TIX -> QueryPerformanceCounter

[Todd Wasson]

I assumed when I saw TIX/END TIX that it would be using QueryPerformanceCounter internally (or the equivalent ASM instruction which escapes me now). This doesn't appears to be the case though. I've replaced QueryPerformanceCounter with TIX in a system and the numerical results returned are radically different, often times displaying negative values as though negative time had passed between two events. Does anyone have some insight into how TIX/END TIX works? QueryPerformanceCounter is supposed to access the same high resolution CPU cycle counter that I thought TIX would use. Where am I wrong?

[Paul Dixon]

Todd,
TIX uses the CPU time stamp counter which runs at CPU clock frequency (GHz and up)
QueryPerformanceCounter uses the internal timer chips which usually run at about 1.7MHz or 3.5MHz

Paul.

[Todd Wasson]

Nevermind this thread. I made a mistake. Feel free to delete it

[Todd Wasson]

Wrote mine while you were posting yours, Paul. Thanks. I made two mistakes then.

QueryPerformanceFrequency on my laptop returns over 14Mhz, btw..

[Arthur Gomide]

The QueryPerformanceFrequency seems not to have any significance to measure the frequency of the processor. I got only one true value. See the results:

    type / motherboard        processor / frequency         QueryPerfFreq   sysinfo
------------------------------------------------------------------------------------
- nb HP Pavilion DV4 1222 & AMD Turion X2 64bit 2.1 GHz ..:    14,282,268 - 2100 MHz
- pc Intel D865perl       & Intel P4 HT 2.8 GHz ..........: 2,793,050,000 - 2793 MHz (true)
- pc Zida 4DPS 64MB       & AMD 486DX 133 MHz ............:     1,193,180 -  133 MHz
- pc Asus K7M PC133       & AMD Athlon 1.0 GHz ...........:     1,193,182 - 1011 MHz
- pc Asus P4M266A-8235    & Intel P4 2.27 GHz ............:     3,579,545 - 2270 MHz
- pc Gigabyte 7VM400M-RZ  & AMD Sempron 1.6 GHz (2.3+) ...:     3,579,545 - 1598 MHz
- pc ECS 761GX-M754 5.0   & AMD Sempron 1.4 GHz (2.5+) ...:     3,579,545 - 1403 MHz