New Contributor
Frequent ContributorI checked your code with 16.1.2; all loops are reported to have an II of one while the loop on "x" is limited to 16 "threads" (i.e. every access to the local buffers in the loop will be stalled at least once every 16 iterations) to limit on-chip memory replication. This information, which is VERY important, seems to have been removed in the newer versions of report. I think what the new report is trying to say something along the lines of "II is an approximation due to stallable accesses/inner loops" or something like that.
