Motorola Droid gets Overclocked.

[1033]

Should be noted for noobs that modern processors are not 'individually' tested or specked. This means some chips can handle OC very easily (and be highly OC'd) while other will not. So if you decide to OC be sure to go in small increments and test the device thoroughly and not jump straight to a high number).

Im pretty sure i7 processors (quite modern) are individually tested, and if they pass they're clocked in to see when they fail. If they work under normal conditions they get the 920 stamp, if they work overclocked maybe 1ghz higher they get the 940 stamp. Now if they can raise the clock up fairly high, and the tempature doesn't exceed some certain threshold they know it made it though the FAB process perfect and the multiplier can be unlocked.

this is from wikipedia:

On May 17, 2005, Sony Computer Entertainment confirmed some specifications of the Cell processor that would be shipping in the forthcoming PlayStation 3 console.[22][23][24] This Cell configuration will have one Power processing element (PPE) on the core, with eight physical SPEs in silicon.[24] In the PlayStation 3, one SPE is locked-out during the test process, a practice which helps to improve manufacturing yields, and another one is reserved for the OS, leaving 6 free SPEs to be used by games' code.[25] The target clock-frequency at introduction is 3.2 GHz.[23] The introductory design is fabricated using a 90-nanometer SOI process, with initial volume production slated for IBM's facility in East Fishkill, New York.[14]

for intel they just make it a 920 or 940 to improve their yields... and make the EE $1000 to compensate. When they all went through the same manufacturing process.

 

[KZIWarrior]

Im pretty sure i7 processors (quite modern) are individually tested, and if they pass they're clocked in to see when they fail. If they work under normal conditions they get the 920 stamp, if they work overclocked maybe 1ghz higher they get the 940 stamp. Now if they can raise the clock up fairly high, and the tempature doesn't exceed some certain threshold they know it made it though the FAB process perfect and the multiplier can be unlocked.

Processors are tested individually while they are still on the wafer before being cut into individual dies (a process called 'wafer sort testing'). But this testing is only a test to make sure the processor handles basic calculations correctly. While certain processor lines (like you i7 example, which is a completely different class of processor) may be further tested for actual clock speeds (and in RARE cases overclocking potential) it is not standard for most processors. And even then processors are not 'individually' tested. One processor per wafer (or sometimes per batch) is tested and based on it's performance the remaining processors are labeled/sorted accordingly. This is way you can have such great variances in the over-clocking potentials seemingly identical chips.

On a different note I found this on the ARM site:
ARM Cortex-A8 - ARM Processor

Of particular interest:
"With the ability to scale in speed from 600MHz to greater than 1GHz, the Cortex-A8 processor can meet the requirements for power-optimized mobile devices needing operation in less than 300mW"

So I wonder if these over-clocking kernels/drivers for the Droid are actually over-clocking the CPU or simply changing the scaling architecture??

 

[cb2000a]

Max speed reached was 1.3 ghz. Not recommended though.

 

[jucytec]

Should be noted for noobs that modern processors are not 'individually' tested or specked. This means some chips can handle OC very easily (and be highly OC'd) while other will not. So if you decide to OC be sure to go in small increments and test the device thoroughly and not jump straight to a high number).

Im pretty sure i7 processors (quite modern) are individually tested, and if they pass they're clocked in to see when they fail. If they work under normal conditions they get the 920 stamp, if they work overclocked maybe 1ghz higher they get the 940 stamp. Now if they can raise the clock up fairly high, and the tempature doesn't exceed some certain threshold they know it made it though the FAB process perfect and the multiplier can be unlocked.

this is from wikipedia:

On May 17, 2005, Sony Computer Entertainment confirmed some specifications of the Cell processor that would be shipping in the forthcoming PlayStation 3 console.[22][23][24] This Cell configuration will have one Power processing element (PPE) on the core, with eight physical SPEs in silicon.[24] In the PlayStation 3, one SPE is locked-out during the test process, a practice which helps to improve manufacturing yields, and another one is reserved for the OS, leaving 6 free SPEs to be used by games' code.[25] The target clock-frequency at introduction is 3.2 GHz.[23] The introductory design is fabricated using a 90-nanometer SOI process, with initial volume production slated for IBM's facility in East Fishkill, New York.[14]

for intel they just make it a 920 or 940 to improve their yields... and make the EE $1000 to compensate. When they all went through the same manufacturing process.

