Galaxy S performance based on Quadrant and Lag fix

[tempest7]

Hi everybody!

I'm a SGS user! Like many other users, my stock Galaxy S had it's share of lag and hangups.

For background information I have the International GT-i9000 version running Android 2.1-update 1.

I love the phone, except the lag and the mysterious shutdowns and freezes were quickly turning my love into HATE. I downloaded the updated 1.5.1 Kies off the support site. Used it to update my phone from I9000ZSJF7 to I9000ZSJH2

That didn't really solve my problems. I didn't even see any improvements with the update. I think my reception/connections actually got worse.

So I turned to the quick and very easy (after you figure out how to properly reboot the device and switch modes) RyanZA Lag fix. So far, the phone's been acting really good. No freezes, seems quicker as well even with all the crap I keep loading onto it. (So I definitely recommend the fix)

From what I've seen almost everyone benchmarks their Android phone using Aurora Softwork's Quadrant.

I see a few posts and screen caps of the post Lag fix results and they're WAY over the Nexus One running 2.2; Froyo.

I guess the Standard and Advance versions of the Quadrant software are different. They don't break down the individual scores. Take a look at these screen shots.

Nexus One (huge CPU score a la Froyo JIT?)

Lag Fixed HK version Galaxy S (huge IO score)

Based on the description on how the lag fix works, the I/O score makes sense, but are we gaming the benchmark? Is this design flaw, the lag, built into every other android system?

 

[thechamp]

Thats a nice study tempest.

But I am not superised with findings
As we all know most of the problems in galaxy is due to poor I/O performance of rfs file system. and what a lag fix actually do is try to replace the same with more efficient file system like EXT2/3/4.\

So whatever boost we see will be in I/O part only.

Thats why mere a very high score dont mean great performance as score is not balanced.

Hopefully froyo will improve some Blue part(CPU performance) as well like it did in case of Nexus one. A lag fix on top of that and we just might have a super phone in our hand.

 

[JiMMaR]

am not sure about this part , but as I remember , the lag fix also makes the phone uses ram instead of always writing to the disk
as it writes to ram , then moves 'em all to disk at once (every 10 entries or something)
and as that we know ram is faster , and here you are actually reducing disk access so you get a faster a device

 

[tempest7]

Most of the time people are using the free, as in Standard, version. It makes it seem that the Nexus One using 2.2 is 2/3rds the speed. Even though the phone seems faster than pre-lag-fix (for me, the important part is that it seems more stable and responsive) we're only seeing the improvement in 1 segment. But, it is still generating a huge score boast.

Also I'm curious, prior to running the bench mark I checked the system information that was reported. As you can see, I have the phone plugged in; charging via USB. It still reports that the current CPU speed is 800 MHz. I guess during the calculations it would throttle up based on load?

I'm obviously not trying to go for the highest results. It's just funny how the community developed a fix that can dramatically improve I/O performance so much. The scores for Quadrant don't seemed to be weighed correctly though if we're scoring so high after just this lag fix? Anyone running the Beta Froyo releases report if they still need, or if they still can apply the lag fix? Does it help?

 

[lobo76]

Also I'm curious, prior to running the bench mark I checked the system information that was reported. As you can see, I have the phone plugged in; charging via USB. It still reports that the current CPU speed is 800 MHz. I guess during the calculations it would throttle up based on load?

I would think so. I've installed a CPU+Temp app and it goes as low as 100Mhz when it's doing nothing.