Database Users

CPU under clock

Here's something I noticed while testing my phone in smart bench. I dialed my CPU down to 816mhz and 912mhz and look at these scores
{
"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"
}
Then I set it back to 1ghz and look at the scores.
Doesn't make sense. We need 2.3(2.4) to truely run out dual core at its best.
But thats just my 2 cents
Sent from my Atrix

drock212 said:
Here's something I noticed while testing my phone in smart bench. I dialed my CPU down to 816mhz and 912mhz and look at these scores
View attachment 591817
View attachment 591818
Then I set it back to 1ghz and look at the scores.
View attachment 591819
Doesn't make sense. We need 2.3(2.4) to truely run out dual core at its best.
But thats just my 2 cents
Sent from my Atrix
Click to expand...
Click to collapse
strange..!!

I find I get better scores on smartbench when I have tons of apps running and less free ram, how odd is that???

These benchmark tools are bogus. In the end what really matters is your experience, do you feel it fast? So it's fast...

drock212 said:
Here's something I noticed while testing my phone in smart bench. I dialed my CPU down to 816mhz and 912mhz and look at these scores
View attachment 591817
View attachment 591818
Then I set it back to 1ghz and look at the scores.
View attachment 591819
Doesn't make sense. We need 2.3(2.4) to truely run out dual core at its best.
But thats just my 2 cents
Sent from my Atrix
Click to expand...
Click to collapse
Are you running single tests or multiple?

I've noticed the same results and thought it quite strange and I ran multiple tests. Its interesting also that the more tests you run the better the scores get until a certain amount of times, around 6 or so on mine, that the scores drop dramatically to a lower score than the first run score.
Sent from my MB860 using XDA App

Benchmarks need to be repeatable for them to be reliable. Any outfit worth its salt does benchmarks on a freshly installed system with all unnecessary services disabled and nothing running etc.
Otherwise you're going to get some unseen interaction that will screw with your benchmark result.

Lieutenant_Dan said:
Are you running single tests or multiple?
Click to expand...
Click to collapse
Ran multiple test to make sure the results were consistent and always got the same results
But thats just my 2 cents
Sent from my Atrix

Remember all kinds of stuff comes into factor. Heat from multiple tests, heat from charging, heat from pocket, rebooting before test, less apps being used, data turned off... and tons more. You cant really compare tests back too back, you must run tons of them with the same variables and take a median not a mean.

gHz show-off

Hey all,
Just wanted to show you how great the Nexus S is!! "img.tapatalk.com/5658d096-434f-704f.jpg" (just copy everything between " " to your toolbar) its running at a stunning 1.6gHz!!! It became a bit too overclocked sometimes, so I stepped back to 1.3gHz... not bad either, but still less. Show how far you can push your Nexus S (4G) in this thread by posting a screenshot that shows on how much gHz you Nexus S can run.
IF YOUR PHONE ISNT WORKING ANYMORE, BECAUSE YOU PUSHED IT TOO FAR. ITS YOUR OWN MISTAKE. IF YOU POINT AT ME BECAUSE YOU THINK ITS MY FAULT... I WILL BE LAUGHING AT YOU.
Greetzz, Jojoost

jojoost said:
Hey all,
{
"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"
}
Click to expand...
Click to collapse
fixed......

this 1.7ghz was from almost a year ago, it ran for almost a whopping 30 seconds before freezing.
but 1.6ghz ran stable, just battery life was terrible

What's the default clock speed? I run my 1.2 at 1782...
Sent from my ICS Splashed using xda premium

Can't go past 1.2 without rebooting on intense CPU usage
Heck I can't even run Trinity kernel.
Sent from my Nexus S 4G using XDA App. Developer of brickROM, and OP of XDA Thread of The Year 2011.

