Simulator stats
Select a stat to view. One chart at a time.
Other
Total GPU cycles (incl. kernel-launch latency)
Instructions executed (thread-level)
Warp instructions issued
IPC, instructions per cycle (whole run)
Achieved occupancy, percent (last kernel)
L1 global reads: accesses
L1 global reads: hits
L1 global reads: merged into an outstanding miss (MSHR hits)
L1 global writes: accesses
L1 global writes: hits
L2 global reads: accesses
L2 global reads: hits
L2 global writes: accesses
L2 global writes: hits
L2 global writes: allocated a new line
L2 local-memory writes: hits
L2 local-memory writes: misses
L2 local-memory writes: sector misses
L2 local-memory writes: allocated a new line
L2 cross-chiplet reads: accesses
L2 cross-chiplet reads: hits
L2 bandwidth, GB per second (last kernel)
DRAM reads (requests)
DRAM writes (requests)
Shared-memory bank conflicts
L1 bank conflicts
Simulator wall time (seconds)
Simulation speed (GPU cycles per real second)
Simulation speed (instructions per real second)
Slowdown vs real silicon (times)