|
Who's Online
|
There are currently 2568 online users. [Complete List]
» 11 Member(s) | 2557 Guest(s)
|
|
|
|
Thread Rating:
- 0 Vote(s) - 0 Average
- 1
- 2
- 3
- 4
- 5
|
CPU minimum clock rate
|
Posts: 104
Threads: 24
Joined: 2013-07-20
Points:
8
Reputation: +81
What are the advantages/disadvantages to raising the CPU minimum clock rate?
I hear raising it drains the battery quicker, but I don't care about that. My CPU drains power so much less than my screen that it doesn't matter much comparatively.
Is there a serious downside in terms of system stability? Can it damage your CPU? I haven't had problems so far running CPUs at minimum equal to maximum.
Is there a performance gain at all when the CPU minimum frequency is raised?
You like this post
Posts: 3,182
Threads: 555
Joined: 2013-02-10
Points:
21,230
Reputation: +3636
It won't hurt your phone or cause any sort of instability.
As far as benefits, you may gain some slight performance improvements, but it could be below the threshold of noticeability.
Make sure your governor is set to hotplug.
You like this post
Two roads diverged in a yellow wood,
And sorry I could not travel both
And be one traveler, long I stood
And looked down one as far as I could
To where it bent in the undergrowth
Posts: 104
Threads: 24
Joined: 2013-07-20
Points:
8
Reputation: +81
(2013-08-01, 16:30)Gizbeat Wrote: It won't hurt your phone or cause any sort of instability.
As far as benefits, you may gain some slight performance improvements, but it could be below the threshold of noticeability.
Make sure your governor is set to hotplug.
Yeah, I noticed from my antutu results that putting the governor on something other than Hotplug, including OnDemand and Performance, puts half my CPU out of business and makes the result drop from 14000+ to less than 7000. Hotplug with minimum frequency at 1.5Ghz has the same effect, but minimum at 1.2Ghz works fine without noticeable performance gains (on benchmark or otherwise).
You and GizBeat like this post
Posts: 104
Threads: 24
Joined: 2013-07-20
Points:
8
Reputation: +81
2013-08-01, 23:17
(This post was last modified: 2013-08-01, 23:18 by Rdmkr.)
I've been wondering what the implications of this strange result are, though. Is the chip advertised as 1.5Ghz quad-core really a 1.5 x 1.5 x 1.2 x 1.2 chip or something?
oh well. I'm pretty happy with the performance regardless.
You like this post
Posts: 3,182
Threads: 555
Joined: 2013-02-10
Points:
21,230
Reputation: +3636
(2013-08-01, 23:17)Rdmkr Wrote: I've been wondering what the implications of this strange result are, though. Is the chip advertised as 1.5Ghz quad-core really a 1.5 x 1.5 x 1.2 x 1.2 chip or something?
oh well. I'm pretty happy with the performance regardless.
It is an interesting result that scores are less when the min frequency is set at 1.5.
It should be 1.5x4, not 1.5x2/1.2x2
The results could be from several things. Perhaps a conflict between the governing app and Antutu.
Are all of the scores cut in half? (eg, GPU,SD Write,CPU etc...)
How does the CPU score when not governed by the app vs. governed max at 1.2 vs governed max at 1.5?
Two roads diverged in a yellow wood,
And sorry I could not travel both
And be one traveler, long I stood
And looked down one as far as I could
To where it bent in the undergrowth
Posts: 104
Threads: 24
Joined: 2013-07-20
Points:
8
Reputation: +81
(2013-08-01, 23:47)Gizbeat Wrote: (2013-08-01, 23:17)Rdmkr Wrote: I've been wondering what the implications of this strange result are, though. Is the chip advertised as 1.5Ghz quad-core really a 1.5 x 1.5 x 1.2 x 1.2 chip or something?
oh well. I'm pretty happy with the performance regardless.
It is an interesting result that scores are less when the min frequency is set at 1.5.
It should be 1.5x4, not 1.5x2/1.2x2
The results could be from several things. Perhaps a conflict between the governing app and Antutu.
Are all of the scores cut in half? (eg, GPU,SD Write,CPU etc...)
How does the CPU score when not governed by the app vs. governed max at 1.2 vs governed max at 1.5?
GPU result looked normal. I think it was just RAM and CPU Int & Float. About 1000 on each instead of the normal 3000 - 4000... So I guess that suggests 3 regular CPU cores were offline..? Or something else entirely.
You like this post
Posts: 3,182
Threads: 555
Joined: 2013-02-10
Points:
21,230
Reputation: +3636
(2013-08-02, 00:39)Rdmkr Wrote: (2013-08-01, 23:47)Gizbeat Wrote: (2013-08-01, 23:17)Rdmkr Wrote: I've been wondering what the implications of this strange result are, though. Is the chip advertised as 1.5Ghz quad-core really a 1.5 x 1.5 x 1.2 x 1.2 chip or something?
oh well. I'm pretty happy with the performance regardless.
It is an interesting result that scores are less when the min frequency is set at 1.5.
