[ros-dev] Re:[ros-svn] [gdalsnes] 18113:-reorderInsertXscendingOrdermacro
argument order and update uses
Jakob Eriksson
jakov at vmlinux.org
Wed Sep 28 23:02:53 CEST 2005
Take a look at the code below. More readable by far IMHO.
Casper Hornstrup wrote:
>
>
>>-----Original Message-----
>>From: ros-dev-bounces at reactos.com [mailto:ros-dev-bounces at reactos.com] On Behalf Of Thomas
>>Weidenmueller
>>Sent: 28. september 2005 17:27
>>To: ReactOS Development List
>>Subject: Re: [ros-dev] Re:[ros-svn] [gdalsnes] 18113:-reorderInsertXscendingOrdermacro argument
>>order and update uses
>>
>>Casper Hornstrup wrote:
>>
>>
>>>
>>>
>>>>The alternative is: do the cleanup at every return, use goto or use
>>>>try/finally.
>>>>1)Cleanup at every return is madness. Most functions in ros do a large
>>>>amount of cleanup at each return and I sometimes spot mistakes where
>>>>one/more return misses some cleanup. Those errors are _hard_ to find.
>>>>
>>>>
>>>The functions are too large then. Use more smaller functions instead.
>>>
>>>
>>I agree with Nathan. Having tons of small functions often isn't a good
>>solution, especially when you'd have to create dozens of small helper
>>functions. That not just only generates slower code but also makes it
>>more difficult to get a picture of the algorithm used. I much more
>>prefer jumping to cleanup labels the way Nathan demonstrated it. Of
>>course I avoid it where it doesn't make sense.
>>
>>- Thomas
>>
>>
>
>There is practically no difference in execution speed.
>
>Casper
>
>
>
>------------------------------------------------------------------------
>
>#include <stdio.h>
>#include <windows.h>
>
>#define NumberOfIterations 60000
>
>void RunCollapsed()
>{
> int i, j;
> for (i = 0; i < NumberOfIterations; i++)
> {
> for (j = 0; j < NumberOfIterations; j++)
> {
> }
> }
>}
>
>static void Helper()
>{
> int j;
> for (j = 0; j < NumberOfIterations; j++)
> {
> }
>}
>
>void RunNonCollapsed()
>{
> int i;
> for (i = 0; i < NumberOfIterations; i++)
> {
> Helper();
> }
>}
>
>void RunOnce(LARGE_INTEGER *collapsedTicks, LARGE_INTEGER *nonCollapsedTicks)
>{
> LARGE_INTEGER start;
> LARGE_INTEGER stop;
>
> QueryPerformanceCounter(&start);
> RunCollapsed();
> QueryPerformanceCounter(&stop);
> collapsedTicks->QuadPart = stop.QuadPart - start.QuadPart;
> printf("Collapsed: %lu ticks\n", collapsedTicks->QuadPart);
>
> QueryPerformanceCounter(&start);
> RunNonCollapsed();
> QueryPerformanceCounter(&stop);
> nonCollapsedTicks->QuadPart = stop.QuadPart - start.QuadPart;
> printf("NonCollapsed: %lu ticks\n", nonCollapsedTicks->QuadPart);
>
> printf("%d%%\n", (nonCollapsedTicks->QuadPart * 100) / collapsedTicks->QuadPart);
>}
>
>int main()
>{
> LARGE_INTEGER collapsedTicks;
> LARGE_INTEGER nonCollapsedTicks;
> LARGE_INTEGER totalCollapsedTicks;
> LARGE_INTEGER totalNonCollapsedTicks;
> int i;
>
> totalCollapsedTicks.QuadPart = 0;
> totalNonCollapsedTicks.QuadPart = 0;
> for (i = 0; i < 10; i++)
> {
> RunOnce(&collapsedTicks, &nonCollapsedTicks);
> totalCollapsedTicks.QuadPart += collapsedTicks.QuadPart;
> totalNonCollapsedTicks.QuadPart += nonCollapsedTicks.QuadPart;
> }
> printf("Average: %d%%\n", (totalNonCollapsedTicks.QuadPart * 100) / totalCollapsedTicks.QuadPart);
>}
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>Ros-dev mailing list
>Ros-dev at reactos.com
>http://reactos.com:8080/mailman/listinfo/ros-dev
>
More information about the Ros-dev
mailing list