KrnlPanic said:
What's the default clock speed? I run my 1.2 at 1782...
Sent from my ICS Splashed using xda premium
Click to expand...
Click to collapse
Nevermind me, didn't realize I wasn't in my subscribed forums list. Sorry.
Sent from my ICS Splashed using xda premium

removed becasue CPU not score comparison

cpu speed isnt everything. after a certain cpu speed, benchmarks will start getting lower/worse. performance overall will become worse if the cpu speed is set to high. personally, my phones sweetspot is between 1.44 and 1.54ghz. above that, performance starts to drop.

simms22 said:
its true, cpu speed isnt everything. after a certain cpu speed, benchmarks will start getting lower/worse. performance overall will become worse if the cpu speed is set to high. personally, my phones sweetspot is between 1.44 and 1.54ghz. above that, performance starts to drop.
Click to expand...
Click to collapse
are you lowering your FSB to achieve higher speeds? dont see overall system slowdown unless you are down clocking your FSB to achieve higher CPU speeds? i mean his CPU is allot higher than mine
IDK im more familiar with desktop CPU not mobile think i might get what you're saying maybe an overheat issue and downclcok for protection..idk

demo23019 said:
removed becasue CPU not score comparison
Click to expand...
Click to collapse
I agree.
And if your phone is getter hot. I'm very sure it is. Then you could be damaging it and that could greatly decrease the life spand of your internal components
Sent from my Nexus S 4G using XDA App. Developer of brickROM, and OP of XDA Thread of The Year 2011.

very funny... tried to OC up to 1540Mhz... the phone start to vibrate (like a female friend) and freezed.... now rebooted... never play again with so high OC

Overclocked bench results

State your rom, current kernel, frequencies, and voltages
Kernel elementalx
Cpu 2052
Gpu 450
Rom arhd 9.4
Volts 1227
Quadrant
{
"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"
}
Antutu
If you would luke to.flash this beautiful kernel it can be had here http://forum.xda-developers.com/showthread.php?t=2249774
Sent from my HTC One using xda premium

What's your bin number?
Average CPU Temperature?
Governor?

Govenor ondemand
Average temp around ,44c with normal use
Pvs 3
Sent from my HTC One using xda premium

Thanks! Great score then! That score is almost the same as the exynos octa core of S4. Although the antutu benchmark got lower score compared to quadrant. Which test did you perform first?

Riyal said:
Thanks! Great score then! That score is almost the same as the exynos octa core of S4. Although the antutu benchmark got lower score compared to quadrant. Which test did you perform first?
Click to expand...
Click to collapse
The quadrant score is definitely higher than my S4 octa (13K) but lower than my antutu (32K). No overclocking, just using stock kernel. That being said, with 4.2.2 on my One, there is definite overall improvement especially in terms of GPU performance.

I think the octa score is due to the powervr its just way more optimized then the adreno.
Sent from my HTC One using xda premium

KindaJackedForANerd said:
I think the octa score is due to the powervr its just way more optimized then the adreno.
Sent from my HTC One using xda premium
Click to expand...
Click to collapse
Ohhh.. I thought it was a Mali GPU? If you're correct then yeah! PowerVR can already be considered a veteran when it comes to mobile GPU. It was the first ever mobile GPU in a phone. It was used in the phones nokia n95 & n82 in the year 2007-2008

"sleeping" CPU cores...anyone else notice this?

Last night I was just messing around and looking at some things trying to figure out the issues with choppy scrolling on our phones since the update and was checking out CPU usage and noticed something interesting...
I saw that two of the CPU cores were being reported as "sleeping" with no activity registered on them...even after running a benchmark there was no activity shown...
A reboot "fixed" the issue and all 4 cores were reporting activity...but then after going to bed, and waking up this morning it is back to the same...I attached a couple of screenshots to show what I'm seeing. You can see in the second picture the "history" clearly shows where the two cores went to sleep...
{
"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"
}
Sent from my XT1650 using Tapatalk

Isn't the 820 a big/little quad core? You'll only have 2 active at one time.

fr4nk1yn said:
Isn't the 820 a big/little quad core? You'll only have 2 active at one time.
Click to expand...
Click to collapse
You can clearly see here...that after rebooting the phone...all 4 cores will register as active for a time...
Or here
Sent from my XT1650 using Tapatalk

