>
> I want to know the reason
>
"panic: Possible Deadlock detected. Aborting!"

First I would build gem5.debug to get a more useful bt. Put a breakpoint on
panic() and then inspect m_last_progress_vector to see what is happening.

Also, if you're not seeing the problem with synth traffic, try upping the
injection rate. Probably there is a problem with your routing algorithm.

Ryan Gambord
<[email protected]>



On Wed, Jun 26, 2019 at 6:55 PM 蔡源 <[email protected]> wrote:

> Hi all,
>   I modified the routing algorithm in the garnet2.0. I used this command
> "./build/NULL/gem5.debug configs/example/garnet_synth_traffic.py
> --num-cpus=64 --num-dirs=64 --network=garnet2.0 --topology=Mesh_XY
> --mesh-rows=4 --mesh-cols=4 --sim-cycles=10000 --synthetic=uniform_random
> --routing-algorithm=2 --vcs-per-vnet=2 --inj-vnet=2
> --injectionrate=0.001", the gem5 ran successfully. However, I used this
> command "./build/X86/gem5.opt configs/example/fs.py --cpu-type=timing -n 64
> --num-dirs=64 --ruby --network=garnet2.0 --topology=Mesh_XY --mesh-rows=4
> --mesh-cols=4 --routing-algorithm=2 --vcs-per-vnet=2", I got some errors:
> panic: Possible Deadlock detected. Aborting!
> version: 0 request.paddr: 0x7c5fc8 m_readRequestTable: 1 current time:
> 530653514000 issue_time: 530403514000 difference: 250000000
>  @ tick 530653514000
> [wakeup:build/X86/mem/ruby/system/Sequencer.cc, line 104]
> Memory Usage: 1190188 KBytes
> Program aborted at tick 530653514000
> --- BEGIN LIBC BACKTRACE ---
> ./build/X86/gem5.opt(_Z15print_backtracev+0x15)[0x772a25]
> ./build/X86/gem5.opt(_Z12abortHandleri+0x36)[0x785786]
> /lib/x86_64-linux-gnu/libpthread.so.0(+0x10330)[0x7f49fba03330]
> /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x37)[0x7f49fa4bac37]
> /lib/x86_64-linux-gnu/libc.so.6(abort+0x148)[0x7f49fa4be028]
> ./build/X86/gem5.opt(_Z15__exit_epilogueiPKcS0_iS0_+0x1ec)[0x98580c]
>
> ./build/X86/gem5.opt(_Z14__exit_messageIIjmmmmmEEvPKciS1_S1_iS1_DpRKT_+0xc7)[0x9253f7]
> ./build/X86/gem5.opt(_ZN9Sequencer6wakeupEv+0x269)[0x922cd9]
> ./build/X86/gem5.opt(_ZN10EventQueue10serviceOneEv+0xb1)[0x77d751]
> ./build/X86/gem5.opt(_Z9doSimLoopP10EventQueue+0x38)[0x7a2fa8]
> ./build/X86/gem5.opt(_Z8simulatem+0x1fb)[0x7a352b]
> ./build/X86/gem5.opt[0x7fb1bd]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x45f7)[0x7f49fbd50607]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalCodeEx+0x80d)[0x7f49fbd52c8d]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x48d8)[0x7f49fbd508e8]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x4b59)[0x7f49fbd50b69]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x4b59)[0x7f49fbd50b69]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalCodeEx+0x80d)[0x7f49fbd52c8d]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalCode+0x32)[0x7f49fbd52f72]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x563e)[0x7f49fbd5164e]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalCodeEx+0x80d)[0x7f49fbd52c8d]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x48d8)[0x7f49fbd508e8]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalCodeEx+0x80d)[0x7f49fbd52c8d]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyEval_EvalCode+0x32)[0x7f49fbd52f72]
>
> /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0(PyRun_StringFlags+0x79)[0x7f49fbc65579]
> ./build/X86/gem5.opt(_Z6m5MainiPPc+0x7f)[0x784fff]
> ./build/X86/gem5.opt(main+0x33)[0x728423]
> /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7f49fa4a5f45]
> ./build/X86/gem5.opt[0x754763]
> --- END LIBC BACKTRACE ---
> Aborted (core dumped)
> I want to know the reason and how to solve this problem.
>
> Thanks in advance for any help.
>
> Yuan Cai
>
> _______________________________________________
> gem5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to