Open LeviMarvin opened 9 months ago
You may need to share your launcher.log for us to confirm (the location of that log is printed above in the output), but I suspect if your virtual device is running for a while and then crashing, is your system perhaps getting tight on memory?
The recommendations for the cpus
flag and the memory_mb
flag are not hard-and-fast rules, they are only some suggested values we've provided as we've found the defaults can be quite slow on the average desktop. You can scale them up and down depending on your host resources.
@prashanthswami I'm sorry for I can't provide the log. I have turned to Windows 11 platform with WSL2. Is there a way to run Cuttlefish on WSL2? And the UI provided by WebRTC has some delay for interaction.
Unfortunately, Cuttlefish is only officially supported on cloud platforms or Linux x86 machines - other variations like WSL2 might work, but I imagine it'd be a brittle experience. Without the launcher logs, there's not much I can suggest, but I'll warn that even if we come up with a fix on this thread, I can't guarantee it won't break elsewhere later. If you're planning to develop, I'd highly suggest using a VM or getting a dedicated Linux machine.
In regards to the UI being delayed, that's unfortunately expected for the moment - QEMU emulation isn't fast and there are long-term projects going on to help make some of the critical operations faster. However, it's hard to beat native emulation and at best, I'd expect that these types of projects will allow us to enable vector instructions without experiencing a huge performance drop on QEMU.
Host details:
Boot command:
But If run the command without
--cpus 6 --memory_mb 8192
, then it can be booted with default cpus and gpus.Boot log:
Boot log without cpus and memory_mb flags: