3rd-party/linux-kernel-module-cheat
1
0
Fork 0
mirror of https://github.com/cirosantilli/linux-kernel-module-cheat.git synced 2026-01-23 02:05:57 +01:00
Code Issues Releases Wiki Activity

readme: rerun gem5-bench-caches and take some conclusions

Browse Source
This commit is contained in:
Ciro Santilli
2018-04-20 00:15:22 +01:00
parent 759fb7b7c8
commit 05888ec5c1
3 changed files with 55 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files

91
README.adoc
View File

@@ -2681,8 +2681,8 @@ gem5:
....
./run -a aarch64 -g -E 'm5 exit'
# Or:
# ./run -a arm -g -E 'm5 exit' -- --cpu-type=HPI --caches
grep sim_insts out/aarch64/gem5/m5out/stats.txt
# ./run -a aarch64 -g -E 'm5 exit' -- --cpu-type=HPI --caches
./gem5-stat -a aarch64 sim_insts
....
Notes:
@@ -3662,37 +3662,48 @@ cat out/aarch64/gem5/bench-cache.txt
which gives:
[options="header",cols="3*>"]
|===
|test size
|cache size
|cycle count
....
n 1000
cmd ./run -a arm -g -- -r 1 --caches --l2cache --l1d_size=1024 --l1i_size=1024 --l2_size=1024 --l3_size=1024 --cpu-type=HPI --restore-with-cpu=HPI
time 24.71
exit_status 0
cycles 52386455
instructions 4555081
cmd ./run -a arm -g -- -r 1 --caches --l2cache --l1d_size=1024kB --l1i_size=1024kB --l2_size=1024kB --l3_size=1024kB --cpu-type=HPI --restore-with-cpu=HPI
time 17.44
exit_status 0
cycles 6683355
instructions 4466051
|1000
|1k
|52.432.956
n 10000
cmd ./run -a arm -g -- -r 1 --caches --l2cache --l1d_size=1024 --l1i_size=1024 --l2_size=1024 --l3_size=1024 --cpu-type=HPI --restore-with-cpu=HPI
time 52.90
exit_status 0
cycles 165704397
instructions 11531136
cmd ./run -a arm -g -- -r 1 --caches --l2cache --l1d_size=1024kB --l1i_size=1024kB --l2_size=1024kB --l3_size=1024kB --cpu-type=HPI --restore-with-cpu=HPI
time 36.19
exit_status 0
cycles 16182925
instructions 11422585
|1000
|1M
|6.328.325
n 100000
cmd ./run -a arm -g -- -r 1 --caches --l2cache --l1d_size=1024 --l1i_size=1024 --l2_size=1024 --l3_size=1024 --cpu-type=HPI --restore-with-cpu=HPI
time 325.09
exit_status 0
cycles 1295703657
instructions 81189411
cmd ./run -a arm -g -- -r 1 --caches --l2cache --l1d_size=1024kB --l1i_size=1024kB --l2_size=1024kB --l3_size=1024kB --cpu-type=HPI --restore-with-cpu=HPI
time 250.74
exit_status 0
cycles 110585681
instructions 80899588
....
|10000
|1k
|141.637.834
We make the following conclusions:
|10000
|1M
|16.969.057
|100000
|1k
|1.034.500.724
|100000
|1M
|121.728.035
|===
* the number of instructions almost does not change: the CPU is waiting for memory all the extra time. TODO: why does it change at all?
* the wall clock execution time is not directionally proportional to the number of cycles: here we had a 10x cycle increase, but only 2x time increase
===== gem5 memory latency
@@ -4892,43 +4903,43 @@ Output on <<p51>> fb317f4778633692b91c9174224dccc6a3a02893:
....
cmd ./run -a x86_64 -E '/poweroff.out'
time 4.00
status 0
exit_status 0
cmd ./run -a x86_64 -E '/poweroff.out' -K
time 1.32
status 0
exit_status 0
cmd ./run -a x86_64 -E '/poweroff.out' -T exec_tb
time 5.00
status 0
exit_status 0
instructions 7014180
cmd ./run -a x86_64 -E 'm5 exit' -g
time 389.39
status 0
exit_status 0
instructions 645396324
cmd ./run -a arm -E '/poweroff.out'
time 1.82
status 0
exit_status 0
cmd ./run -a arm -E '/poweroff.out' -T exec_tb
time 2.11
status 0
exit_status 0
instructions 687532
cmd ./run -a arm -E 'm5 exit' -g
time 88.59
status 0
exit_status 0
instructions 142359498
cmd ./run -a aarch64 -E '/poweroff.out'
time 1.34
status 0
exit_status 0
cmd ./run -a aarch64 -E '/poweroff.out' -T exec_tb
time 1.35
status 0
exit_status 0
instructions 179564
cmd ./run -a aarch64 -E 'm5 exit' -g
time 58.18
status 0
exit_status 0
instructions 116389714
cmd ./run -a aarch64 -E 'm5 exit' -g -- --cpu-type=HPI --caches --l2cache --l1d_size=1024kB --l1i_size=1024kB --l2_size=1024kB --l3_size=1024kB
time 355.81
status 0
exit_status 0
instructions 116508439
....