Memory Allocation

Started by
-1 comments, last by Bashar 21 years, 12 months ago
Hi, I was just timing the "new" function of the CRT. Everything is good, using the hi-res timer. However, for some reason the Retail build is MUCH slower then Debug. Seems counter-intuitive, I''d like to think, no? But does anyone know for sure, or have any ideas why this is? my test is simple: int * x[5000]; for( i = 0; i < 5000; i++ ) { x = new int; } this is the only code I''m timing and I get: Retail: Cycles needed: ~1118000 Debug: Cycles needed: ~611000 This is almost 1/2 the # of cycles? I''m using a PIII 850, 512MB. My current thinking is that the CRT (C-runtime) is does special allocations for debug-headers and what not...but that this "cacheing" could possibly be the source. I''m most likely way off the mark.
People fear what they don''t understand, hate what they can''t conquer!

This topic is closed to new replies.

Advertisement