is sceRtcGetTickResolution a constant?

jonny · Post by **jonny** » Fri Jan 27, 2006 2:25 am

as the subject says, can i assume this?

thanks

Raphael · Post by **Raphael** » Fri Jan 27, 2006 3:16 am

It shoul be, since the time generation is based on these ticks, so if the ticks per second weren't constant, you'd get jumps in times.

PacManFan · Post by **PacManFan** » Fri Jan 27, 2006 3:36 am

I'm not sure of this, but I think it's based on the clock frequency. You might want to run a few tests to see what it is at 166, 266, and 333. Then you can generate your own constants. If you find out what they are offhand, please post them here.

-PMF

Raphael · Post by **Raphael** » Fri Jan 27, 2006 4:00 am

Well, that's what I was implicitly expecting, but should have mentionend. It's constant under a specific clock frequency.

chp · Post by **chp** » Fri Jan 27, 2006 4:18 am

When I looked at sceRtcGetTickResolution(), it was something like 1000000, which sounds like an oddly round number for 222Mhz. I suspect that if the timer is cycle-based, the RTC compensates this in its calculations.

bulb · Post by **bulb** » Fri Jan 27, 2006 4:30 am

My guess is tick resolution should be independent of CPU clock (always the same). Modifing CPU clock only means that some tasks will use more or less cycles, it has nothing to do with resolution. The same is true for Pentium CPUs anyway.

jonny · Post by **jonny** » Fri Jan 27, 2006 6:21 am

seems fixed to 1000000
( i'd bet i have seen it at 1500000 some time ago :S , but i can be wrong here)