Data lines

To force bit flips i n data lines, the test uses two sets of static data, a pseudo-random set and a fixed-pattern set. Each set contains 31 elements-a prime number. The test uses a prime number of elements to avoid coinciding with binary math boundaries. The sets are unique to each addressing mode so as to occupy the full range of bits.

• The test uses the pseudo-random data set to stress the data lines in a repeatable but pattern-less fashion.
• The test uses the fixed-pattern set to force significant numbers of data bits to flip from one access to the next.

The sub-tests execute similarly in that each subtest iterates through static data, writing values to memory. The test combines the three address line approaches with the two data sets to produce six unique sub-tests:

• Sequential with Random Data
• Sequential with Fixed Pattern Data
• Full Range Converging with Random Data
• Full Range Converging with Fixed Pattern Data
• Maximum Invert Convergence with Random Data
• Maximum Invert Convergence with Fixed Pattern Data