I added another layer of hKey lpData is char* applist is CString I got it figured out with a lot more code needed

http://www.contract-developer.tk

I use CString all the time but not the ATL::Cstring It seems to be handy with MessageBox and other system calls so I use it all the time What is the issue with CString that is bothering you?

http://www.contract-developer.tk

So I am digging around deep in the registry So far so good, but I need to access specifics for each subkey

	for (i=0; i<cSubKeys; i++)  { 
		cbName = MAX\_KEY\_LENGTH;
		retCode = RegEnumKeyEx(hKey, i,	achKey, &cbName, 0, 0, 0, &ftLastWriteTime); 
		if (retCode == ERROR\_SUCCESS) {
			RegQueryValueEx(hKey,L"DisplayName",NULL,NULL,(LPBYTE) lpData, &buffersize);
			applist.push\_back(lpData);
		}
	}

Anyway all I get are a bunch of "-" so am I getting the right type etc?

http://www.contract-developer.tk

Putting that module into a thread was to make the program able to quit at any time. This was to make sure the message handler is able to get messages.

http://www.contract-developer.tk

I am going to use MessageBox all over the module and see how far it gets. One question, do I need to use the exitthread anywhere?

http://www.contract-developer.tk

typedef std::basic_string<wchar_t> tstring; tstring output; The output will eventually be placed on the clipboard as a text

http://www.contract-developer.tk

void showtime(void *dummy) {
computing = true;
output.clear();
// Display the copyright
output = _TEXT("Fibonacci 3.0.0 for Windows");
output += ToString("/n/l/n/l");
// f << name << ": " << time << " seconds" << std::endl << std::endl;
output += _TEXT("(c) 2010-1 583727 BC Ltd, All Rights Reserved.");
// f << name << ": " << time << " seconds" << std::endl << std::endl;
output += _TEXT("/n/l");
output += _TEXT("Free for peronsal use only");
output += _TEXT("/n/l/n/l");
output += _TEXT("Commercial and government use requires a license");
output += _TEXT("/n/l/n/l");
output += _TEXT("To acheieve consistant results, run this program in the clean boot mode");
output += _TEXT("/n/l");
output += _TEXT("Click on start, in the seach box enter 'msconfig' and choose the startup tab, disable everyting and reboot");
output += _TEXT("/n/l/n/l");
output += _TEXT("RAM Performance ");
output += ToString(benchmark(fib2));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Scalar Performance ");
output += ToString(benchmark(fib3));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("CPU Large Multiplication Performance ");
output += ToString(benchmark(fib4));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
v.resize(0); t.resize(0); // free up some RAM
output += _TEXT("Recursive Factorial Performance ");
output += ToString(benchmark(RunFact));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Bubble Sort Performance ");
output += ToString(benchmark(RunBubbleSort));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Insertion Sort Performance ");
output += ToString(benchmark(RunInsertionSort));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Shell Sort Performance ");
output += ToString(benchmark(RunShellSort));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Quick Sort Performance ");
output += ToString(benchmark(RunQuickSort));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Merge Sort Performance ");
output += ToString(benchmark(RunMergeSort));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Heap Sort Performance ");
output += ToString(benchmark(RunHeapSort));
output += _TEXT(" seconds");
output += _TEXT("/n/l");
output += _TEXT("Parallel Bitwise Sieve

I have this benchmark program that started off as a console program and I am now well into making work with a window instead. It's all Win32 for the most part with a smattering of C++ STL for some containers. I launch the background via a keyboard command ctrl-r

			case (0x52): // R key
				if(GetKeyState(VK\_LCONTROL) || GetKeyState(VK\_RCONTROL)) {
					// crtl-r
					// showtime(); //?
					\_beginthread(showtime, NULL, NULL);
				}
				break;

So what am I doing wrong, the background thread runs for while but crashes, used to work fine.

http://www.contract-developer.tk

I have been working on a project for the last year and I have make much progress but much remains. First I will provide headers of what I have done so far.

template <class T> class vector; // vector space
template <class T> class matrix; // derived from vector
template <typename T> T norm(vector<T> v);
template <typename T> T norm(matrix<T> m);
template <typename T> void transpose(matrix<T> &m);
template <typename T> void conjugate_transpose(matrix<T> &m);
template <typename T> bool hermetian(matrix<T> &a, matrix<T> &b);
template <typename T> matrix<T> tensor_product(vector<T> row, vector<T> col);
template <typename T> matrix<T> minor(matrix<T> &source, int row, int col);
template <typename T> bool diagonal_dominant(matrix<T> a);
template <typename T> matrix<T> gauss_elimination(matrix<T> &m);
template <typename T> matrix<T> gauss_jordan(matrix<T> &m);
template <typename T> void gauss_siedel(matrix<T> a, vector<T> b, vector<T> x, double tolerance);

Now I can use vector and use it with math equations fine. Same for the matrix template. So given my work to make these helpful templates, I am now out of my math book big time and into the realm of advanced linear algebra. This is my Gauss elimination, and I was looking to leverage it to do LU decomposition etc to solve Ax = b more expediently.

