Thread: Evo 4g/Evo 3D
View Single Post
  #6 (permalink)  
Old 06-11-2011, 11:43 AM
BlackDynamite's Avatar
BlackDynamite
VIP Member
Offline
 
Join Date: Sep 2006
Posts: 1,839
Reputation: 1190
BlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on repBlackDynamite is halfway to VIP status based on rep
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Re: Evo 4g/Evo 3D

Quote:
Originally Posted by gTen View Post
As I said earlier, Quadrant and Linpack are defective benchmarks.

Quadrant is defective because without the pro version you cant exactly see whats going on and even with the pro version it sometimes inflates values. The higher score though is due to the higher FPS the video was playing on the SGS2 while the sensation was frame skipping, hence why it finished faster. Either way people should really stop using Quadrant altogether, there are better benchmarks.

Linpack is also a failed benchmark. Its only use is for benching same devices. Simply because snapdragon phones cheat this benchmark, Linpack works by calculating Floating points. Snapdragon has this thing called VFP, aka Vector Floating Point,its a chip designed specifically for doing floating points.

ARM architecture - Wikipedia, the free encyclopedia

Which kinda defeats the purpose of Linpack..I am unsure whether or not the Exynos has a VFP too or not, I know the previous hummingbird did not, but either way this benchmark is just a measurement of how fast the VFP works for those phones that ave it, not how fast the CPU works as intended.


Smartbench on the other-hand is a decent benchmark. the extra pixels is irrelevant, its not down-scaling, most likely it works similar to the density apps where it tricks the phone to think its WVGA. As I said, considering Smartbench did this, most likely its correct and even comparison, they are the pros at this and know better then both of us in the very least on this subject. (smartbench is made by a team on xda)
Smartbench is no better than linpack. Smartbench runs at a resolution not compatible with qhd devices. Period. Whatever has to be done to make the Sensation think it's only wvga, does not have to be done on the Galaxy S 2. Period. There is no debating this and it's funny that you keep trying. Also it's no coincidence that this is the one benchmark where the Galaxy S 2 is faster, and it's the only one you give any credence. I don't see any reason why smartbench is any better than linpack. I don't put much stock in either of them, but at least linpack is running at the proper resolution on both devices.

Last edited by BlackDynamite; 06-11-2011 at 11:47 AM.
Reply With Quote