Performance monitoring of mulss and imul on SMT

Software Tuning, Performance Optimization & Platform Monitoring

Discussion regarding monitoring and software tuning methodologies, Performance Monitoring Unit (PMU) of Intel microprocessors, and platform updating.

Performance monitoring of mulss and imul on SMT

464 Views

I am trying to understand port utilization of sandy-bridge while running multiplication.

I am running three versions of multiplication, in one version 2 sibling SMT thread is running floating point multiplication(mulss)(case 1)(port 0). Another version performs integer multiplication(imul)(case 2)(port 1) in sibling SMT threads and in final version, one sibling SMT thread is running mulss(port 0) and another thread running imul(case 3)(port 1). When I measure port utilization of port 0 and 1 using UOPS_DISPATCHED_PORT it seems that port 0 and 1 utilization is similar for case 1 and case 3. But it was expected that port 1 should be more utilized in case 3 compared to case 1 as port 1 performs imul operation.

UOPS_DISPATCHED_PORT:PORT_1 measures cycles per thread, does it mean it can observe only one thread and cannot report about the other sibling SMT thread?

Link Copied

0 Replies

Community support is provided Monday to Friday. Other contact methods are available here.

Intel does not verify all solutions, including but not limited to any file transfers that may appear in this community. Accordingly, Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or usage in trade.

For more complete information about compiler optimizations, see our Optimization Notice.