This is a benchmark for the Bristol University Docking Engine. A very flat profile with almost 0% time spent in front-end misses and very low L2 access and as a result high average GHz. The time bottleneck seems to be backend CPU cycles at least on AMD.
AMD metrics. The amount of frontend time is surprisingly low (0.1%), the backend CPU is higher than average. This is a floating point code that does much better on my AMD processor than Intel processor.
elapsed 860.823
on_cpu 0.959 # 15.34 / 16 cores
utime 13199.154
stime 4.191
nvcsw 3153 # 1.55%
nivcsw 200170 # 98.45%
inblock 9416 # 10.94/sec
onblock 1472 # 1.71/sec
cpu-clock 13203663272263 # 13203.663 seconds
task-clock 13203751860729 # 13203.752 seconds
page faults 149139 # 11.295/sec
context switches 207434 # 15.710/sec
cpu migrations 327 # 0.025/sec
major page faults 9 # 0.001/sec
minor page faults 149130 # 11.295/sec
alignment faults 0 # 0.000/sec
emulation faults 0 # 0.000/sec
branches 1812286189855 # 36.357 branches per 1000 inst
branch misses 3373308787 # 0.19% branch miss
conditional 1736412032224 # 34.835 conditional branches per 1000 inst
indirect 347384335 # 0.007 indirect branches per 1000 inst
cpu-cycles 56442460721050 # 4.11 GHz
instructions 49855526461212 # 0.88 IPC
slots 112896497428296 #
retiring 16597169840414 # 14.7% (18.6%)
-- ucode 182636319 # 0.0%
-- fastpath 16596987204095 # 14.7%
frontend 133841423134 # 0.1% ( 0.1%)
-- latency 103818842496 # 0.1%
-- bandwidth 30022580638 # 0.0%
backend 72647925519692 # 64.3% (81.2%)
-- cpu 72220434205857 # 64.0%
-- memory 427491313835 # 0.4%
speculation 52940901992 # 0.0% ( 0.1%)
-- branch mispredict 52901934851 # 0.0%
-- pipeline restart 38967141 # 0.0%
smt-contention 23464406142871 # 20.8% ( 0.0%)
cpu-cycles 56319475365345 # 4.12 GHz
instructions 49846507153415 # 0.89 IPC
instructions 16619408448807 # 0.564 l2 access per 1000 inst
l2 hit from l1 8974505532 # 0.41% l2 miss
l2 miss from l1 25839841 #
l2 hit from l2 pf 387375438 #
l3 hit from l2 pf 8204553 #
l3 miss from l2 pf 4449609 #
instructions 16612896802749 # 90.745 float per 1000 inst
float 512 79 # 0.000 AVX-512 per 1000 inst
float 256 1115000285842 # 67.117 AVX-256 per 1000 inst
float 128 392531488237 # 23.628 AVX-128 per 1000 inst
float MMX 0 # 0.000 MMX per 1000 inst
float scalar 0 # 0.000 scalar per 1000 inst
Intel metrics. Somewhat of a contrast including more frontend latency and a lower IPC. It would be interesting to see the mix of floating point. The on-core amount is also less.
