if you think this is not cheating, why not apply it to all use cases instead of only target at particular benchmark apps, if its such a wonderful behaviour to have?
Battery Life Only Benchmark that Matters
Since the 801 -- which still speeds along my 1+1 plenty fast for me -- I typically skip all the benchmark scores (boring!) and look for the one thing that matters most to me -- the battery life experienced by the reviewer. And according to ArsTech*, the battery life on the 1+5 is stellar!
* https://arstechnica.com/gadgets/2017/06/oneplus-5-review-the-best-sub-500-phone-you-can-buy/2/
Who cares? Does it affect your use of the phone? No. Does it make oneplus look bad? Yes.
Benchmark
Just done a benchmark rated top
Cheating? Is a Volkswagen Diesel inside?
Sorry.. couldn´t resist.
I don't get it, why this is cheating? I want to know what the hardware is capable of. Even if all time max frequency is not daily usable.
oVeRdOsE. said:
I don't get it, why this is cheating? I want to know what the hardware is capable of. Even if all time max frequency is not daily usable.
Click to expand...
Click to collapse
Because every other phone doesn't do it so you get a false benchmark score for OP compared to others.
Batfink33 said:
Because every other phone doesn't do it so you get a false benchmark score for OP compared to others.
Click to expand...
Click to collapse
makes sense.
But this phone still showing hardware benchmarking, I mean they don't edit the software to change the numbers.
If other phone makers don't do this, well I understand.
But in theory with a custom kernel, and cpu governer, it's possible to max out the cpu/gpu and using the phone like so. with 30 min of battery...
oVeRdOsE. said:
makes sense.
But this phone still showing hardware benchmarking, I mean they don't edit the software to change the numbers.
If other phone makers don't do this, well I understand.
But in theory with a custom kernel, and cpu governer, it's possible to max out the cpu/gpu and using the phone like so. with 30 min of battery...
Click to expand...
Click to collapse
The cheating is something other phone makers have done before. Samsung did it a few years ago but stopped when they were found out. One Plus have been found out and continued to do it, Google have even called them out on it and told them to stop. They're maxing the CPU to give an inflated benchmark result and make them top of the benchmark scores.
OnePlus is a bunch of developers that got together and made their own phone right? Soooooo is it a surprise they bumped the scores a little? Who freakin cares it's an awesome phone. (I'd prolly not be as pleased if I had the jelly scrolling issue that is "normal"......)
I'll never understand this "cheating" nonsense...benchmarking a computer or any type of technology's CPU/GPU, etc was done to gauge the potential of that hardware. When people go to sites like 3dmark and look at the leader charts, those PCs are overclocked beyond your wildest imagination using things like liquid nitrogen and dry ice...the video cards are striped of their stock coolers and have exotic water cooling blocks installed.....if you simply wanted to benchmark something at it's basic stock speed, the scores would literally be the same across every product using that CPU/GPU...what's the point? OP sets their phone to go to max performance when benchmarking and it's cheating?? Samsung used to get cherry picked Snapdragons in the past that were able to clock higher than the rated spec safely...is that cheating? If OP clocked the Snapdragon 835 higher than it's rated spec out of the box, would that be considered cheating? Are all the millions of enthusiasts who overclock their PCs to achieve the highest possible score in 3dmark also cheating?
All in all, this article tells us about just how good Oneplus 5 optimizations happen to be. When the Cupertino phone company can drive the heck out of optimizations between each of their software & hardware components we should thank that at least there now is an Android phone team that does the same.
Plus the Benchmarks are not there to imitate real life usage, they are there to see how well software and hardware components are intertwined to pull up a Score based on performance. Real life usage can only be perceived by a real life usage and in that department OnePlus 5 is second to None Literally.
Not really. Any phone manufacturer could easily do the same thing here (increase clock speeds when benchmarks are detected). There's nothing really being "optimized" by OnePlus.
Sent from my ONEPLUS A5000 using Tapatalk
Hello, I bought the phone yesterday (Indian global rom) and after an auto-update for MIUI 12 (12.0.2.0 QFKINXM), installed geekbench 5 and antutu (8.4.4) to try them out. I got substantially lower score in antutu (compared to reviews, YouTube videos) and breakdown showed it's the CPU causing the low score. Upon further benchmarking with Geekbench 5, geekbench 4 and 3dmark it was evident I was getting low score across all the benchmarks where CPU was involved and fantastic score in gpu-only tests (like multiple scenarios in gfxbench).
