[Kernel] [2.6.38.8 (CM7)] HAVS/SVS + 1152MHz (09/17/11)

[cdmove]

OP,

thanks for doing this and I'd like to request a .34 uv + oc kernel without the audio boost tweak!!

Thanks again1!

 

[dictionary]

OP,

thanks for doing this and I'd like to request a .34 uv + oc kernel without the audio boost tweak!!

Thanks again1!

Be a bit more specific. 800mV? 925mV? BFS? CFS?

Just trying to help ya.

 

[wildmonks]

OP,

thanks for doing this and I'd like to request a .34 uv + oc kernel without the audio boost tweak!!

Thanks again1!

Be a bit more specific. 800mV? 925mV? BFS? CFS?

Just trying to help ya.

I'd gladly make one for you! But yah.. please specify. Unfortunatly it takes me almost 50 minutes to compile each version... I dont have a fast computer

 

[daveid]

Here yah go:
2.6.33.5_AVS-925mV_CFS_NoAudBoost_v2_060210.zip

I think all those years of listening to loud music has affected my hearing so I need the audio boost

I tested it out and it seems audio levels are back to standard (noticeably lower) but let me know if it doesn't work or if there are any other problems with it.

Update: Doh! I found a bug that I introduced into this no audio boost version. Seems that I broke music playback O_O. I fixed it and updated the above link. Sorry to those that download the other version -_- . . .

Thank you, flashing soon, I look forward to being able to hear callers without distortion and discomfort.

 

[nazrin313]

Interesting observation you all are getting. Must be the combination of BFS and OC? But I would not be surprised if battery life suffers a little... But very cool nonetheless

I flashed the CFS 925 froyo version

Just checked my phone...slept for 5 or so hours and the battery is down 94%.....not so bad I guess...will be using it (my daily usage pattern) today

 

[irishrally]

Did anyone else see that the "clear" button in the notifications pull down is gone? I'm using 925CFS? I'd rather it be there.

EDIT: The "clear" button is there sometimes, but you can't clear usb debugging, usb mount or voicemail.

 

[eViL D:]

Trying to compile, anyone know of this error?

No rule to make target `net/ipv4/netfilter/ipt_ecn.o', needed by `net/ipv4/netfilter/built-in.o'. Stop.

 

[cdmove]

is the sd remount still a problem with .34?

 

[Daswolven]

I got a rather significant performance boost in benchmarks on the 800 BFS Froyo over pershoots OC. Anyone care to enlighten me on why this is?

 

[nazrin313]

Found a bug...now my phone keeps on randomly rebooting...happened 3 times already today....

Battery life is still good..7 hours of moderate and heave use...(approx: display now on for a good one hour - random reboots resetted my counter) and im still on 76%

 

[irishrally]

Just got 1717 on the Quadrant benchmark, higher than any other kernel I've tried. Using CFS 925mV.

 

[wildmonks]

Did anyone else see that the "clear" button in the notifications pull down is gone? I'm using 925CFS? I'd rather it be there.

EDIT: The "clear" button is there sometimes, but you can't clear usb debugging, usb mount or voicemail.

The USB debugging and USB mount notification always stayed there for me when I have my Nexus connected to the computer but they disappear when its not connected. I thought that was normal behavior on stock google roms. I don't know about that voicemail notification though... Have you tried clearing your voicemail out to see if it goes away? (I use google's voice service so I don't see the standard voicemail notification)

I got a rather significant performance boost in benchmarks on the 800 BFS Froyo over pershoots OC. Anyone care to enlighten me on why this is?

I honestly don't really know. My only guess is because of the BFS as that is the only difference. The overclock code itself should be nearly identical. It could also be the difference in compilers.

Found a bug...now my phone keeps on randomly rebooting...happened 3 times already today....

Battery life is still good..7 hours of moderate and heave use...(approx: display now on for a good one hour - random reboots resetted my counter) and im still on 76%

That could be caused by the cpu not receiving enough voltage. I can try increasing voltage to 1300mV. Currently its set to max 1275mV which matches pershoot max. The reboots can also be caused by the BFS itself. Like what daveid said, BFS still has a long way to go which is why you'll probably never see it on any of Google's official builds.

 

[wildmonks]

Trying to compile, anyone know of this error?

No rule to make target `net/ipv4/netfilter/ipt_ecn.o', needed by `net/ipv4/netfilter/built-in.o'. Stop.

What source are you using? Cyanogen's or intersectRaven's? If your using intersectRaven's try using these configs.

If your still having errors, maybe its the compiler you are using. I used to get a lot of similar errors when I tried using codeSourcery's compiler. Finally had to give up and use the standard gcc compilers (currently using gcc 4.5).