I was wrong. The SD820 is not big/little it's asymmetrical.
With that said I'm running Geekbench in splitscreen and the highspeed cores are not showing as active.
Going to reboot and try again. If the score is significantly higher.

fr4nk1yn said:
I was wrong. The SD820 is not big/little it's asymmetrical.
With that said I'm running Geekbench in splitscreen and the highspeed cores are not showing as active.
Going to reboot and try again. If the score is significantly higher.
Click to expand...
Click to collapse
My AnTuTu scores don't seem to be significantly lower when the cores are "sleeping" maybe 1000 points give or take (on total scores of 150,000+ so not significant) but it's still odd for all of these apps to be having such a hard time accurately reporting the activity on such a popular chipset ya know?
Sent from my XT1650 using Tapatalk

I have noticed the same phenomenon on Xperia XZ (Snapdragon 820 also) after Nougat update. Have not done any benchmarking though.

Whats the path to get to the screenshots you posted???? Or is it an app????

You can also see the utilization is much less. Maybe it's automatic that they wake or sleep based on demand?

I've noticed this too. Have opened up games and many other apps but cant seem to get those 2 cores "awake".

Here's an interesting app I stumbled across: https://play.google.com/store/apps/details?id=com.mediatek.multicoreobserver
With the last update its definitely not turning off any cores. All 4 are staying active and scaling at all times.

Is This the Highest Benchmark Score the Note 3 Will Get?

I want to preface this with the fact that I have run several roms over the course of my time spent in rooting and flashing. Out of all of the roms, CM12.1 is the most stable and currently runs the fastest. I'm wondering if the Note 7 or S7/S7E ports will be able to handle overclocking because every time I have tried, the device will reboot. If I can't get these types of roms to be stable and fast, I won't be able to test them with Antutu benchmark. That said, I'm wondering if there is anyone who has a higher score than this. If so, what rom did you use and how stable is it?
I should note that I am currently using these specs:
-ROM: CM12.1
-Kernel: IdleKernel
-Screen Size: 2560 X 1440 (My phone seems to handle this incredibly well, though I can't quite get it up to 4K without having problems.)
{
"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 Killing Joke said:
If I can't get these types of roms to be stable and fast, I won't be able to test them with Antutu benchmark.
Click to expand...
Click to collapse
Yeah, this joke is really killing me...

Dejan Sathanas said:
Yeah, this joke is really killing me...
Click to expand...
Click to collapse
The questions were posed because I want to know if anyone has had better luck than I did. I'm merely curious as to what the limits are of this type of phone.

The Killing Joke said:
The questions were posed because I want to know if anyone has had better luck than I did. I'm merely curious as to what the limits are of this type of phone.
Click to expand...
Click to collapse
This score means nothing if you do not specify used cpu governor, hotplug driver, i/o scheduler and o.c value.
With an S7E port (default settings, no optimizations and no OC), i can get +62K and a 3D score higher than yours.

zlazaar said:
This score means nothing if you do not specify used cpu governor, hotplug driver, i/o scheduler and o.c value.
With an S7E port (default settings, no optimizations and no OC), i can get +62K and a 3D score higher than yours.
Click to expand...
Click to collapse
As expected, most N7/S7E ports will have a higher 3D value than CM. The govenor I used was interactive. Overclocked to 2.76 ghz, gpu was also overclocked to 450 mhz. I chose not to mess with the i/o scheduler. Hotplugging was turned off.

The Killing Joke said:
As expected, most N7/S7E ports will have a higher 3D value than CM. The govenor I used was interactive. Overclocked to 2.76 ghz, gpu was also overclocked to 450 mhz. I chose not to mess with the i/o scheduler. Hotplugging was turned off.
Click to expand...
Click to collapse
Well, if you are satisfied with your rom so keep it.
Anyway, in raw speed Touchwiz can't compete with CM/AOSP.
A little tip for higher score : use Fiops schedular with read ahead buffer = 256kb