PixInsight computer specs minimum and middle

Hi,

I want to know what is the minimum requirements for PixInsight as computer to have without much of bottlenecking with processing?

I am not sure i will build a very strong computer for processing with PixInsight [or even for planetary captures stacking], but i don't want to live with so old models and waste more times in processing things, for now i don't heavy data to work with, but i doubt that will last longer sooner or later, and with APS-C file sizes [not sure if a full frame will be added later] so i need to make sure that the machine can handle something, i don't mind some time waiting as i can do a lot of other stuff while waiting, but i don't know if that will take very very long time actually if a lot of data files [or huge planetary capturing files].

All I can say is that I have a core-i7 8700 with 32G RAM and dual NVme drives and it works fine on anything I can throw at it. My PC is not exactly state-of-the-art; I built it myself about 3 years ago.

Maybe you should give us some idea about what you have and what you can afford.

Here are the official PI recommendations:

https://pixinsight.com/sysreq/index.html

The simplest recommendation is to get the most powerful machine you can afford.

Here are the official PI recommendations:

https://pixinsight.com/sysreq/index.html

The simplest recommendation is to get the most powerful machine you can afford.

Hardly a recommendation at all, isn't it?

Hello Tareq,
Just some background for me. I had an i5 6400 with 16gigs of ram, 4gig GPU and I switched the HD to an SSD drive and I was surprised at how well it actually did. The biggest problem if you want to call it a problem was it just took so long to do many of the processes. But it wasn’t until I started doing a lot of “post processing” and specifically in PI the machine would crash after about 2-3hrs using it. During that time I would always make sure I saved my project every 1/2 hour 45 minutes so that if it crashed I’d loose as little as possible. In some cases I’d close out and reboot the PC and start back into PI. I did this for about a year or so and decided I was going to have one built.

In the end I ended up with a 5950x Ryzen 9 CPU with a 12gig 3060 GPU 64gigs or DDR4 viper ram and of course an SSD drive. When I first got this machine after about 4hrs of solid working in PI it would crash, so I added the extra 32gigs for a total of 96gigs. I’m happy to say that since I added the extra ram its been an incredible machine.

I know many use i7 configurations and have good luck with them so you do not need to go crazy with a machine. I think probably an average i7 would do fine. I’m far from a computer techy person as many will attest to here LOL but this has been my experience is all with PI.

Dale

I use an Intel i5 with 4 cores, 16 GB RAM and 1GB SSD drive. It works reasonably well and has never crashed while using PI. The images I process are 16 MP.

If budget is tight, my minimum recommendation would be to get at least 16 GB ram and a 500 GB sata SSD for working with the 16 Megapixel files from your QHY163M.

If you plan to stack lots of subs (100+ per filter), then 32 GB ram would be a good upgrade to improve stacking speed and to prevent crashes from memory overflow. Most of the common post-processing steps in pixinsight do not require much processing power.

None of the native pixinsight processes are affected by the GPU so you can definitely save money here. A dedicated GPU is only useful to accelerate the processing time of 2 plug-ins (StarXterminator and Starnet2). It must be an Nvidia card. Both star removal programs run without GPU by default.

Edit: Investing in a good monitor with IPS panel and accurate color reproduction could also improve your computer experience.

As for the usefulness of a GPU, I would like to say that the plugins StarXtermimator/Startools are very useful. Once I started using them, I would not like to be without them. There is also the NoiseXterminator plugin which also benefits from a GPU, and which is also very useful. And the time gain when using the GPU to boost these plugins is substantial, even with a modest GPU. It does not need to be a super expensive one. The computer I am using costed about 800$ including a GPU and an extra SSD disk. And secondly, I really hope that the number of processes in PI that can use GPU processing will increase rather soon. That the PI developers have not taken the effort to implement this is one of the biggest flaws there is with PI.

I have no doubt about the usefulness of StarXterminator and Starnet for image processing. Running Starnet v2 on a 4656x3520 (16 Megapixel) image takes 1m40sec on my 7-year old i5 4670k processor. I typically only need to run Starnet once or twice during the processing of an entire dataset, so I'm personally not really bothered by that delay from a minimum-spec perspective.

@Tareq Abdulla what type of computer are you currently using for processing astrophotos?

Andy Wray:
All I can say is that I have a core-i7 8700 with 32G RAM and dual NVme drives and it works fine on anything I can throw at it. My PC is not exactly state-of-the-art; I built it myself about 3 years ago.

Maybe you should give us some idea about what you have and what you can afford.

Ok cool, i will

For now my main computers for processing are the following three, all old models:

1. Asus RIVE + i7-3930K + 32GB RAM and GTX Titan

2. Apple Mac Mini late 2011 [or late 2012 i don't remember], it has 16GB RAM and whatever CPU or GPU i don't know

3. My last gaming laptop, HP Pavilion, i7-7700HQ, 32GB RAM, 1050 GPU

All above are installed with either 1TB or 512GB SSD, only laptop has NVMe M.2 512, Mac Mini 512GB 2.5" and Asus RIVE 1TB 2.5".

I am thinking about buying an extra mini PC but with either AMD Ryzen 9 5900HX or 9 5980HX and crank RAM/SSD to highest i can have, that will be like 2 in 1, or even 3 in 1, so it will replace my laptop HP and Asus RIVE.

Here are the official PI recommendations:

https://pixinsight.com/sysreq/index.html

The simplest recommendation is to get the most powerful machine you can afford.

Thank you

If i can't afford that?

andrea tasselli:
Here are the official PI recommendations:

https://pixinsight.com/sysreq/index.html

The simplest recommendation is to get the most powerful machine you can afford.

Hardly a recommendation at all, isn't it?

