Better messages

Author: joseemoreira

Code/matmul/memory.cc

@@ -19,6 +19,7 @@
 #define xxpermdi(XT, XA, XB, IM)	asm("xxpermdi " #XT "," #XA "," #XB "," #IM)
 #define xxmfacc(XT)			asm("xxmfacc " #XT)
 #define xvadddp(XT, XA, XB)		asm("xvadddp " #XT "," #XA "," #XB)
+#define nop()				asm("nop ")
 
 void memory_load_1KiB
 (
@@ -27,6 +28,7 @@ void memory_load_1KiB
 )
 {
     asm("mtctr 4");
+    asm("nop");
     asm("LOOP17:");
 
     lxvp( 0+ 0, 3,   0+  0); lxvp( 0+ 2, 3,  32+  0);
@@ -83,8 +85,7 @@ double run_kernel
 #pragma omp parallel for
     for (int j=0; j<nthreads; j++)
     {
-	uint32_t reps = count;
-	for(; reps; reps--)
+	for(uint32_t reps = count; reps; reps--)
 	{
 	    kernel(A[j], r);
 	}
@@ -112,7 +113,7 @@ void run_kernel_and_report
     volatile double GB = (1.0e-09)*N*sizeof(double)*count*omp_get_max_threads();
     elapsed = run_kernel(kernel, count, N, n);
     std::cout << std::setprecision(6);
-    std::cout << "Time to run " << std::setw(30) << name << " (" << std::setw(9) << N << ") " << std::setw(6) << count << " times = " << std::setw(10) << std::fixed << elapsed << " seconds (" << std::setw(10) << std::scientific << GB/elapsed << " GB/s)" << std::endl;
+    std::cout << "Time to run " << std::setw(30) << name << " (" << std::setw(9) << N << " doubles) " << std::setw(9) << count << " times = " << std::setw(10) << std::fixed << elapsed << " seconds (" << std::setw(10) << std::scientific << GB/elapsed << " GB/s)" << std::endl;
 }
 
 #define RUN_KERNEL(kernel, count, N, n) run_kernel_and_report(kernel, count, #kernel, N, n)
@@ -127,8 +128,9 @@ int main
 
     volatile double elapsed;
 
+    int nthreads = omp_get_max_threads();
     std::cout << "=========================================================================================================================" << std::endl;
-    std::cout << "Running on " << omp_get_max_threads() << " threads" << std::endl;
+    std::cout << "Running on " << nthreads << ((nthreads > 1) ? " threads" : " thread") << std::endl;
 
     for (uint32_t i=1; i<1000000; i *= 2) 
 	RUN_KERNEL(memory_load_1KiB, memory_load_count/i, i*1024, 128);