[AMD Official Use Only - General]
I see, thanks again. The verification should have passed it that case...
The docker image is *only* for building applications, so that the user does not
need to install ROCm locally to build applications. You do *not* run gem5
simulations in a docker for full system GPU.
-Matt
From: Rajesh Shashi Kumar <reachrajesh...@gmail.com>
Sent: Wednesday, December 7, 2022 11:32 AM
To: Poremba, Matthew <matthew.pore...@amd.com>
Cc: The gem5 Users mailing list <gem5-users@gem5.org>
Subject: Re: [gem5-users] GPU-FS simulation progress
Caution: This message originated from an External Source. Use proper caution
when opening attachments, clicking links, or responding.
By older ROCm I was referring to this v4.2 which is mentioned in the
gem5-resources documentation. I used the following command to build square. Not
sure if this makes a difference, but do you run the simulation inside the
docker image or on the host machine?
docker run --rm -v ${PWD}:${PWD} -w ${PWD} gcr.io/gem5-test/gpu-fs:latest bash
-c 'make clean; HCC_AMDGPU_TARGET=gfx900 make'
On Wed, Dec 7, 2022 at 11:03 AM Poremba, Matthew
<matthew.pore...@amd.com<mailto:matthew.pore...@amd.com>> wrote:
[AMD Official Use Only - General]
Thanks Rajesh,
That is good to know. I don't think there is an list anywhere of which CPUs
work with KVM.
Which older ROCm do you mean here? Was square compiled with an older version?
Ideally the verification should be passing as well. At least, it does on my
local setup so it would be difficult for me to debug why it does not work for
other folks.
-Matt
From: Rajesh Shashi Kumar
<reachrajesh...@gmail.com<mailto:reachrajesh...@gmail.com>>
Sent: Tuesday, December 6, 2022 5:49 PM
To: Poremba, Matthew <matthew.pore...@amd.com<mailto:matthew.pore...@amd.com>>
Cc: The gem5 Users mailing list
<gem5-users@gem5.org<mailto:gem5-users@gem5.org>>
Subject: Re: [gem5-users] GPU-FS simulation progress
Caution: This message originated from an External Source. Use proper caution
when opening attachments, clicking links, or responding.
Okay turns out the issue was indeed using a slow local machine somehow (AMD
Ryzen 7 5800H)
I ran the same thing on a "AMD EPYC 7451 24-Core Processor x2" and I am able to
run square now within 10minutes or so. I guess the last two lines could be
because of using an older rocm version.
Running ../../gpu/square/bin/square
info: running on device Vega 10 XTX [Radeon Vega Frontier Edition]
info: architecture on AMD GPU device is: 900
info: allocate host and device mem ( 7.63 MB)
info: launch 'vector_square' kernel
info: check result
error: 'hipErrorUnknown'(999) at square.cpp:82
./script.sh: line 13: 548 Segmentation fault ./myapp
Thank you again for your time on this!
--
Rajesh Shashi Kumar
On Tue, Dec 6, 2022 at 7:39 PM Rajesh Shashi Kumar
<reachrajesh...@gmail.com<mailto:reachrajesh...@gmail.com>> wrote:
Thank you for your time. I tried using the provided example for booting Ubuntu
from a disk-image.
./build/X86/gem5.opt configs/example/gem5_library/x86-ubuntu-run.py
With this, I see that the boot did complete with this example and kvm-ok
returns as expected on my machine.
Also, I should mention that I'm using the pre-compiled image/kernel for GPU-FS
to rule out any uncertainty there.
Term output:
Welcome to Ubuntu 18.04.2 LTS!
systemd[1]: Set hostname to <gem5-host>.
systemd[1]: File /lib/systemd/system/systemd-journald.service:36 configures an
IP firewall (IPAddressDeny=any), but the local system does not support
BPF/cgroup based firewalling.
systemd[1]: Proceeding WITHOUT firewalling in effect! (This warning is only
shown for the first loaded unit using IP firewalling.)
random: systemd: uninitialized urandom read (16 bytes read)
systemd[1]: Reached target Remote File Systems.
[ OK ] Reached target Remote File Systems.
random: systemd: uninitialized urandom read (16 bytes read)
systemd[1]: Created slice System Slice.