 

[wildmonks]

Just got 1717 on the Quadrant benchmark, higher than any other kernel I've tried. Using CFS 925mV.

Wow, that's really awesome! Uh.. that kinda blows my theory of BFS having to do with the improvement out of the water....

 

[zizoty]

Gotta try this kernel since i only get 1414 from pershoot kernel. 1717 is a crazy ass number.
Edit: Just tried the cfs 925 and clocked 1490. I don't know how you get 1717. What Rom are you using?

 

[irishrally]

Gotta try this kernel since i only get 1414 from pershoot kernel. 1717 is a crazy ass number.
Just tried the cfs 925 and clocked 1490. I don't know how you get 1717. What Rom are you using?

Paul's Froyo release with CM Quick Rooter and the 925CFS. I ran it 4 or 5 times and averaged around 1600.

There is a lot more to a kernel than doing well on benchmarks though, but it would be nice to understand what causes the increase, if it turns out to be present in many N1s.

EDIT: Try setting min and max to 1113mhz in setcpu and also set to performance, not on demand. I was conncected to wifi, not sure if that requires less cpu resources than 3G data running in the background.

 

[gn00my]

awesome kernels thanks for your work


oh btw , got a score of 1434 on quadrant with 2.6.34_AVS-925mV_CFS_1113Mhz_060110 using froyo

 

[OrganizedFellow]

I've done some benchmark testing.
Device is my Nexus One with FRF50 with radio from Modaco.

This is my daily use device, so nothing was done to optimize anything (with SetCPU as the only exception).
I thought about doing this with a fresh Froyo install, but I wanted to get "real world"/daily use values, with a device running widgets, etc.
I have 33-35 apps installed, a few widgets, and with average daily use (9am til atleast 11pm).
I plug my phone into charger close to 12am each night when it has about 10-20% battery.
Plenty of SMS/MMS throughout the day, 2-3 hours of web browsing, no YouTube, lots of Facebook updates via text & replies via m.facebook.com, Gmail, no Twitter, no games except for Solitaire
----- ----- ----- ----- ----- ----- ----- ----- ----- -----
A few days ago, I downloaded and installed all four of the kernels offered by intersectRaven's Kernel - http://bit.ly/d4rk1R, ofcourse, one by one.
Tonight I tested just one from Hybrid AVS + 1113Mhz Overclock - http://bit.ly/cOvFEe, the CFS:925 version. As that one seemed the most recommended/used.

I used the following apps:
BenchmarkPi http://www.cyrket.com/p/android/gr.androiddev.BenchmarkPi/
Linpack http://www.cyrket.com/p/android/com.greenecomputing.linpack/
Quadrant http://www.cyrket.com/p/android/com.aurorasoftworks.quadrant.ui.standard/

All tests performed while charging and at a minimum of 90%.

I did three tests with each kernel (15 tests).
Test#1: reboot, wait 60 seconds, run Test#1.
Test#2: run immediately after exiting Test#1, run Test#2.
Test#3: kill all tasks, wait 5 seconds, run Test#3.
I did the above for each of the five kernels, with each of the three apps.
The numbers shared below are the averages of Test#1, Test#2 and Test#3, per app, per kernel.

(Almost forgot, SetCPU is set at max/max, and Governor set to performance for all tests.)

These are my results:
intersectRaven's Kernel BFS:925mV BenchmarkPi 1,373.75
intersectRaven's Kernel BFS:800mV BenchmarkPi 1,287.75
intersectRaven's Kernel CFS:925mV BenchmarkPi 1,399.25
intersectRaven's Kernel CFS:800mV BenchmarkPi 1,247.50
HybridAVS+1113MhzOver...CFS:925mV BenchmarkPi 734.50

intersectRaven's Kernel BFS:925mV Linpack 17.58
intersectRaven's Kernel BFS:800mV Linpack 15.83
intersectRaven's Kernel CFS:925mV Linpack 16.55
intersectRaven's Kernel CFS:800mV Linpack 15.54
HybridAVS+1113MhzOver...CFS:925mV Linpack 34.60

intersectRaven's Kernel BFS:925mV Quadrant 1,480.33
intersectRaven's Kernel BFS:800mV Quadrant 1,410.33
intersectRaven's Kernel CFS:925mV Quadrant 1,363.33
intersectRaven's Kernel CFS:800mV Quadrant 1,422.33
HybridAVS+1113MhzOver...CFS:925mV Quadrant 1,391.67
----- ----- ----- ----- ----- ----- ----- ----- ----- -----
Obviously the overclocked kernel kicks ass in nearly all the tests.
It doubled the average scores in BenchmarkPi & Linpack.