Then I installed Aida64 and "CPU Monitor" app to further investigate the issue and find possible correlation with temperature etc. The temperature seemed to hover around 55/65 deg while stress testing and 35/40 deg when idle.
I then checked the CPU clockspeed with CPU Monitor app in split screen mode while running Geekbench. I noticed the fastest Kryo gold 485 core (with 2.84 GHz speed never reached the peak). I then ran the Antutu 15-min stress test, and also "CPU Throttle Test" app and "CPU Load Generator" app while keeping the "CPU Monitor" app open in split-screen and noticed the same thing, the most powerful core never reached its peak and mostly runs at idle clock speed of ~800 MHz even when benchmarking. Further, even when that core boosts, it seems capped at 2.4 GHz which is thehe maximum clockspeed of other 3 Kryo gold 485 cores. I know 2.84 GHz isn't supposed to be a sustained speed but the fact that it never reached it and gets capped at the same speed as the other 3 cores implies something is wrong here.
Also I don't really know if the poor score is solely caused by the poor performance of that 2.84 GHz core but I do know from observation that the core isn't running as intended as I explained above.
Now if it's a software issue then i wouldn't worry much as it'll possibly get fixed in next updates or by installing custom rom. But I'm worried if the issue is related to hardware itself (problematic CPU core or heatsink issue).
So I'm writing to know if any of you guys have had such issue or what do u think it is caused from or possible fix. Also
Any constructive suggestion will be highly welcome.
Geekbench 5 CPU score: single core ~600, multi-core ~2200 to 2300
Antutu 8 CPU score: ~90k (sometimes much less)
I have an mi 9t pro with the global miui 12.0.1 and it's similar to what you're saying, so it's not a hardware issue, it's the kernel, it seems that Xiaomi it's very conservative about the battery and throttles the prime core to 2.4ghz max
The only thing that have worked for me was to change the kernel to another one, I'm trying Yuki kernel and it's very good, also there's a new kernel, quantic, and it's very good too, the battery doesn't seems to take a big hit, it's fine for me, like the original kernel, sorry if my english it's not perfect, I'm from Mexico
Omarrrc1 said:
I have an mi 9t pro with the global miui 12.0.1 and it's similar to what you're saying, so it's not a hardware issue, it's the kernel, it seems that Xiaomi it's very conservative about the battery and throttles the prime core to 2.4ghz max
The only thing that have worked for me was to change the kernel to another one, I'm trying Yuki kernel and it's very good, also there's a new kernel, quantic, and it's very good too, the battery doesn't seems to take a big hit, it's fine for me, like the original kernel, sorry if my english it's not perfect, I'm from Mexico
Click to expand...
Click to collapse
Hello, thanks for your reply and your English is awesome.
Now regarding the kernel change, all the reviewers and YouTubers benchmarked it in stock and most of them didn't face such problem. Their score seems reasonably similar to each other.
Also do u remember if your CPU benchmark score was substantially lower than standard before changing kernel?
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
60c is where CPU starts to throttle, prime core never runs at max freq for more than 3-4 mins regardless of temperature to save battery and wear from the heat it's producing on stock kernel.
Also never forget how much a impact ambient temps play a part in phone performance, I've put my phone in the freezer for 5-10 mins and saw a 500 multi core score although that was on a warm 40c Aus summer day lol
As others have suggested try custom rom if you really need max performance although it is completely unnecessary and a waste of time pursuing max performance, take it from someone who's been there.
The result after a month of trying every combination of roms and kernels lol
hoopsnake said:
View attachment 507762560c is where CPU starts to throttle, prime core never runs at max freq for more than 3-4 mins regardless of temperature to save battery and wear from the heat it's producing on stock kernel.
Also never forget how much a impact ambient temps play a part in phone performance, I've put my phone in the freezer for 5-10 mins and saw a 500 multi core score although that was on a warm 40c Aus summer day lol
As others have suggested try custom rom if you really need max performance although it is completely unnecessary and a waste of time pursuing max performance, take it from someone who's been there.
The result after a month of trying every combination of roms and kernels lol
Click to expand...
Click to collapse
Thanks for your reply. However my issue is it never reached 2.84 GHz and is artificially capped at 2.4 ghz. I'm not talking about the core being constant at 2.84 GHz.
Also I'm not necessarily looking for absolute max performance as I'm not a gamer but just wanna investigate the issue to make sure it's not hardware issue.
Btw if you've come back to stock rom/kernel, can you pls test and mention the geekbench 5, antutu 8, 3dmark scores for K20 pro?