Correct inf loop defect in fast clock init.

Each time delay is grown we need to reset elapsed.  We want to take
the most accurate sample from the set at fixed delay.

Without this we will hang if there is ever an insufficiently accurate,
high unit clock read.

Change-Id: Ic65f364067789ac85a6572d67af2d77528e265bb


[ROCm/ROCR-Runtime commit: 4e9849034d]

projects/rocr-runtime/runtime/hsa-runtime/core/util/timer.cpp

@@ -57,9 +57,11 @@ fast_clock::init::init() {
 
   // calibrate clock
   fast_clock::raw_rep min = 0;
-  clock::duration elapsed = clock::duration::max();
+  clock::duration elapsed;
 
   do {
+    elapsed = clock::duration::max();
+
     for (int t = 0; t < 10; t++) {
       fast_clock::raw_rep r1, r2;
       clock::time_point t0, t1, t2, t3;