elapsed 2532.034
on_cpu 0.768 # 12.29 / 16 cores
utime 31128.271
stime 1.284
nvcsw 23228 # 10.24%
nivcsw 203670 # 89.76%
inblock 9776 # 3.86/sec
onblock 1464 # 0.58/sec
cpu-clock 31128499474749 # 31128.499 seconds
task-clock 31128588659384 # 31128.589 seconds
page faults 144833 # 4.653/sec
context switches 239360 # 7.689/sec
cpu migrations 31310 # 1.006/sec
major page faults 62 # 0.002/sec
minor page faults 144770 # 4.651/sec
alignment faults 0 # 0.000/sec
emulation faults 0 # 0.000/sec
branches 2938735691997 # 29.701 branches per 1000 inst
branch misses 63751468128 # 2.17% branch miss
conditional 2938735707069 # 29.701 conditional branches per 1000 inst
indirect 702385471417 # 7.099 indirect branches per 1000 inst
slots 191258364400442 #
retiring 78324202863149 # 41.0% (41.0%)
-- ucode 9916757825840 # 5.2%
-- fastpath 68407445037309 # 35.8%
frontend 33150398566636 # 17.3% (17.3%)
-- latency 31618399330327 # 16.5%
-- bandwidth 1531999236309 # 0.8%
backend 76230829114597 # 39.9% (39.9%)
-- cpu 70409289474916 # 36.8%
-- memory 5821539639681 # 3.0%
speculation 3245583067706 # 1.7% ( 1.7%)
-- branch mispredict 3245437890649 # 1.7%
-- pipeline restart 145177057 # 0.0%
smt-contention 0 # 0.0% ( 0.0%)
cpu-cycles 104562031071524 # 2.59 GHz
instructions 140985466084116 # 1.35 IPC
l2 access 11979732638 # 0.167 l2 access per 1000 inst
l2 miss 263932623 # 2.20% l2 miss
Process structure is straighforward
454 processes
96 bude 210888.16 15.80
64 clinfo 11.20 4.48
38 vulkaninfo 0.76 1.14
6 glxinfo:gdrv0 0.11 0.05
4 vulkani:disk$0 0.08 0.12
6 php 0.07 0.09
2 glxinfo 0.06 0.03
2 glxinfo:cs0 0.06 0.03
2 glxinfo:disk$0 0.06 0.03
2 glxinfo:sh0 0.06 0.03
2 glxinfo:shlo0 0.06 0.03
2 llvmpipe-0 0.04 0.06
2 llvmpipe-1 0.04 0.06
2 llvmpipe-10 0.04 0.06
2 llvmpipe-11 0.04 0.06
2 llvmpipe-12 0.04 0.06
2 llvmpipe-13 0.04 0.06
2 llvmpipe-14 0.04 0.06
2 llvmpipe-15 0.04 0.06
2 llvmpipe-2 0.04 0.06
2 llvmpipe-3 0.04 0.06
2 llvmpipe-4 0.04 0.06
2 llvmpipe-5 0.04 0.06
2 llvmpipe-6 0.04 0.06
2 llvmpipe-7 0.04 0.06
2 llvmpipe-8 0.04 0.06
2 llvmpipe-9 0.04 0.06
6 clang 0.04 0.05
1 lspci 0.00 0.03
91 sh 0.00 0.00
13 gcc 0.00 0.00
11 gsettings 0.00 0.00
9 stty 0.00 0.00
8 stat 0.00 0.00
8 systemd-detect- 0.00 0.00
6 llvm-link 0.00 0.00
6 minibude 0.00 0.00
5 phoronix-test-s 0.00 0.00
4 gmain 0.00 0.00
2 cc 0.00 0.00
2 lscpu 0.00 0.00
2 uname 0.00 0.00
2 which 0.00 0.00
2 xset 0.00 0.00
1 date 0.00 0.00
1 dconf worker 0.00 0.00
1 dirname 0.00 0.00
1 dmesg 0.00 0.00
1 dmidecode 0.00 0.00
1 grep 0.00 0.00
1 ifconfig 0.00 0.00
1 ip 0.00 0.00
1 lsmod 0.00 0.00
1 mktemp 0.00 0.00
1 ps 0.00 0.00
1 qdbus 0.00 0.00
1 readlink 0.00 0.00
1 realpath 0.00 0.00
1 sed 0.00 0.00
1 sort 0.00 0.00
1 systemctl 0.00 0.00
1 template.sh 0.00 0.00
1 wc 0.00 0.00
1 xrandr 0.00 0.00
0 processes running
47 maximum processes
The core parts start one on each core
283675) minibude cpu=6 start=46.75 finish=83.83
283676) bude cpu=10 start=46.75 finish=83.82
283677) bude cpu=2 start=46.76 finish=83.82
283678) bude cpu=3 start=46.76 finish=83.82
283679) bude cpu=4 start=46.76 finish=83.82
283680) bude cpu=9 start=46.76 finish=83.82
283681) bude cpu=0 start=46.76 finish=83.82
283682) bude cpu=5 start=46.76 finish=83.82
283683) bude cpu=6 start=46.76 finish=83.82
283684) bude cpu=1 start=46.76 finish=83.82
283685) bude cpu=11 start=46.76 finish=83.82
283686) bude cpu=12 start=46.76 finish=83.82
283687) bude cpu=8 start=46.76 finish=83.82
283688) bude cpu=9 start=46.76 finish=83.82
283689) bude cpu=13 start=46.76 finish=83.82
283690) bude cpu=14 start=46.76 finish=83.82
283691) bude cpu=7 start=46.76 finish=83.82