search

andikleen / pmu-tools

Intel PMU profiling tools
GNU General Public License v2.0
2.04k stars 340 forks source link

toplev --globals omits <== marking of critical nodes #446

Closed aayasin closed 1 year ago

aayasin commented 1 year ago

Toplev --global reproducer on my TGL:

$ ./pmu-tools/toplev.py --no-desc --quiet --global -- ./pmu-tools/workloads/BC1s
# 4.5-full-perf on 11th Gen Intel(R) Core(TM) i7-11700B @ 3.20GHz [tgl/icelake]
allFE             Frontend_Bound                      % Slots                      15.4   [29.7%]
allBAD            Bad_Speculation                     % Slots                      31.3   [29.7%]
allFE             Frontend_Bound.Fetch_Bandwidth      % Slots                      10.2   [29.7%]
allBAD            Bad_Speculation.Branch_Mispredicts  % Slots                      31.3   [29.7%]

$ ./pmu-tools/toplev.py --no-desc --quiet  -- ./pmu-tools/workloads/BC1s
# 4.5-full-perf on 11th Gen Intel(R) Core(TM) i7-11700B @ 3.20GHz [tgl/icelake]
FE             Frontend_Bound                      % Slots                      15.3   [30.0%]
BAD            Bad_Speculation                     % Slots                      31.7   [30.0%]
FE             Frontend_Bound.Fetch_Bandwidth      % Slots                      10.1   [30.0%]
BAD            Bad_Speculation.Branch_Mispredicts  % Slots                      31.6   [30.0%]<==

perf-tools$ make lspmu
icelake
Model name:                      11th Gen Intel(R) Core(TM) i7-11700B @ 3.20GHz

dungeon@CPIE-CH-104064:~/ayasin/perf-tools$ head setup-system.log
Linux CPIE-CH-104064 5.15.0-52-generic #58-Ubuntu SMP Thu Oct 13 08:03:55 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
PRETTY_NAME="Ubuntu 22.04.1 LTS"
NAME="Ubuntu"
VERSION_ID="22.04"
VERSION="22.04.1 LTS (Jammy Jellyfish)"
ID=ubuntu
/sys/kernel/mm/transparent_hugepage/enabled : always [madvise] never
kernel.randomize_va_space = 2
vm.nr_hugepages = 0
vm.nr_overcommit_hugepages = 0

Maybe or may not be related to https://github.com/andikleen/pmu-tools/issues/444