-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kvm full system workloads fail to run on x86 #1126
Comments
Have you tried setting "allow_hybrid" to false on the KVM object? |
No, I have not tried that |
Changing gem5/src/cpu/kvm/BaseKvmCPU.py Line 76 in 10b24dc
to false allows me to run simulation again. |
Related PRs:
The data we have right now is that on an i7-13700 (I think; definitely 13th gen with both P and E cores) after #1065 KVM breaks. KVM works on that machine with gem5 stable. Given that #1065 and #1040 were merged around the same time, it's unclear which one (if it's either or both or one of them) that are causing the regression. @Harshil2107, please reply to this with which commits work and don't work on the three machines we talked about (Intel with P/E cores, Intel without P/E cores, and AMD). Please give all the information you have now, and fill in more as you get it. |
I will keep updating this message as I get updates
KVM has failed with 2 different errors so far I will refer to them as error 1 and error 2 Error 1:
Error 2
|
#1138 fixes "Error 2". Not sure about the other error, have not seen it. |
@Harshil2107 can you run with Matt's new PR (#1138) and see what happens? If all cases fail with #1065, I think we should revert it until @nmosier can resubmit. |
The script with KVM runs on AMD when I checkedout #1138 |
Describe the bug
This bug was introduced in #1065
KVM full system workload fail with the following error:
The test config used is the following:
The above is an test script that should get 3 exits.
Affects version
develop for 24.0
To Reproduce
Steps to reproduce the behavior. Please assume starting from a clean repository:
scons build/X86/gem5.opt
Expected behavior
We should get 3 m5 exit events.
The text was updated successfully, but these errors were encountered: