Hi Giraph user,
I want to get timings for communication and compute within a superstep. I
ran SSSP example with -Dgiraph.metrics.enable=true . I got following log in
master thread
--- METRICS: superstep 334 ---
*superstep time*
mean: 236.875 ms
slowest: 241 ms from orion-05.local_39
fastest: 215 ms from orion-03.local_10
*compute all partitions*
mean: 123.65 ms
slowest: 157 ms from orion-03.local_25
fastest: 101 ms from orion-01.local_18
*network communication time*
mean: 0.0 ms
slowest: 0 ms from orion-03.local_5
fastest: 0 ms from orion-03.local_5
time to first message
mean: 0.0 us
slowest: 0 us from orion-03.local_5
fastest: 0 us from orion-03.local_5
wait requests time
mean: 727.25 us
slowest: 4212 us from orion-03.local_15
fastest: 164 us from orion-04.local_32
I have a doubt : is compute all partitions + network communication time +
synchronization overhead =superstep time or is network communication time
is included in the compute time itself?
for worker threads I got following log
--- METRICS: superstep 1028 ---
superstep time: 250 ms
compute all partitions: 119 ms
network communication time: 0 ms
time to first message: 0 us
wait on requests time: 109 us
4/11/15 9:16:33 AM
=============================================================
giraph.superstep.1028:
communication-time-ms:
value = 0
compute-all-ms:
value = 119
compute-per-partition-ms:
sum = 113.00
min = 2.00
max = 5.00
mean = 2.83
stddev = 0.64
median = 3.00
75% <= 3.00
95% <= 4.00
98% <= 5.00
99% <= 5.00
99.9% <= 5.00
count = 40
local-requests:
count = 1
message-bytes-sent:
count = 8305
messages-sent:
count = 321
percent-local-requests:
value = 2.5
Can some one point out what exactly these superstep time,compute all
partitions, network communication time exactly mean?
Thanks
Ravikant
