s00601.nersc.gov.197008.0
powered by IPM
command: ./a3.out
codename: A3_important
state: completed
username: wanda
group: mpccc
host: s01409 (006006294C00_AIX)
mpi_tasks: 1024 on 64 hosts
start: 02/02/05/09:47:10
wallclock: 7.18162e+02 sec
stop: 02/02/05/09:59:04
%comm: 28.4401624498546
max memory: 63.8729528 gbytes
total gflop/sec: 65.14704
switch(send): 863.826758228242 gbytes
switch(recv): 863.644127734005 gbytes

Computation

Event Count Pop
PM_CYC 262179900846682 *
PM_FPU0_CMPL 16689229674229 *
PM_FPU1_CMPL 10902407880206 *
PM_FPU_FMA 19194494702948 *
PM_INST_CMPL 158313092396780 *
PM_LD_CMPL 54175636509519 *
PM_ST_CMPL 34614909037689 *
PM_TLB_MISS 168005302522 *

Communication

% of MPI Time
Compute Event Statistics
  Ntasks Avg Min Max
PM_CYC * 256035059420.59 254601209559 (597) 256546541925 (359)
PM_FPU0_CMPL * 16298075853.74 16175702110 (118) 16484412911 (0)
PM_FPU1_CMPL * 10646882695.51 10531428188 (959) 10765377343 (896)
PM_FPU_FMA * 18744623733.35 18741516233 (908) 18927736447 (64)
PM_INST_CMPL * 154602629293.73 133806636299 (597) 160458095051 (194)
PM_LD_CMPL * 52905895028.83 45652061473 (597) 54950626501 (194)
PM_ST_CMPL * 33803622107.12 28777685927 (597) 35216724917 (610)
PM_TLB_MISS * 164067678.24 154798056 (510) 617533675 (733)
Communication Event Statistics (100.00% detail, 3.5445e-02 error)
  Buffer Size Ncalls Total Time Min Time Max Time %MPI %Wall
MPI_Allreduce 8 3278848 124132.547 0.000 120.130 59.35 16.88
MPI_Comm_rank 0 -1 43439.102 0.000 43.350 20.77 5.91
MPI_Wait 98304 13221888 15710.953 0.000 11.551 7.51 2.14
MPI_Wait 196608 13221888 5331.236 0.000 3.611 2.55 0.72
MPI_Wait 589824 206848 5166.272 0.000 7.764 2.47 0.70
MPI_Issend 196608 6610944 4556.737 0.000 2.959 2.18 0.62
MPI_Issend 98304 6610944 3395.876 0.000 2.571 1.62 0.46
MPI_Wait 1310720 258048 3252.752 0.000 0.295 1.56 0.44
MPI_Bcast 588 3072 1553.856 0.000 8.321 0.74 0.21
MPI_Wait 1179648 202752 1536.354 0.000 5.934 0.73 0.21
MPI_Irecv 98304 6610944 443.843 0.000 0.160 0.21 0.06
MPI_Allreduce 4 11264 344.220 0.000 0.333 0.16 0.05
 
Load balance by task: HPM counters
by MPI rank, by MPI time
Load balance by task: memory, flops, timings
by MPI rank, by MPI time
 
Communication balance by task (sorted by MPI rank)
Communication balance by task (sorted by MPI time)
 
Communication Topology : point to point data flow
time spent , ~rate , map_data file map_adjacency file
Communication Topology : Connectivity (sorted by # neighbors)
 
Message Buffer Size Distributions: time
abs, hist
Message Buffer Size Distributions: Ncalls
abs, hist
Message Buffer Size Distributions: data volume
abs, hist
Switch volume by node
 
Memory usage by node




powered by IPM