I've been overclocking since the Intel OverDrive days... While I'm certain that Intel CPUs do undergo some testing, I can assure you they don't OVERCLOCK their chips prior to leaving the their manufacturing facility. At most the i7 920/940 chips would be tested under stress in their supported turbo mode in a Sample Batch of some sort. They rest would probably be inspected visually by some advance computer scans which would detect anomalies at the wafer level. Turbo mode just allows the clock MULTIPLIER to be up by 1x. This is different then upping the clock frequency, which is the more common overclocking terminology. For example, an i7 runs at 133MHz clock x 20 multiplier = 2660MHz in turbo mode is x21 which would equate 2793MHz My system at home is running 171MHz x 21 @ 3591MHz Overclocking usually requires a LOT more power (as high as 1.375v for i7 from the stock 1.2), and results in heat related failures if the CPU's not properly cooled. The Extreme Edition that someone mentioned allows the multiplier to go up to 24... which is bad a$$, but not too many of consumers can spare 1000$ on a CPU. ARM Cortex A8 is a different story tho, it uses around the same voltage, a little lower, 1.2v or LESS but it is NOT computing the same number of calculations. also according to the ARM Cortex-A8 spec, it can actually run 1100MHz, and according to the little chart on the bottom, 1100MHz manages to use 1v using Dynamic Power. IF they the chip was set at 550MHz for longer standby time, then using SETCPU to set the ARM Chip at 125MHz would make it run Possibly even longer than stock while being overclocked. There are other information on the Web about leakage controls which explains the relationship of power/performance gains/loss. If you follow the thread on Alldroid, you'll see that some people are reporting LONGER battery life on 1GHz OC. I've been at 1GHz as well for a few days now, EVERYTHING is stable AND I had hard time draining the battery with regular/some-what-heavy use. I'm loving it so far...

 

[KZIWarrior]

? for you OverClockers

Are you all using cases (Seidio, BodyGlove ect) or running naked?? Was wondering if cases seam to have any affect on the temp.

 

[JhankG]

Are you all using cases (Seidio, BodyGlove ect) or running naked?? Was wondering if cases seam to have any affect on the temp.

I'm using the Seido case and holster running at 900 mhz and haven't had any heat issues. I ran the Navigation with the case on for about 20 minutes straight yesterday and couldn't feel any change in the temp. I was actually wondering about this same thing earlier this morning.

 

[jucytec]

Are you all using cases (Seidio, BodyGlove ect) or running naked?? Was wondering if cases seam to have any affect on the temp.

i have a Seidio. i've also put Just In Time compiler library which really didn't help...

 

[JhankG]

Are you all using cases (Seidio, BodyGlove ect) or running naked?? Was wondering if cases seam to have any affect on the temp.

i have a Seidio. i've also put Just In Time compiler library which really didn't help...

So not much of a difference with JIT? Even overclocked to 1.2 ghz, the Droid's Mflop score is well below the Nexus One's with the CyanogenMod v5.0

 

[jucytec]

Are you all using cases (Seidio, BodyGlove ect) or running naked?? Was wondering if cases seam to have any affect on the temp.

i have a Seidio. i've also put Just In Time compiler library which really didn't help...

So not much of a difference with JIT? Even overclocked to 1.2 ghz, the Droid's Mflop score is well below the Nexus One's with the CyanogenMod v5.0

Either dalvikvm was optimized for Nexus one or... JIT kicks ass on Snapdragon. on mine it was almost NO performance gain

 

[hacku]

On a different note I found this on the ARM site:
ARM Cortex-A8 - ARM Processor

Of particular interest:
"With the ability to scale in speed from 600MHz to greater than 1GHz, the Cortex-A8 processor can meet the requirements for power-optimized mobile devices needing operation in less than 300mW"

So I wonder if these over-clocking kernels/drivers for the Droid are actually over-clocking the CPU or simply changing the scaling architecture??

Hmmm... this is very interesting.

 

[m21knight]

I've said it in other threads before, but I will say it here as well. Outside of academic curiosity, I would strongly discourage people from overclocking their mobile device for normal operation. These devices are nothing like regular computers they dont have the cooling systems necissary to maintain a safe operating temprature.

The folks who overclock in the lab setting are not testing their systems outside of the room temprature range, try doing that on an 85 degree day with no AC and your toast...yes that was a terrible pun.

 

[m21knight]

Or at least your Droid will be toast... :laugh:

 

[JhankG]

Guys, we know the risks of overclocking. For those who didn't know the risk, I am sure they read one of the dozens of warnings from other people on this forum that have popped up almost daily.

 

[victorythagr8]

The thing about overclocking the droid is that they actually over clocking it without messing with the voltage which is probably why there barely is any heat increase. As long you have SetCPU and have scale at 125 as the minimal that you will not see any damage.

 

[CorvetteAustin24]

The thing about overclocking the droid is that they actually over clocking it without messing with the voltage which is probably why there barely is any heat increase. As long you have SetCPU and have scale at 125 as the minimal that you will not see any damage.

+1. People think electronics are rather fragile items. They will shut off before damage occurs these days, and the odds of that are minute because the voltage is not increasing. Overclocking will show a slow linear temperature increase the higher the Mhz, overvolting will show an exponential increase of temperature.