Post Benchmark Screen Shots Here!!!!

linpack-11.png


My new best in linpack:p
 
You people keep posting your silly little Quadrant scores with ROMs that use the stagefright hack, which on this device pretty much DOUBLES your score. I went from a 500 average stock rom to 1000 average stock rom by simply enabling that. :D

The Stagefright hack tricks Quadrant into thinking that you completed this one CPU test in 0 seconds because it skips it, so therefore your scores go up! Show us a Quadrant without stagefright hack on for it to be considered even close to 'accurate' :P
 
  • Like
Reactions: zefie
You people keep posting your silly little Quadrant scores with ROMs that use the stagefright hack, which on this device pretty much DOUBLES your score. I went from a 500 average stock rom to 1000 average stock rom by simply enabling that. :D

The Stagefright hack tricks Quadrant into thinking that you completed this one CPU test in 0 seconds because it skips it, so therefore your scores go up! Show us a Quadrant without stagefright hack on for it to be considered even close to 'accurate' :P

Okay, we'll go with realistic & accurate this time, 748mhz, which pretty much everybody can hit, stagefright off, /data on sdcard, ubuntdroid on optimus t:
ext2-2.png
 
Can we just exclude the bogus stagefright results? The only reason it passes so high is because Quadrant uses a timer to see how fast the encoding and decoding takes.

However, the encoding step is FAILING with stagefright enabled, implying a false "zomg you encoded that in half a second!" benchmark. Consequently, the decoding fails just as fast since it has nothing to decode (since encoding failed).

Notice how it's not that high in CM7? CM7 is forced stagefright, but the encoder is working.

Froyo Stagefright benchmarks are fake.

Okay, we'll go with realistic & accurate this time, 748mhz, which pretty much everybody can hit, stagefright off, /data on sdcard, ubuntdroid on optimus t:

Class 10? Hmm.. This may be accurate. The cpu section seems valid, despite scaled wrong in Quadrant. The I/O looks nice, and the internal nand is slow as crap so this is very possible.
 
Last edited:
Can we just exclude the bogus stagefright results? The only reason it passes so high is because Quadrant uses a timer to see how fast the encoding and decoding takes.

However, the encoding step is FAILING with stagefright enabled, implying a false "zomg you encoded that in half a second!" benchmark. Consequently, the decoding fails just as fast since it has nothing to decode (since encoding failed).

Notice how it's not that high in CM7? CM7 is forced stagefright, but the encoder is working.

Froyo Stagefright benchmarks are fake.



Class 10? Hmm.. This may be accurate. The cpu section seems valid, despite scaled wrong in Quadrant. The I/O looks nice, and the internal nand is slow as crap so this is very possible.

Class 6, formatted my sd card with mke2fs -b 4096, just changed init.rc to mount it from mmcblk0p2.
 
Okay, we'll go with realistic & accurate this time, 748mhz, which pretty much everybody can hit, stagefright off, /data on sdcard, ubuntdroid on optimus t:
ext2-2.png

You used the samsung galaxy dirty i/o (aka galaxy lag fix) which inflates your I/O causing for a higher, inaccurate quadrant score. Also, it was this same hack, which they did on the internal drive, which caused hardware to start failing. [Reference] The life of your SD card my vary doing this.

Still, the screen shot with this hack and stagefright was cool to see lol :) you should edit that image, cut the status bar out and make that your phones background wallpaper lol. You could just walk around handing people your phone so they can see it.
 
Last edited:
You used the samsung galaxy dirty i/o (aka galaxy lag fix) which inflates your I/O causing for a higher, inaccurate quadrant score. Also, it was this same hack, which they did on the internal drive, which caused hardware to start failing. [Reference] The life of your SD card my vary doing this.

Bingo. Most of the work I've done with kernels, roms, and recoveries has been on the samsung moment & intercept, so I became familiar with ways to get rid of rfs. I know about the instability, that's why I have it on the sd card, so the nand doesn't fail, and why I haven't released it yet, since it may be dangerous. And overly inflated or not, the phone is noticeably faster this way.
 

Trending Posts

Forum statistics

Threads
956,811
Messages
6,970,119
Members
3,163,627
Latest member
Vicarious1