Open Eboubaker opened 3 months ago
For now i had to modify the qemu run command
mem=$(free -m | grep -oP '\d+' | head -n 1)
cpus=$(nproc)
qemu-system-x86_64 -machine q35 -m $mem -smp cpus=$cpus -cpu qemu64 -drive if=pflash,format=raw,read-only=on,file=$PREFIX/share/qemu/edk2-x86_64-code.fd -netdev user,id=n1,dns=8.8.8.8,hostfwd=tcp::2222-:22 -device virtio-net,netdev=n1 -nographic alpine.img
I wanted you to change it in the readme to be the default, the emulation was very slow before i changed this.
Thanks for the feedback! While increasing resources can improve performance for some users, it's not necessary for everyone. The default configuration aims for wider compatibility. We can definitely add your script as an optional optimization in the README for users who want to utilize more resources. Would you like to contribute by submitting a pull request with this information?
Termux bench
~/sysbench $ sysbench cpu run
sysbench 1.1.0 (using bundled LuaJIT 2.1.0-beta3)
Running the test with following options:
Number of threads: 1
Initializing random number generator from current time
Prime numbers limit: 10000
Initializing worker threads...
Threads started!
CPU speed:
events per second: 753647.69
Throughput:
events/s (eps): 753647.6891
time elapsed: 10.0008s
total number of events: 7537057
Latency (ms):
min: 0.00
avg: 0.00
max: 0.33
95th percentile: 0.00
sum: 2695.12
Threads fairness:
events (avg/stddev): 7537057.0000/0.00
execution time (avg/stddev): 2.6951/0.00
Default docker
alpine:~# sysbench cpu run
sysbench 1.0.20 (using system LuaJIT 2.1.1710398010)
Running the test with following options:
Number of threads: 1
Initializing random number generator from current time
Prime numbers limit: 10000
Initializing worker threads...
Threads started!
CPU speed:
events per second: 31.17
General statistics:
total time: 10.0382s
total number of events: 315
Latency (ms):
min: 27.69
avg: 31.26
max: 49.39
95th percentile: 39.65
sum: 9847.27
Threads fairness:
events (avg/stddev): 315.0000/0.00
execution time (avg/stddev): 9.8473/0.00
Docker with unlocked cpu and mem
alpine:~# sysbench cpu run
sysbench 1.0.20 (using system LuaJIT 2.1.1710398010)
Running the test with following options:
Number of threads: 1
Initializing random number generator from current time
Prime numbers limit: 10000
Initializing worker threads...
Threads started!
CPU speed:
events per second: 34.42
General statistics:
total time: 10.0125s
total number of events: 347
Latency (ms):
min: 27.04
avg: 28.50
max: 40.94
95th percentile: 30.26
sum: 9890.67
Threads fairness:
events (avg/stddev): 347.0000/0.00
execution time (avg/stddev): 9.8907/0.00
Yea still low speed, my nproc=6,mem=2806
My phone has 8 cores and 8G of memory but when running lscpu and free -h in the alpine box it shows 2 cores and 1G of memory can you modify the qemu run command so that it uses all available resources?