According to data manual (SPRS582B, sec 7.12.3) it seems that tw(HSTBL) + tw(HSTBH) are the specifications that define how fast the interface can operate. The value for 1.2 GHz CPU is 25 ns or 40 MHz which in HPI32 mode gives 160 MB/sec theoretical maximum. Is this a correct interpretation?
Is there any benchmark information on HPI throughput for real world implementation? Would like to see how close to theoretical max throughput is practical.
Thanks.