Welcome to gem5 Q&A, where you can ask questions and receive answers from other members of the community.

8 or more CPUs in aarch64 configuration: ArchTimer problem

0 votes

Hey,

I'm running the latest gem5 version using a the dedicated gem5 ARM 4.3 Linux kernel in a aarch64 configuration with the default ubuntu-trusty-headless.img. I attached a second SCSI disk image with several microbenchmarks. The configuration is using no caches.

When I boot the configuration with 4 cores the benchmarks execute as expected. However when I switch to 8 (or more cores by enabling gem5_extensions) I occasionally get this error:

gem5.opt: build/ARM/sim/eventq_impl.hh:44: void EventQueue::schedule(Event*, Tick, bool): Assertion `when >= getCurTick()' failed.
Program aborted at tick 18446696982122995000
--- BEGIN LIBC BACKTRACE ---
gem5.opt(_Z15print_backtracev+0x15)[0xe8b3b5]
gem5.opt(_Z12abortHandleri+0x39)[0xea7c19]
/usr/lib64/libpthread.so.0(+0xf370)[0x2ae213d4d370]
/usr/lib64/libc.so.6(gsignal+0x37)[0x2ae2155c61d7]
/usr/lib64/libc.so.6(abort+0x148)[0x2ae2155c78c8]
/usr/lib64/libc.so.6(+0x2e146)[0x2ae2155bf146]
/usr/lib64/libc.so.6(+0x2e1f2)[0x2ae2155bf1f2]
gem5.opt(_ZN9ArchTimer13updateCounterEv+0x207)[0xde2fa7]
gem5.opt(_ZN6ArmISA3ISA10setMiscRegEiRKmP13ThreadContext+0x1475)[0x9b7c15]
gem5.opt(_ZNK10ArmISAInst5Msr647executeEP11ExecContextPN5Trace10InstRecordE+0x375)[0x1a59c25]
gem5.opt(_ZN15AtomicSimpleCPU4tickEv+0x26d)[0xbed29d]
gem5.opt(_ZN10EventQueue10serviceOneEv+0xb1)[0xe98831]
gem5.opt(_Z9doSimLoopP10EventQueue+0x38)[0xec3828]
gem5.opt(_Z8simulatem+0x1fb)[0xec3dbb]
gem5.opt[0xc944dd]
gem5.opt(PyEval_EvalFrameEx+0x534d)[0x1dcf3dd]
gem5.opt(PyEval_EvalCodeEx+0x89c)[0x1dd0f7c]
gem5.opt(PyEval_EvalFrameEx+0x5130)[0x1dcf1c0]
gem5.opt(PyEval_EvalCodeEx+0x89c)[0x1dd0f7c]
gem5.opt(PyEval_EvalFrameEx+0x5130)[0x1dcf1c0]
gem5.opt(PyEval_EvalCodeEx+0x89c)[0x1dd0f7c]
gem5.opt(PyEval_EvalFrameEx+0x54a5)[0x1dcf535]
gem5.opt(PyEval_EvalCodeEx+0x89c)[0x1dd0f7c]
gem5.opt(PyEval_EvalFrameEx+0x5130)[0x1dcf1c0]
gem5.opt(PyEval_EvalCodeEx+0x89c)[0x1dd0f7c]
gem5.opt(PyEval_EvalCode+0x19)[0x1dd1069]
gem5.opt(PyRun_StringFlags+0xf5)[0x1df8ac5]
gem5.opt(_Z6m5MainiPPc+0x7f)[0xea748f]
gem5.opt(main+0x33)[0x8dae93]
/usr/lib64/libc.so.6(__libc_start_main+0xf5)[0x2ae2155b2b35]
gem5.opt[0x93575f]

Has anyone else experienced this recently? Is this a known problem?
I found this thread from several months ago, but it remained unanswered.
https://www.mail-archive.com/gem5-users@gem5.org/msg13982.html

It seems to be a very similar if not exact the same problem.
I would really appreciate if someone could help me out on this for my thesis work!

asked May 28 in Configuration by timaa2k (120 points)

Please log in or register to answer this question.

...