i need to open the localhost page each time to turn it on or i get that red icon and no connection
Hi @laurentlavigne , I’m sorry to hear that you are still having problem with running Accelerator. Could you provide us more information about what’s going on? For example, screenshots of the red icon and how you know that there is no connection, what you turn on when you go to localhost page, and the endpoint of the localhost page that you are referring to, …etc
it’s our lucky day: vfx graph crashed my machine again so i had to reboot
here is the sequence, i didn’t press any button, just opened localhost in chrome
in chrome
back in unity

PM-ed i the log
After taking a look at the log, we did not find any obvious issue. A couple of things that could affect your setup includes Firewall setup and network connection within the machine.
Also, there’s little benefit to running the Accelerator locally when using the new V2 import pipeline since it already keeps a database in the library folder to make switching more efficient. The Accelerator is better to be used for a team sharing the same local network and it should be set up on a separate machine. For a single user, the Asset Database local caching (which is part of the Editor) should already be enough in most cases.
I did not know that!
Can you expand on “little” and “most cases”?
Yep. I will first clarify the case for a single user. If you mainly develop on one machine, the Asset Database local caching is helpful for the asset import process. If you use the Accelerator in this case, the only benefit is when your local cache does not have certain imported assets that the Accelerator has. If the Accelerator is set up in a different machine, you would suffer from network latency and connection. And if you have Accelerator set up on the same machine, this could affect your machine’s performance as the Accelerator will try to fight for resources with the Editor, which can affect your workflow.
If multiple machines are involved, the Accelerator would be good for those other machines to get imported assets rather than importing from raw assets. This also applies for when multiple users are involved.
(it works fine now)
what does this mean?
v1 or v2? v2 has many performance regressions so i’d rather not use it.
V2 stores it in the project library folder, does this mean the another project on the same disk will not be sharing the same cache and just make a new cache in that projects library folder?
This does not sound better than what v1 did, but please let me know if I am mistaken to assume this?
yeah that’s a big bonus of cache server, it was hashing assets across projects (i think, i hope)
@laurentlavigne , when an Accelerator is running, if the disk space is low, the Accelerator will try to evict items from cache and this means a lot of Read/Write operations being performed to the Disk. If you run this on the same machine with the Unity Editor or other software, the Accelerator will try to gain system resources, which can increase the CPU usage/Memory usage, and make the Editor performing slowly. The Asset Database local caching I mentioned was for Asset Database pipeline V2. We are aware that there are some performance regression issues and we are currently working on them. If you want to use Asset Database pipeline v1, please install Unity Cache Server instead of the Accelerator.
@Lars-Steenhoff , yes, V2 has local database per project and it is not shared across the different project. The difference is that V1 did not have this and in V1 you would have to reimport everything when you switch build targets. If you want different projects to share the cached data, the Accelerator would be a good choice here. It can be set up on the same machine as the Editor, but as I explained above, it’s better to set it up on a separate machine.
My network is not fast enough to install accelarator on a seperate machine, ( and I dont have a seperate machine )
So what I did is get an external 1tb nvme ssd and set that as the cache location for the accelerator.
Would that eliminate the fighting for resources?
Or should I stick to v1 with local cache server if I want shared cache between projects on the same machine?
Seems v1 is deprecated in 2020.2 that I’m using.
So I’m using the accelerator now locally.
In the old v1 I did enable local cache and that made platform switching so fast that I did not mind anymore.
It feels like v2 will make every new project cache everything again, and a therefore not as efficient as v1.
The worst thing for me is the texture compression importing part, I hate that with a passion. ( yes I did disable compress on import )
When making a build it compressed the whole project folder ( don’t need all textures in build )
I cannot wait for the “only compress textures that are actually used needed in the build” option.
If you only have one project, V2 alone would be enough.
If you have multiple projects, Accelerator should be used as V2 does not have a shared cache database in the Editor.
Installing Accelerator with cache location on an external drive would help. However, we have seen some problems with using the external drive in the past so please let us know if you have any problem installing with an external drive.
I did not have any problems yet, I did not set the location in the editor but made a symbolic link to the folder on the external drive. so the editor thinks everything its at its normal place