Droid speed comparisons

[iFrostyy]

New thread here
http://androidforums.com/motorola-droid/43257-new-droid-speed-comparisons.html

 

[spazma7ik]

Wow, nice tests, a lot faster! And can you give me a bit of info on what setting the OC to 800/250 compared to 800/800 does differently? I know that with 800/250 it allows the clock speed to be dramatically decreased, but when does this happen? Cause obviously it would be when you're using the phone with the screen on, but why would I want the speed to go down that low when I'm using it? I'm truly just confused, so I would enjoy some input on how and where this helps out?

 

[iFrostyy]

From what i understand, i could be wrong but im pretty sure that when you are programs that dont use as high a frequency it goes down in mhz. For example if you are replying to a text less mhz is needed compared to using a program such as Google Navigation. If you have it at 800/800 it will stay at 800 mhz to send a text rather then going down in mhz using an un needed amount of mhz. But thats what i understand, i could be completly wrong

 

[alostpacket]

Nice tests! Thanks for sharing.

Did you clear the cache before testing? Still a useful test, but I was just curious. I almost think the OC'd would be even better on a fresh download, though maybe only a tiny bit.

 

[iFrostyy]

No i didnt clear cache but i did however re-add each kernal before testing. Again i am going to make more tests but right now i dont have time. To busy watching the Superbowl

 

[jasonb1985]

nice! i'm running at 800. i knew it seemed like it rendered web pages quicker.

 

[xmr405o]

nice! i'm running at 800. i knew it seemed like it rendered web pages quicker.

yeah, thanks man. I was wondering how an OC'd Droid compared to a stock Droid. Besides browser tests, what other tests can we see between stock vs OC'd.

 

[monsterenergy22]

not sure why a faster processor would improve internet speed

 

[alostpacket]

not sure why a faster processor would improve internet speed

Because it can process and render the data faster.

 

[Shadowtech]

I just recently rooted my Droid and installed the Bugless Beast on it. Made me a little nervous I might add cause I was just waiting for it to brick or something. Anyway, I had noticed my browser did seem a lot faster as well... along with everything else on the phone. Installing apps from the market is almost instant now, which is great.. it used to freeze up my phone every now n then when I added a new one.

 

[doomedromance]

My 950mhz on VZW 3G beat the stock on wifi in your video. Haha. O.C. 4TW
Add that to the test. 1GHz on 3g v. 550 on wifi. haha.

 

[Psychokitty]

Besides browser tests, what other tests can we see between stock vs OC'd.

You can increase video bitrate & framerate in the recorder.

 

[oakley1261]

That's a very skewed benchmark. If your going to use the same Wifi/Internet connectiong to test with, you need to time the results per phone, you can't do both at the same time. You are better off doing a comparison with 3g. I already know that overclocking, or having a faster processor in general will increase the load time of web pages. but this hardly identifies how much faster it will load.

 

[CRPercodani]

^^What he said. You have both Droid's sharing the same WiFi connection. Try to time them next time doing individual tests.

 

[iFrostyy]

Ya that does make sense, silly me. Once i get my friends Droid back i will redo the tests and add some more tests. Any other comparisons people want to see?

 

[alostpacket]

That's a very skewed benchmark. If your going to use the same Wifi/Internet connectiong to test with, you need to time the results per phone, you can't do both at the same time. You are better off doing a comparison with 3g. I already know that overclocking, or having a faster processor in general will increase the load time of web pages. but this hardly identifies how much faster it will load.

I dont see why doing them both at the same time would matter much -- especially if he is on 802.11g. It's not like one device gets exclusive use of the bandwidth on an internet connection. Each web page load is broken down into somewhere between several dozen to a couple hundred HTTP requests which are further broken down into however many more TCP packets.

Unless there is something inherent about Wi-Fi that I haven't heard about that causes bottle necking, there shouldn't be any problem running these thing simultaneously.

 

[Fabolous]

I think it really depends on the quality of your wireless router.

Anyways, we've all probably seen that infamous Engadget test where they tested 3 devices on the same wifi concurrently and failed big time.

It doesn't hurt just to do them one by one and time them individually, especially if you're using a crapo router.

 

[iFrostyy]

I have on of the best routers on the market. But you never know who can be stealing my internet :S

You guys are correct on the individual testing+ timing because it eliminates a couple of variables that could skew the results.

 

[alostpacket]

Secure your router with WPA2 and a long password and no one should be able to jack it.

But yeah, I guess scientifically speaking it is better to time them. Also even to do many tests over a period of several days and weeks and average them. But I think doing them at the same time is good enough for a YouTube test heh. But if you want to try the timing method, have at it.

 

[iFrostyy]

It would be the best way to satisfy the experts ;D

 

[Fabolous]

It would be the best way to satisfy the experts ;D

For the geeks, some may want to see side by side BenchmarkPi and Linpack results.

They're not perfect benchmarks, but they do show the difference in theoretical processing power quite well.

 

[iFrostyy]

I wouldnt know how to do that if they asked xD

 

[alostpacket]

One way to show if they are hogging each other's bandwidth is to do a side by side speedtest.net test.

Those should come out close to the same running side by side because it doesn't process the data it gets back just times it.

 

[alostpacket]

I wouldnt know how to do that if they asked xD

Just search for becnchmark and linpack in the market, they are both free

 

[iFrostyy]

If i where to perform these tests, how would you suggest i set them up?