I am writing a tool that uses the Pentium 4 performance counters to measure events. I want to measure the DTLB miss rate. The performance counter documentation lists the DTLB_load_miss and DTLB_store_miss events in table A5. I also need to know the total number of DTLB accesses to calculate the percent of DTLB load misses. I do not see an event for all DTLB references. What event or events should I use to count all DTLB references?
For more complete information about compiler optimizations, see our Optimization Notice.