[ OK ] Created slice System Slice.
...
On Tue, Dec 6, 2022 at 6:30 PM Poremba, Matthew
<matthew.pore...@amd.com<mailto:matthew.pore...@amd.com>> wrote:
[AMD Official Use Only - General]
At this point I would check if the other KVM scripts are working for you (there
are some simple tests somewhere like boot Ubuntu and exit). KVM works on some
CPUs better than others, I believe, or at least this was true in the past. I
have a few other ideas to try, but I would like to see if any other scripts are
working first and understand your setup to see if other folks might run into
the same issue in the future.
-Matt
From: Rajesh Shashi Kumar
<reachrajesh...@gmail.com<mailto:reachrajesh...@gmail.com>>
Sent: Tuesday, December 6, 2022 4:09 PM
To: Poremba, Matthew <matthew.pore...@amd.com<mailto:matthew.pore...@amd.com>>
Cc: The gem5 Users mailing list
<gem5-users@gem5.org<mailto:gem5-users@gem5.org>>
Subject: Re: [gem5-users] GPU-FS simulation progress
Caution: This message originated from an External Source. Use proper caution
when opening attachments, clicking links, or responding.
Thank you for your response.
I double checked my image and kernel, I don't think KVM is hanging but the
progress seems to be a character printed on the term every once in a while. I
assume this is even before it could finish booting. Not sure if fastfoward
could help here
My term output:
==== m5 terminal: Terminal 0 ====
[ 0.000000] Linux version
Thanks,
Rajesh
On Tue, Dec 6, 2022 at 4:20 PM Poremba, Matthew
<matthew.pore...@amd.com<mailto:matthew.pore...@amd.com>> wrote:
[AMD Official Use Only - General]
Hi Rajesh,
I looks like no progress has been made since a very early tick number (the
timestamp print by Linux is equal to the current simulation tick / 1 trillion).
For reference it should take no more than 1-3 wall clock minutes to full boot
Linux and begin running the application with the KVM CPU. I have seen fairly
rarely where the KVM simply hangs and makes no progress but simply running
again fixed this. Your command looks correct though.
Maybe someone who knows more about debugging KVM can comment how to see what
the KVM CPU is doing.
-Matt
From: Rajesh Shashi Kumar via gem5-users
<gem5-users@gem5.org<mailto:gem5-users@gem5.org>>
Sent: Tuesday, December 6, 2022 2:06 PM
To: gem5 users mailing list <gem5-users@gem5.org<mailto:gem5-users@gem5.org>>
Cc: Rajesh Shashi Kumar
<reachrajesh...@gmail.com<mailto:reachrajesh...@gmail.com>>
Subject: [gem5-users] GPU-FS simulation progress
Caution: This message originated from an External Source. Use proper caution
when opening attachments, clicking links, or responding.
Hi,
I followed the instructions on running gpu-fs square using the gem5-resources
repository. My simulation has been stuck here for a while
...
build/VEGA_X86/arch/x86/kvm/x86_cpu.cc:1561: warn: kvm-x86: MSR (0xc0010015)
unsupported by gem5. Skipping.
build/VEGA_X86/arch/x86/kvm/x86_cpu.cc:1561: warn: kvm-x86: MSR (0x4b564d05)
unsupported by gem5. Skipping.
build/VEGA_X86/dev/x86/pc.cc:117: warn: Don't know what interrupt to clear for
console.
16964000000000: system.pc.com_1.device: attach terminal 0
I tried attaching a terminal on a different tab using the following but I'm not
sure if my image has booted or if the application is running:
$ util/term/m5term localhost 3456
==== m5 terminal: Terminal 0 ====
[ 0.00000
Any advice is appreciated!
My run command:
build/VEGA_X86/gem5.opt configs/example/gpufs/vega10_kvm.py --disk-image
../disk-image/rocm42/rocm42-image/rocm42 --kernel ../vmlinux-5.4.0-105-generic
--gpu-mmio-trace ../vega_mmio.log --app ../../gpu/square/bin/square
Thanks,
Rajesh
_______________________________________________
gem5-users mailing list -- gem5-users@gem5.org
To unsubscribe send an email to gem5-users-le...@gem5.org
