Michael and I would like to run this on a few different machines, just to get an idea :P
here's what i get on my machine, running a Prescott w/htt, XP MCE2005 SP3
Sleep(1) delay 1.953 ms
Sleep(10) delay 10.742 ms
Sleep(1) delay 1.953 ms
On my laptop:
Intel(R) Pentium(R) 4 CPU 3.20GHz (SSE2)
Sleep(1) delay 1.953 ms
Sleep(10) delay 10.742 ms
Sleep(1) delay 1.953 ms
Dave.
AMD Phenom II X6 1100T 3.31GHz on Windows 7 Pro x64
Sleep(1) delay 15.587 ms
Sleep(10) delay 15.610 ms
Sleep(1) delay 1.029 ms
thanks guys
ok sinsi - nothing ambiguous about that run - lol
leaves me thinking, "now, i don't know what to think" :redface:
P3, Windows 2000:
Sleep(1) delay 10.014 ms
Sleep(10) delay 10.014 ms
Sleep(1) delay 1.013 ms
P4 Northwood, Windows XP HE Version 2002 SP3:
Sleep(1) delay 15.624 ms
Sleep(10) delay 15.613 ms
Sleep(1) delay 1.965 ms
P4 3GHz XP SP3
Sleep(1) delay 3.906 ms
Sleep(10) delay 11.718 ms
Sleep(1) delay 1.969 ms
Why Sleep(1) = 3.906 ms and = 1.969 ms ?
OS : Windows v5.1.2600 Service Pack 3
Processor : (4x) Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
Sleep(1) delay 1.953 ms
Sleep(10) delay 10.742 ms
Sleep(1) delay 1.953 ms
Sleep(1) delay 8.706 ms
Sleep(10) delay 15.625 ms
Sleep(1) delay 1.965 ms
Press any key to continue ...
Hi,
Redirection to a file was not successful. Cut & paste.
P-III, Windows 2000.
G:\WORK>MICHAELS
Sleep(1) delay 10.014 ms
Sleep(10) delay 10.014 ms
Sleep(1) delay 1.012 ms
Press any key to continue ...
P-MMX, Windows 98.
A:\>michaels
Sleep(1) delay 4.994 ms
Sleep(10) delay 13.792 ms
Sleep(1) delay 1.973 ms
Press any key to continue ...
Regards,
Steve N.
Win7 64-bit Pro, SP1
AMD Phenom II X6 1035T 2.80 GHz
16GB
Sleep(1) delay 1.000 ms
Sleep(10) delay 10.000 ms
Sleep(1) delay 1.000 ms
thanks everyone :t
By setting the interval to 1 before any timings I got clive's numbers exactly.
The interval is a global windows setting, so any program that sets it has to reset it, maybe some apps don't.
Quote from: sinsi on June 19, 2012, 12:25:24 AMBy setting the interval to 1...
how did you do that ?
Over the years I have tested this on a fair number of systems, not just my systems but other systems that I had worked on or otherwise had access to, and the results have been very uniform. Now they appear to be all over the map.
Perhaps this is a side effect of the change to multi-core systems. My P4 Northwood system may behave differently than Dave's Prescott system because at 3.06 GHz it was the first P4 with HT.
perhaps it is the test code that is giving us fits
maybe review it to see if it is sound
of course, we don't really know how time-slices are devised
at least, i couldn't find much about it on msdn