你好,
一个32位Windows系统上你可以使用一个内核函数。我测试它,它工作得非常准确,硬件定时器。不幸的是我还没有找到一个解决方案涉及64位。欢迎任何提示。
干杯
安德烈亚斯
这里为32位Windows的一个例子:
如果~ libisloaded (“QPerf”)
loadlibrary (“kernel32.dll”@QueryPerformance,“别名”,“QPerf”)
结束
价值。LowPart = uint32 (0);
价值。HighPart = uint32 (0);
QStruct = libstruct (“s_ULARGE_INTEGER”、价值);
[xval, QFreq] = calllib (“QPerf”,“QueryPerformanceFrequency”,QStruct);
频率= QFreq。HighPart * 2 ^ 32 + QFreq.LowPart;
[~,QCounter] = calllib (“QPerf”,“QueryPerformanceCounter”,QStruct);
t0 = (QCounter。HighPart * 2 ^ 32 * 1000 + QCounter.LowPart) /频率;
t1 = t0;
而t1 - t0 < 1000
[~,QCounter] = calllib (“QPerf”,“QueryPerformanceCounter”,QStruct);
t1 = (QCounter。HighPart * 2 ^ 32 * 1000 + QCounter.LowPart) /频率;
结束
% - - - - - -
函数[methodinfo、结构、enuminfo] = QueryPerformance
ival ={细胞(1,0)};
fcns =结构(“名字”ival,“calltype”ival,“韩”ival,“皇家”ival,“别名”,ival);
结构= [];enuminfo = []; fcnNum = 1;
fcns.name {fcnNum} =“QueryPerformanceCounter”;
fcns.calltype {fcnNum} =“stdcall”;
fcns.LHS {fcnNum} =“int32”;
fcns.RHS {fcnNum} = {“s_ULARGE_INTEGERPtr”};
fcnNum = fcnNum + 1;
fcns.name {fcnNum} =“QueryPerformanceFrequency”;
fcns.calltype {fcnNum} =“stdcall”;
fcns.LHS {fcnNum} =“int32”;
fcns.RHS {fcnNum} = {“s_ULARGE_INTEGERPtr”};
fcnNum = fcnNum + 1;
structs.s_ULARGE_INTEGER.packing = 8;
structs.s_ULARGE_INTEGER.members =结构(“LowPart”,“uint32”,“HighPart”,“uint32”);
methodinfo = fcns;
