Originally Posted by Tr1cK
I wonder how I managed to pull of the exact same #5 as you in your first test Night? What does #5 do?
Test #1.. bitmap rescaling using NEAREST sampling.. (fully software)
Test #2.. bitmap rescaling using LINEAR sampling.. (also fully software)
Test #3.. fills to the memory
Test #4.. Translucency rendering 640x480 bitmap on a 1024x768 framebuffer (50%)
Test #5.. Renders 8000 texts into the framebuffer using a bitmap font.
cool part of it is that different brand of CPU results in completely different rates:
example: (Xp2600+ vs P4 2.4Ghz/533)
AthlonXP 2600+@stock, 1024MB DDR400@DDR333
Test #1 score: 336
Test #2 score: 252
Test #3 score: 194
Test #4 score: 410
Test #5 score: 28469
This system scores a g-mark score of: 1557
Intel Pentium IV 2.4Ghz @ 2.4Ghz, 1024MB DDR333
Test #1 score: 421
Test #2 score: 128
Test #3 score: 248
Test #4 score: 505
Test #5 score: 14209
fps (p4 does very bad compared to an Athlon in this test)
This system scores a g-mark score of: 1335
it seems the P4 scores good in the bad-ass memory filling department and the Athlon system scores extremely well in the fast-randomly into the memory accesses.
scores designated in red are slower compared to the rival. green is faster. this makes me wonder how a Pentium IV 3.2Ghz/800 will score? and prescott vs Northwood (512kB/1M cache). even ClawHammer with 1M cache vs NewCastle/Winchester with 512kB cache..
I am wondering how much influence cache has on the score.