It should be 1.5x4, not 1.5x2/1.2x2
The results could be from several things. Perhaps a conflict between the governing app and Antutu.
Are all of the scores cut in half? (eg, GPU,SD Write,CPU etc...)
How does the CPU score when not governed by the app vs. governed max at 1.2 vs governed max at 1.5?
GPU result looked normal. I think it was just RAM and CPU Int & Float. About 1000 on each instead of the normal 3000 - 4000... So I guess that suggests 3 regular CPU cores were offline..? Or something else entirely. It might be as I suggested, that the CPU being governed at 1.5 by the app clashes with the way Antutu tests. It's strange it's not happening when being governed at 1.2, but the clash is still a possibility.
Download "Micro CPU Monitor". Check MCM settings and make sure it's set to monitor all CPUs individually. It will run in the corner and show you their use in real time while the test runs.
You like this post
Two roads diverged in a yellow wood,
And sorry I could not travel both
And be one traveler, long I stood
And looked down one as far as I could
To where it bent in the undergrowth
Posts: 104
Threads: 24
Joined: 2013-07-20
Points:
8
Reputation: +81
2013-08-02, 00:50
(This post was last modified: 2013-08-02, 00:59 by Rdmkr.)
(2013-08-02, 00:47)Gizbeat Wrote: (2013-08-02, 00:39)Rdmkr Wrote: (2013-08-01, 23:47)Gizbeat Wrote: It is an interesting result that scores are less when the min frequency is set at 1.5.
It should be 1.5x4, not 1.5x2/1.2x2
The results could be from several things. Perhaps a conflict between the governing app and Antutu.
Are all of the scores cut in half? (eg, GPU,SD Write,CPU etc...)
How does the CPU score when not governed by the app vs. governed max at 1.2 vs governed max at 1.5?
GPU result looked normal. I think it was just RAM and CPU Int & Float. About 1000 on each instead of the normal 3000 - 4000... So I guess that suggests 3 regular CPU cores were offline..? Or something else entirely. It might be as I suggested, that the CPU being governed at 1.5 by the app clashes with the way Antutu tests. It's strange it's not happening when being governed at 1.2, but the clash is still a possibility.
Download "Micro CPU Monitor". Check MCM settings and make sure it's set to monitor all CPUs individually. It will run in the corner and show you their use in real time while the test runs.
Ok, I'll try that. Another thing though: I ran antutu again on Hotplug + minimum at 1.5 Ghz and got a normal 14000+ result. I'd have sworn this went wrong the last time. It's possible there is something independent of the CPU control that's interfering with the result... I'll run it on the Performance governor again to be sure...
Result on Performance, 1.5Ghz minimum: 6829
RAM: 987
Int: 1076
Float: 791
2D: 684
3D: 2595
So about 1/4 of normal on the first 3, normal on the last 2.
Not 100% sure yet, but I'm guessing it's mostly just the Hotplug governor that is a must.
You like this post
Posts: 3,182
Threads: 555
Joined: 2013-02-10
Points:
21,230
Reputation: +3636
(2013-08-02, 00:50)Rdmkr Wrote: (2013-08-02, 00:47)Gizbeat Wrote: (2013-08-02, 00:39)Rdmkr Wrote: GPU result looked normal. I think it was just RAM and CPU Int & Float. About 1000 on each instead of the normal 3000 - 4000... So I guess that suggests 3 regular CPU cores were offline..? Or something else entirely. It might be as I suggested, that the CPU being governed at 1.5 by the app clashes with the way Antutu tests. It's strange it's not happening when being governed at 1.2, but the clash is still a possibility.
Download "Micro CPU Monitor". Check MCM settings and make sure it's set to monitor all CPUs individually. It will run in the corner and show you their use in real time while the test runs.
Ok, I'll try that. Another thing though: I ran antutu again on Hotplug + minimum at 1.5 Ghz and got a normal 14000+ result. I'd have sworn this went wrong the last time. It's possible there is something independent of the CPU control that's interfering with the result... I'll run it on the Performance governor again to be sure... 
It's also possible your governor was still on a governor other than hotplug when you ran the 1.5 test the first time through. Not saying that's what it was... Just saying it's possible.
You like this post
Two roads diverged in a yellow wood,
And sorry I could not travel both
And be one traveler, long I stood
And looked down one as far as I could
To where it bent in the undergrowth
Posts: 104
Threads: 24
Joined: 2013-07-20
Points:
8
Reputation: +81
2013-08-02, 01:23
(This post was last modified: 2013-08-02, 01:30 by Rdmkr.)
Ok, ran it on hotplug 1.5Ghz minimum again and got slightly below 14000, which qualifies as normal. So at this point I'm thinking it is in fact the non-hotplug governors that are the problem.
For the record, though, my Freelander PD800 Rockchip RK3188 tablet has no trouble getting 16000+ results on Antutu with Ondemand or Performance turned on using the same app (No Frills CPU Control). So it's not 100% an app incompatibility problem independent of CPU type.
|
Users browsing this thread: 1 Guest(s)
|
|
![[-]](https://www.chinaphonearena.com/forum/images/collapse.png "[-]")
