-int input_available()
-{
- static HANDLE inh = NULL;
- static bool usePipe = false;
- INPUT_RECORD rec[256];
- DWORD nchars, recCnt;
-
- if (!inh)
- {
- inh = GetStdHandle(STD_INPUT_HANDLE);
- if (GetConsoleMode(inh, &nchars))
- {
- SetConsoleMode(inh, nchars & ~(ENABLE_MOUSE_INPUT | ENABLE_WINDOW_INPUT));
- FlushConsoleInputBuffer(inh);
- } else
- usePipe = true;
- }
-
- // When using Standard C input functions, also check if there
- // is anything in the buffer. After a call to such functions,
- // the input waiting in the pipe will be copied to the buffer,
- // and the call to PeekNamedPipe can indicate no input available.
- // Setting stdin to unbuffered was not enough. [from Greko]
- if (stdin->_cnt > 0)
- return 1;
-
- // When running under a GUI the input commands are sent to us
- // directly over the internal pipe. If PeekNamedPipe() returns 0
- // then something went wrong. Probably the parent program exited.
- // Returning 1 will make the next call to the input function
- // return EOF, where this should be catched then.
- if (usePipe)
- return PeekNamedPipe(inh, NULL, 0, NULL, &nchars, NULL) ? nchars : 1;
-
- // Count the number of unread input records, including keyboard,
- // mouse, and window-resizing input records.
- GetNumberOfConsoleInputEvents(inh, &nchars);
-
- // Read data from console without removing it from the buffer
- if (nchars <= 0 || !PeekConsoleInput(inh, rec, Min(nchars, 256), &recCnt))
- return 0;
-
- // Search for at least one keyboard event
- for (DWORD i = 0; i < recCnt; i++)
- if (rec[i].EventType == KEY_EVENT)
- return 1;
-
- return 0;
+/// get_group() retrieves logical processor information using Windows specific
+/// API and returns the best group id for the thread with index idx. Original
+/// code from Texel by Peter Ă–sterlund.
+
+int get_group(size_t idx) {
+
+ int threads = 0;
+ int nodes = 0;
+ int cores = 0;
+ DWORD returnLength = 0;
+ DWORD byteOffset = 0;
+
+ // Early exit if the needed API are not available at runtime
+ HMODULE k32 = GetModuleHandle("Kernel32.dll");
+ if ( !GetProcAddress(k32, "GetLogicalProcessorInformationEx")
+ || !GetProcAddress(k32, "GetNumaNodeProcessorMaskEx")
+ || !GetProcAddress(k32, "SetThreadGroupAffinity"))
+ return -1;
+
+ // First call to get returnLength. We expect it to fail due to null buffer
+ if (GetLogicalProcessorInformationEx(RelationAll, nullptr, &returnLength))
+ return -1;
+
+ // Once we know returnLength, allocate the buffer
+ SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *buffer, *ptr;
+ ptr = buffer = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)malloc(returnLength);
+
+ // Second call, now we expect to succeed
+ if (!GetLogicalProcessorInformationEx(RelationAll, buffer, &returnLength))
+ {
+ free(buffer);
+ return -1;
+ }
+
+ while (ptr->Size > 0 && byteOffset + ptr->Size <= returnLength)
+ {
+ if (ptr->Relationship == RelationNumaNode)
+ nodes++;
+
+ else if (ptr->Relationship == RelationProcessorCore)
+ {
+ cores++;
+ threads += (ptr->Processor.Flags == LTP_PC_SMT) ? 2 : 1;
+ }
+
+ byteOffset += ptr->Size;
+ ptr = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)(((char*)ptr) + ptr->Size);
+ }
+
+ free(buffer);
+
+ std::vector<int> groups;
+
+ // Run as many threads as possible on the same node until core limit is
+ // reached, then move on filling the next node.
+ for (int n = 0; n < nodes; n++)
+ for (int i = 0; i < cores / nodes; i++)
+ groups.push_back(n);
+
+ // In case a core has more than one logical processor (we assume 2) and we
+ // have still threads to allocate, then spread them evenly across available
+ // nodes.
+ for (int t = 0; t < threads - cores; t++)
+ groups.push_back(t % nodes);
+
+ // If we still have more threads than the total number of logical processors
+ // then return -1 and let the OS to decide what to do.
+ return idx < groups.size() ? groups[idx] : -1;