Droid X successfully overclocked

[luv2increase]

Sorry guys, haven't been able to get a successful compile working since I don't have the actual source that they're working off of. I've tried a few ways so far to try to trick it into installing, but so far no luck. We may have to wait for the official release from Moto so that they will post the source code so I can use it to recompile the module.

I'll keep at it, but don't get your hopes up before the official 2.2 release.

Yeah, I along with everyone at the DroidXForums really like what you have done!!! We really appreciate it! Keep up the good work!

 

[coolbho3000]

Sorry guys, haven't been able to get a successful compile working since I don't have the actual source that they're working off of. I've tried a few ways so far to try to trick it into installing, but so far no luck. We may have to wait for the official release from Moto so that they will post the source code so I can use it to recompile the module.

I'll keep at it, but don't get your hopes up before the official 2.2 release.

This is how I made a kernel module work under a similar situation (no source).

You can actually take the kernel source from kernel.org with the same version as the one you want to modify (eg. take 2.6.32 if you want your module to load into a 2.6.32 Motorola kernel), pull the config from the Motorola kernel for which you do not have the source, attempt to compile the kernel using this config (you'll likely get an error - the compile does not have to actually finish), then compile your module against the kernel. Hex edit the vermagic string in the module to fit the kernel you want the module to load it in (you can find the proper vermagic string by attempting to load the unedited module then running a dmesg).

As far as I can tell, as long as the main kernel version is the same as the one the module was compiled on, this method should be safe. Otherwise, the system will most likely crash. I've tested this extensively with HTC devices and my original Droid.

It's hacky, and ignores grave warnings about kernel ABI compatibility, but it works.

 

[xsmooth69x]

i have no idea what you just said but its awsome

 

[Elkay]

Sorry guys, haven't been able to get a successful compile working since I don't have the actual source that they're working off of. I've tried a few ways so far to try to trick it into installing, but so far no luck. We may have to wait for the official release from Moto so that they will post the source code so I can use it to recompile the module.

I'll keep at it, but don't get your hopes up before the official 2.2 release.

This is how I made a kernel module work under a similar situation (no source).

You can actually take the kernel source from kernel.org with the same version as the one you want to modify (eg. take 2.6.32 if you want your module to load into a 2.6.32 Motorola kernel), pull the config from the Motorola kernel for which you do not have the source, attempt to compile the kernel using this config (you'll likely get an error - the compile does not have to actually finish), then compile your module against the kernel. Hex edit the vermagic string in the module to fit the kernel you want the module to load it in (you can find the proper vermagic string by attempting to load the unedited module then running a dmesg).

As far as I can tell, as long as the main kernel version is the same as the one the module was compiled on, this method should be safe. Otherwise, the system will most likely crash. I've tested this extensively with HTC devices and my original Droid.

It's hacky, and ignores grave warnings about kernel ABI compatibility, but it works.

Yep, it's quite possible to do it that way, too. Usually works fine. I have vanilla 2.6.32 right now compiled and I'm working off that. Unfortunately, unless I'm missing some code somewhere it looks like they may have changed how the power management works. A bunch of stuff that was in 2.6.29 seems to be missing. I have a fully built 2.6.32 kernel without errors, so I think it's more-so changes than missing pieces. Can't even get the source for the overclock to compile. Tiago and I are both looking at it.

 

[El Poblano]

Ok i have a new question someone may be able to help me with. I've been successfuly undervolting now for a while. I've found my most stable clock speed at 1.1 ghz, but my question is about voltage. I've been dropping the voltage slowly every other day or so (to ensure stability). Currently my settings are: 1.1 - 32 vsel. 800 - 30 vsel. 600 - 28 vsel. 450 - 26 vsel. Is it just me or does this seem WAY low yo be stable? (66 is stock at 1 ghz) i've done the mpu_opps and it confirms the voltage, but i still wonder. Any ideas?

 

[Jedb]

is there anyway someone could make a video of setting up how to overclock?

 

[mirage]

Milestone/Droid Overclock 1.4

Just to let you know I've released Milestone/Droid Overclock version 1.4. Supports Droid X (2.1) and autoload on boot, so you don't have to mess with boot scripts. It's on the market as usual and on the project site.