template <typename T> matrix<T> gauss_elimination(matrix<T> &mat) {
int i = 0; int j = 0; int maxi;
matrix<T> a = mat;
int m = (int)a.size();
int n = (int)a[0].size();
while ((i < m) && (j < n)) {
maxi = i;
// sweep down the column to find the pivot element
for (int k = i + 1; k < m; k++) // partial pivoting
if (std::abs(a[k][j]) > std::abs(a[maxi][j])) maxi = k;
// eliminate
if (a[maxi][j] != 0) {
swap(a[i], a[maxi]); // row op 1, swap
a[i] = a[i] / a[i][j]; // row op 2, scale the pivot to 1
for (int u = i + 1; u < m; u++)
a[u] = a[u] - (a[i] * a[u][j]); // row op 3, zero out the column below
i++;
}
j++;
}
return a;
}

Suggestions?

http://www.contract-developer.tk

Each client will need the appropriate run-time packages for the version of Visual Studio you are using. I have a bunch of the C++ ones in my forum for my Fibonacci project. This way users can download them as needed.

http://www.contract-developer.tk

Then I guess I need to change the function a tad, so that its a drop in comparable to std::sort. I changed the code while (pos < last) { from <= That has it working now, thanks for spotting that. Now I think I will post the code on Wilipedia as the algorithm is not super clear.

http://www.contract-developer.tk

here is the call routine

void RunGnomeSort(void) {
size_t size = 10000;
std::vector<int> data;
data.resize(size);
for (int i = 0; i < size; i++)
data[i] = (int)rand();
sort::gnome(data.begin(),data.end());
//for (int i = 0; i < size; i++)
//std::cerr << data[i] << std::endl;
data.resize(0);
}

http://www.contract-developer.tk

After a few moments looking over the book I figured out that I need to watch the pointers close when using the STL method of iterators. This is not well covered in any of the books I have. No big deal, that is what forums are for. Here is what I ended up with that does work properly.

template <typename iterator> void bubble(iterator first, iterator last) { // O(n^2)
iterator i, j;
for (i = first; i != last; ++i)
for (j = first; j < i; ++j)
if (*i < *j)
std::iter_swap(i, j); // or std::swap(*i, *j);
}

Now I have this other sort that is annoying me.

template <class iterator> void gnome(iterator first, iterator last){ // O(n^2)
iterator pos = first;
iterator temp;
while (pos <= last) {
if ((pos == first) || (*pos >= *(pos-1)))
++pos;
else {
//{int tmp = ar[i]; ar[i] = ar[i-1]; ar[--i] = tmp;}
std::iter_swap(pos, pos-1);
--pos;
}
}
}

This one does not work, not sure what the problem is.

http://www.contract-developer.tk

OK I found a sort algorithm on Wikipedia, that is done like the std::sort so I figured why not convert the who lot that I have to STL. Blah. First the new sort

template<class ForwardIterator> void comb(ForwardIterator first, ForwardIterator last){
static const double shrink_factor = 1.247330950103979;
typedef typename std::iterator_traits<ForwardIterator>::difference_type difference_type;
difference_type gap = std::distance(first, last);
bool swaps = true;
while ( (gap > 1) || (swaps == true) ){
if (gap > 1)
gap = static_cast<difference_type>(gap/shrink_factor);
swaps = false;
ForwardIterator itLeft(first);
ForwardIterator itRight(first); std::advance(itRight, gap);
for ( ; itRight!=last; ++itLeft, ++itRight ){
if ( (*itRight) < (*itLeft) ){
std::iter_swap(itLeft, itRight);
swaps = true;
}
}
}
}

So now the old bubble sort

template <typename T> void bubble(std::vector<T> &a) { // O(n^2)
size_t size = a.size() - 1;
for (int pass = 0; pass < size; pass++) {
for (int n = 0; n < size; n++) {
if (a[n] > a[n+1])
std::swap(a[n], a[n+1]);
}
}
}

and this is as far as I have gone so far

template<class ForwardIterator> void bubble(ForwardIterator first, ForwardIterator last){
typedef typename std::iterator_traits<ForwardIterator>::difference_type difference_type;
difference_type size = std::distance(first, last);
}

http://www.contract-developer.tk

Looks like a glaring omission, as I have searched MSDN and found nothing, Google, no good and on Soureforge the stuff is specific to certain boards only. I guess I will have to ask users to watch the temps manually when using my benchmarks.

http://www.contract-developer.tk

If needed I guess another template/class can act as an API. When in doubt, drop down to assembler. Then the whole machine is accessible.

http://www.contract-developer.tk

I was looking for a code snippet to read the temperature of a CPU. I have a project that trashes the machine and watching the CPU temp seems prudent. I speculate there is a Winbond chip integrated into the motherboard. So the is obviously direct calls to the chip. Perhaps there is an API I can call? Direct reading of the hardware is less desirable than an API as the chip may not be the same in all machines.

http://www.contract-developer.tk

One other idea that comes to mind try { cout << mystring << endl; catch wcout << mystring << endl; }

http://www.contract-developer.tk

I tend to use assert a lot to check on things as its one way to keep code from crapping out.

http://www.contract-developer.tk

That was another waste of time. Sure seems VC 2010 is hostile to data processing.

http://www.contract-developer.tk