Lonnie Cumberland <lonnie@...>
toggle quoted messageShow quoted text
for the NUC that I think that I am getting:
Intel NUC NUC7i7DNHE Mini PC/HTPC, Intel Quad-Core i7-8650U Upto 4.2GHz, 32GB DDR4, 512GB SSD, WiFi, Bluetooth, 4k Support, Dual Monitor Capable, Windows 10 Pro (32GB Ram + 512GB SSD)
- 8th Generation Intel Quad-Core i7-8650U 1.9GHz With Turbo Boost Upto 4.2GHz, 8MB Cache
- 32GB DDR4 2400MHz, 512GB Solid State Drive SATA III (You Could Add A m.2 2nd Drive)
- Intel UHD Graphics 620, Wifi, Bluetooth 4.2, Gigabit Ethernet, 4k Support, Dual Monitor Capable
- Windows 10 Professional 64Bit, 4x USB 3.0, 2x HDMI, 1x Headphone/Microphone Combo Jack
It has the "Intel UHD Graphcis 620" which I am hoping to be able to share, at least for simple testing as vGPU's for some host OS's like Windows and Linux for development and testing.
Quick question on the GPU side: are you planning to share the GPU between different User VMs? Or will it be dedicated to one?
Well, I may have narrowed down the spec for the test rig:
1. Intel NUC NUC7i7DNKE Mini PC/HTPC, Intel Quad-Core i7-8650U Upto 4.2GHz, 32GB DDR4, 512GB m.2 SSD, WiFi, Bluetooth, 4k Support, Dual Monitor Capable, Windows 10 Pro (32GB Ram + 512GB SSD) -- $754.99
2. Intel NUC NUC7i3DNHE Mini PC/HTPC, Intel Dual-Core i3-7100U 2.4GHz, 16GB DDR4, 512GB SSD, WiFi, Bluetooth, 4k Support, Dual Monitor Capable, Windows 10 Professional 64Bit (16GB Ram + 512GB SSD) -- $439.99
This one is a little cheaper as well a sa bit less powerful and may not support all of the ACRN scenarios but still might be a good deal
My criteria was basically:
B.) Reasonable RAM, and HDD (or SSD)
C.) Good Graphics Card for some GPU work and Dual HDMI
Any Other suggestions for a possible selection would be greatly appreciated?
Do you have any recommendations for a small inexpensive test rig (possibly NUC) with reasonable graphics that can run ACRN? I think that instead of jumping to trying to get AMD working initially, that I should get a small Intel NUC test
rig setup and start working with ACRN itself after which I can perhaps move to work towards making it work on AMD systems. Baby steps first are better.
I am trying to decipher the acceptable hardware for a reasonable test rig with a reasonable GPU and a minimum of 8 - 16GB RAM, and perhaps a 500GB Drive or better but not sure of the Intel processor that is good. I see a lot of NUC systems
with Core-i3, Core-i5, and Core-i7. Seems like so many choices.
That reminds me that perhaps we should do a refresh of that video :-/
We do not use ClearLinux anymore inside the Service VM. We have switched to Ubuntu (for reasons linked to the way we boot, see below for a bit more details ). Now, that’s not
to say you cannot run a different OS in the Service VM, the key will be to see if you can install all the right dependencies in it in order to run the Device Model (acrn-dm) which is the component providing the drivers (shared devices) for the User VMs. You
can have a rough idea about those dependencies by looking at this section:
https://projectacrn.github.io/latest/getting-started/rt_industry_ubuntu.html#build-the-acrn-hypervisor-on-ubuntu. These are the tools and development libraries needed to *build* ACRN. The runtime dependencies list is a little smaller, i.e. I don’t
believe you need things like bison, flex, pkg-config, python3, git, make (and possibly others). I do not know TinyCore myself but if those packages are readily available for it, it may be a relatively straightforward exercise.
 Staged removal of deprivileged boot mode support. ACRN has supported deprivileged boot mode to ease the integration of Linux distributions such as Clear Linux. Unfortunately,
deprivileged boot mode limits ACRN’s scalability and is unsuitable for ACRN’s hybrid hypervisor mode. In ACRN v2.2, deprivileged boot mode is no longer the default and will be completely removed in ACRN v2.3. We’re focusing instead on using multiboot2 boot
I actually just came across a great video that cleared up the way that ACRN runs:
It helps a lot, but I see that Clear Linux is used a lot and for my project goals, it is still much too large. I need to be able to run something like TinyCore Linux or something
with an Xserver that is around 20 - 50 MB max.
It’s great to see you interested to learn and potential contribute to ACRN!! :-)
The most flexible pre-defined scenario that we have is the “industry” scenario. It runs up to 8 VMs (inc. the Service VM), so that’s 7 User VMs – see also this table
What do you mean by “bare-metal GUI”? Are you looking for a GUI that allows you to control ACRN (something akin to virt-manager for example)?
Have a nice week-end too!
I am beginning my exploration for ACRN as I think that it could offer what I need for my project which is based upon a nUltra-Litghtweight Hypervisor & Virtualizer like ACRN that
will run a number of pre-loaded and dynamically loaded VM's which are composed of streamlined Unikernel applications all of which is run in RAM.
After searching for a VERY long time for possible candidates to use as a starting point with the strong criteria of:
2. Able to run commodity guest OS's
3. Reliable and stable to semi-stable performance
4. Monolithic in design (as much as possible)
5. Would prefer it to run on Intel & AMD X86_64 based systems (others to follow)
I was able to narrow things down to:
A.) NOVA Hyervisor (Still Experimental and not mucd development, used in the Genode project)
Both ACRN and Xvisor seem to be progressing and can offer many features but need development in some areas as well.
I like the very small footprint of ACRN and also that it is industrial grade for security which makes me want to use it as a starting point.
My goal is to do what I can to see if I can get it to run on AMD systems as well as the currently supported Intel systems and then give the patch back to the ACRN project.
I would also be interested to know if there is a scenario in which ACRN can dynamically load and start VM's as I did not read that in the scenario list.
Also, maybe finding a bare-metal GUI and adding it at a later time or perhaps in a VM as well.
Anyway, I look forward to seeing what I can learn and contribute to the ACRN project.
Cheers and have a great weekend,