 

[phnxdestroyer]

does this overclock work on leaked 2.2? the overclock.sh and .ko

 

[hijackerjack]

does this overclock work on leaked 2.2? the overclock.sh and .ko

i think you may need to have a look through the past few pages of this thread.... lol

and thats a no...

 

[L4WD0GG]

psha, i7 930 at 4.28 on air Droid X at 1.GHz still, actually, scratch that, I've got it downclocked to 600MHz currently for battery life. Performance with Launcher Pro is still excellent, the 512RAM is the current bottleneck for me.

Does the Droid X have 512MB RAM? I read somewhere that it comes with only 256MB while the Droid 2 comes with 512MB. As I understand, it has RAM built into the GPU. This is why it can output a 720p picture.

Thoughts?

Any results on that overclocked Droid X? Does said device also have froyo?

 

[tobytl]

Yea, it does have 512MB RAM. cat /proc/meminfo reports 488748 kB total ram for the linux environment (the rest being allocated to the radio from what I gather).

 

[givmedew]

Alright, I managed to get my Droid X overclocked last night. Initial testing is not looking very promising.

I wasn't able to get Quadrant to even do a full run without rebooting the phone at 1.1Ghz. Looks like Motorola has this thing clocked surprisingly close to its frequency limit on stock voltages out of the box unless I ended up with a bad sample. Keep in mind that this was with stock voltages, though, so not all hope is lost yet. I didn't get this working until about 4am so I didn't have a chance to mess with the voltages since I had work this morning.

I'll work on it some more tonight and report back with additional findings.

Now that this thread is so very very long shouldn't you be editing the original post to have updated information. You can't expect people to read each and every response to find what is important and realvant, this is the way pretty much every rom developer does it. Could you please step up to the plate as a leader in this topic and edit your orignal post to contain the most up to date information? Also include something like Updated on 08/31/10 at the top. For example I like skyraiders rom for the the DINC so I have his thread bookmarked I can quickly click on it and see if any updates are out there been following him for a long time now.

I've been interested but not willing to follow you for quite some time.

Please not just for me but for everyone PLEASE update your original post.

 

[6thgear]

i just overclocked to 1.15ghz using droid milestone, can see a noticeable difference, snappier web browsing and overall processing speeds. wow cant wait to see this thing at 2ghz!

 

[fmkaiba]

i just overclocked to 1.15ghz using droid milestone, can see a noticeable difference, snappier web browsing and overall processing speeds. wow cant wait to see this thing at 2ghz!

I don't think u will
you MAY see 1.3Ghz, but anything about 1.2Ghz right now seems impossible.

The droid u could overclock about double the stock setting because the stock was scaled down so far from what the chip was capable of.

The droid X on the other hand, motorola has done a good job of clocking the chip to the stable maximum. I got mine up to 1.15Ghz before instability occurred in the navigation application

 

[stevil]

Ok something just doesn't seem right. I am trying my best to undervolt the DX, but I turn down the voltage and and I just don't think its taking. These voltages are pretty damn low, and the stress tests I've run haven't locked it up yet.

Yes I've placed overclock.ko in /system/lib/modules/, yes I've placeed setscaling.sh in /system/bin/ and yes I am running SetCPU. And if you look at the output below you'd think commands would have taken.

#cat /proc/overclock/*
freq_table[0] index=0 frequency=1000000
freq_table[1] index=1 frequency=800000
freq_table[2] index=2 frequency=600000
freq_table[3] index=3 frequency=300000
0xc0573ffc
cpumin=300000 cpumax=1000000 min=300000 max=1000000 usermin=300000 usermax=1000000
clk_get_rate=1000000 omap_pm_cpu_get_freq=1000000000
1000000
24
mpu_opps[4] rate=1000000000 opp_id=4 vsel=24
mpu_opps[3] rate=800000000 opp_id=3 vsel=20
mpu_opps[2] rate=600000000 opp_id=2 vsel=18
mpu_opps[1] rate=300000000 opp_id=1 vsel=14
0xc0574e24
0x0
1.2

Thoughts?