You have it partially correct, but for the EDR3 data you need to add gedr3-1.0.0-01.xpsd as well. The databases go progressively in order of magnitude so you need everything from "-01" up to whatever number you've settled on as the appropriate magnitude limit for your imaging setup.
PixInsight 1.8.8-7 Released
- Thread starter Juan Conejero
- Start date
Also, in order to use those files for plate solving, make sure he Image Solver script is configured to use them.
The default Automatic catalog option should be fine (but it's possible you had changed it in the past). Alternatively, set the explicit Local XPSD server to Gaia EDR3 (XPSD) but normally the automatic option is better.
The default Automatic catalog option should be fine (but it's possible you had changed it in the past). Alternatively, set the explicit Local XPSD server to Gaia EDR3 (XPSD) but normally the automatic option is better.
Attachments
The installer should have put them there, my fresh install has them (see below) so in case you run any antivirus software, you may want to disable it temporarily and reinstall PI and check if they are the files are there before re-enabling the antivirus.
Directory of C:\Program Files\PixInsight\library
2020-12-17 06:28 PM 221,418 ca-bundle.crt
2020-12-17 06:28 PM 8,873 default.filters
2020-12-17 06:28 PM 2,337 default.ssm
2020-12-17 06:28 PM 9,085 default.xwrf
2020-12-19 12:34 PM 217,718,902 mono_starnet_weights.pb
2020-12-19 12:34 PM 217,743,491 rgb_starnet_weights.pb
Directory of C:\Program Files\PixInsight\library
2020-12-17 06:28 PM 221,418 ca-bundle.crt
2020-12-17 06:28 PM 8,873 default.filters
2020-12-17 06:28 PM 2,337 default.ssm
2020-12-17 06:28 PM 9,085 default.xwrf
2020-12-19 12:34 PM 217,718,902 mono_starnet_weights.pb
2020-12-19 12:34 PM 217,743,491 rgb_starnet_weights.pb
Ya, strange... they're not there for me. I guess I'll try reinstalling and see what happens.
Reinstalled, checked the directory still weren't there... however when I started PI there were 4 updates and low and behold the pb files are now there
AwesomeI ran an update to 1.8.8-7 but I can not find the StarNet process. Are there any pre-requisites to get this process displayed in the list of available processes( min. processor type, min memory...)?
When you install the 4 updates you get starnet. You can then look for databases it needs. They are also downloaded, but (for me) need to be installed manually.
min processor type is an x86 CPU with AVX2 and FMA instructions, though usually if you are in this situation PI crashes while it's trying to load the module (rather than silently ignoring the module which does happen in other circumstances.)
rob
Hi Jürgen,
I have always defended the necessity to support FreeBSD as the only really open-source operating system IMO, as well as the only pure UNIX platform. I love FreeBSD and will try to support it again for current PixInsight versions, but the amount of work that I have to deal with may be excessive to achieve this goal. Nothing has been decided yet. I'll keep you informed.
I really hate to say this, but we are considering the possibility to drop FreeBSD as a supported platform for PixInsight, at least for some time. At present we have to deal with an overwhelming workload. This includes new important tools requiring a lot of design, implementation and testing work, the GPU support project (CUDA support), and the new Apple Silicon processors, which we must support as soon as possible as a completely new platform. Support for FreeBSD has necessarily to be a non-priority task, especially considering the extremely small number of users working on this operating system.
I have always defended the necessity to support FreeBSD as the only really open-source operating system IMO, as well as the only pure UNIX platform. I love FreeBSD and will try to support it again for current PixInsight versions, but the amount of work that I have to deal with may be excessive to achieve this goal. Nothing has been decided yet. I'll keep you informed.
PixInsight 1.8.8-7 works out-of-the-box without any problems on Kubuntu 20.04 and Ubuntu 20.04 in our tests. If you can provide some information to describe the errors you're getting, such as some terminal output, that would help us diagnose the problem. Otherwise we are unable to know what happens.
Hi Juan,
thank you very much for your response. The background of my question was to identify in which direction i need to go with my setup. For the time being i use a FreeBSD server with quite good hardware. 12 cores, 96GB memory, 20TB storage and so on. FreeBSD is my first choice for a server like that using jails and bhyve VMs all on ZFS as this is manageable quite easily. And stable! I will not change the setup there.
On the other hand i have a 12 year old iMac (office computer also my wife is using). I want to replace that. Knowing that you are planning to support Apple Silicon processors i may get a new MacMini M1 then.
Don't get me wrong, i love Linux as well, i use it as VMs and on smaller projects on raspberries. But for PI you need powerful hardware. Buying separate hardware just for PI is something i want to avoid. I may try it on a Linux VM on the big server. Let's see.
So please keep up the good work! I really appreciate it. I would fully understand that it may take some time to get PI on FreeBSD in the future or even that this platform will not be supported anymore. It's a small user group.... Although i would love to see it there.
Jürgen
thank you very much for your response. The background of my question was to identify in which direction i need to go with my setup. For the time being i use a FreeBSD server with quite good hardware. 12 cores, 96GB memory, 20TB storage and so on. FreeBSD is my first choice for a server like that using jails and bhyve VMs all on ZFS as this is manageable quite easily. And stable! I will not change the setup there.
On the other hand i have a 12 year old iMac (office computer also my wife is using). I want to replace that. Knowing that you are planning to support Apple Silicon processors i may get a new MacMini M1 then.
Don't get me wrong, i love Linux as well, i use it as VMs and on smaller projects on raspberries. But for PI you need powerful hardware. Buying separate hardware just for PI is something i want to avoid. I may try it on a Linux VM on the big server. Let's see.
So please keep up the good work! I really appreciate it. I would fully understand that it may take some time to get PI on FreeBSD in the future or even that this platform will not be supported anymore. It's a small user group.... Although i would love to see it there.
Jürgen
Hi all, It's unclear to me whether enabling the CUDA acceleration fields on the Parallel Processing and Threads tab of Global Preferences, will expedite StarNet in the same profound way (10x!), as following Darkarchon's instructions from 2020- https://darkskies.space/pixinsight-starnet-cuda/. As I'd already replaced the tensorflow dll and updated the NVIDIA card per Stan's instructions, I can't test this. I was under the impression that Stanley's changes were still valid, but perhaps I'm wrong? I note that the tool tip in preferences says that the feature is not yet available, but a colleague believes that it's all that's needed for SN acceleration. Thanks for clarifying!