MY QUESTION IS:
But why does it get such similar scores as the others when benchmarked with Quadrant


***MY SILLY DISCLAIMER***
There are many factors not mentioned here (running apps, how many beers I've had, etc.
et cetera yadda yadda yaaa ... ya know?

 

[flak0]

Paul's Froyo release with CM Quick Rooter and the 925CFS. I ran it 4 or 5 times and averaged around 1600.

There is a lot more to a kernel than doing well on benchmarks though, but it would be nice to understand what causes the increase, if it turns out to be present in many N1s.

EDIT: Try setting min and max to 1113mhz in setcpu and also set to performance, not on demand. I was conncected to wifi, not sure if that requires less cpu resources than 3G data running in the background.

I can confirm this also. I'm getting over 1600 in quadrant. proof below

got 60+fps during the planets sequence in quadrant.

this along with the picture above was posted from my nexus using opera mini 5.

 

[irishrally]

I've done some benchmark testing.
Device is my Nexus One with FRF50 with radio from Modaco.

This is my daily use device, so nothing was done to optimize anything (with SetCPU as the only exception).
I thought about doing this with a fresh Froyo install, but I wanted to get "real world"/daily use values, with a device running widgets, etc.
I have 33-35 apps installed, a few widgets, and with average daily use (9am til atleast 11pm).
I plug my phone into charger close to 12am each night when it has about 10-20% battery.
Plenty of SMS/MMS throughout the day, 2-3 hours of web browsing, no YouTube, lots of Facebook updates via text & replies via m.facebook.com, Gmail, no Twitter, no games except for Solitaire
----- ----- ----- ----- ----- ----- ----- ----- ----- -----
A few days ago, I downloaded and installed all four of the kernels offered by intersectRaven's Kernel - http://bit.ly/d4rk1R, ofcourse, one by one.
Tonight I tested just one from Hybrid AVS + 1113Mhz Overclock - http://bit.ly/cOvFEe, the CFS:925 version. As that one seemed the most recommended/used.

I used the following apps:
BenchmarkPi http://www.cyrket.com/p/android/gr.androiddev.BenchmarkPi/
Linpack http://www.cyrket.com/p/android/com.greenecomputing.linpack/
Quadrant http://www.cyrket.com/p/android/com.aurorasoftworks.quadrant.ui.standard/

All tests performed while charging and at a minimum of 90%.

I did three tests with each kernel (15 tests).
Test#1: reboot, wait 60 seconds, run Test#1.
Test#2: run immediately after exiting Test#1, run Test#2.
Test#3: kill all tasks, wait 5 seconds, run Test#3.
I did the above for each of the five kernels, with each of the three apps.
The numbers shared below are the averages of Test#1, Test#2 and Test#3, per app, per kernel.

(Almost forgot, SetCPU is set at max/max, and Governor set to performance for all tests.)

These are my results:
intersectRaven's Kernel BFS:925mV BenchmarkPi 1,373.75
intersectRaven's Kernel BFS:800mV BenchmarkPi 1,287.75
intersectRaven's Kernel CFS:925mV BenchmarkPi 1,399.25
intersectRaven's Kernel CFS:800mV BenchmarkPi 1,247.50
HybridAVS+1113MhzOver...CFS:925mV BenchmarkPi 734.50

intersectRaven's Kernel BFS:925mV Linpack 17.58
intersectRaven's Kernel BFS:800mV Linpack 15.83
intersectRaven's Kernel CFS:925mV Linpack 16.55
intersectRaven's Kernel CFS:800mV Linpack 15.54
HybridAVS+1113MhzOver...CFS:925mV Linpack 34.60

intersectRaven's Kernel BFS:925mV Quadrant 1,480.33
intersectRaven's Kernel BFS:800mV Quadrant 1,410.33
intersectRaven's Kernel CFS:925mV Quadrant 1,363.33
intersectRaven's Kernel CFS:800mV Quadrant 1,422.33
HybridAVS+1113MhzOver...CFS:925mV Quadrant 1,391.67
----- ----- ----- ----- ----- ----- ----- ----- ----- -----
Obviously the overclocked kernel kicks ass in nearly all the tests.
It doubled the average scores in BenchmarkPi & Linpack.

MY QUESTION IS:
But why does it get such similar scores as the others when benchmarked with Quadrant


***MY SILLY DISCLAIMER***
There are many factors not mentioned here (running apps, how many beers I've had, etc.
et cetera yadda yadda yaaa ... ya know?

Your OC results for benchmarkpi seems unusually low, as do your non-OC results for Linpack.