I agree, kinda making my question here just .... useless.

Dale Penkala:
Hello Tareq,
Just some background for me. I had an i5 6400 with 16gigs of ram, 4gig GPU and I switched the HD to an SSD drive and I was surprised at how well it actually did. The biggest problem if you want to call it a problem was it just took so long to do many of the processes. But it wasn’t until I started doing a lot of “post processing” and specifically in PI the machine would crash after about 2-3hrs using it. During that time I would always make sure I saved my project every 1/2 hour 45 minutes so that if it crashed I’d loose as little as possible. In some cases I’d close out and reboot the PC and start back into PI. I did this for about a year or so and decided I was going to have one built.

In the end I ended up with a 5950x Ryzen 9 CPU with a 12gig 3060 GPU 64gigs or DDR4 viper ram and of course an SSD drive. When I first got this machine after about 4hrs of solid working in PI it would crash, so I added the extra 32gigs for a total of 96gigs. I’m happy to say that since I added the extra ram its been an incredible machine.

I know many use i7 configurations and have good luck with them so you do not need to go crazy with a machine. I think probably an average i7 would do fine. I’m far from a computer techy person as many will attest to here LOL but this has been my experience is all with PI.

Dale

I don't know what were people using before those i9 or Ryzen 9 CPUs came out, and i am sure those i9/R9 will be also outdated next year or in 2 years, and i can't keep thinking about upgrading, i only started astro in 2017, and i started to use PixInsight since 2018 but didn't have any heavy data for that, but i did heavy capturing of planetary in entire 2018 and the files are in GB per file video, but now with my DSO setup getting bigger or more and i will start imaging soon then i will have a lot of DSO data to process then, and by years i might grow with more data to process, so i don't know when or at what point my current or very soon computer can do the job efficiently.

I use an Intel i5 with 4 cores, 16 GB RAM and 1GB SSD drive. It works reasonably well and has never crashed while using PI. The images I process are 16 MP.

And i use my i7-7700HP laptop which is able to process my files just fine, i even fed it with my new cameras IMX571 ~50MB file and it was fine, but because i use this laptop for all my things [surfing, gaming, chatting, watching YT,...etc] it definitely slowing down, and i can't wait down the processing to finish so i continue my general all round things, and with my Asus RIVE i can live it to finish processing fine, but it is connected to my big 4K TV so i don't like to turn on the TV always and keep it on for all my processing to finish, and the desktop is so noisy anyway.

Victor Van Puyenbroeck:
If budget is tight, my minimum recommendation would be to get at least 16 GB ram and a 500 GB sata SSD for working with the 16 Megapixel files from your QHY163M.

If you plan to stack lots of subs (100+ per filter), then 32 GB ram would be a good upgrade to improve stacking speed and to prevent crashes from memory overflow. Most of the common post-processing steps in pixinsight do not require much processing power.

None of the native pixinsight processes are affected by the GPU so you can definitely save money here. A dedicated GPU is only useful to accelerate the processing time of 2 plug-ins (StarXterminator and Starnet2). It must be an Nvidia card. Both star removal programs run without GPU by default.

Edit: Investing in a good monitor with IPS panel and accurate color reproduction could also improve your computer experience.

Let me just say that my gaming laptop has 32GB RAM and NVMe 1TB SSD, so that helped a lot, but this laptop i bought mainly for gaming and my general stuff, suddenly i started to use it for processing because it is my top performance computer i have so far, but with my other things it kinda of slowing the processing and my general stuff too, so i am thinking about a PC cheap for processing mainly that i can connect to my monitor or TV as necessary, which i do for my gaming laptop and new mini PC, and now i saw many mini pcs that can have nice performance and i was thinking about buying one so that can be dedicated to processing even if it is not matching a better desktop, but it will smash outperforms all my current computers anyway, i asked about it once somewhere and many people just tried hardly with me to STAY AWAY from this powerful mini pc so i spend more elsewhere which i can't or not planning to do any soon.

As for the usefulness of a GPU, I would like to say that the plugins StarXtermimator/Startools are very useful. Once I started using them, I would not like to be without them. There is also the NoiseXterminator plugin which also benefits from a GPU, and which is also very useful. And the time gain when using the GPU to boost these plugins is substantial, even with a modest GPU. It does not need to be a super expensive one. The computer I am using costed about 800$ including a GPU and an extra SSD disk. And secondly, I really hope that the number of processes in PI that can use GPU processing will increase rather soon. That the PI developers have not taken the effort to implement this is one of the biggest flaws there is with PI.

And i want to settle with something until that PI can go final stable version so i don't have issues with any latest machine i have or keep thinking about upgrading time to time.

Victor Van Puyenbroeck:
I have no doubt about the usefulness of StarXterminator and Starnet for image processing. Running Starnet v2 on a 4656x3520 (16 Megapixel) image takes 1m40sec on my 7-year old i5 4670k processor. I typically only need to run Starnet once or twice during the processing of an entire dataset, so I'm personally not really bothered by that delay from a minimum-spec perspective.

@Tareq Abdulla what type of computer are you currently using for processing astrophotos?

I answered that to Andy, so you will see what i am using as my main top 3 computers for now.

I was thinking about starting to build a powerful computer as i did in long time past ago with Asus RIVE, but that took me nearly 1 year and things went outdated really fast, the machine is still a workhorse for me for some things, but nowadays with like 12/13 gen of CPUs i feel like i will never keep using that forever, but if i start a new build i will also take so long time when things will get outdated too but i can't keep changing components all the time, so i keep thinking about a powerful performance mini pc i found or just start a build and keep building for long time until done.

I would like to add that if your looking for power regaurdless of what componants your going to use, I’d stay in the PC configuration instead of a laptop. PC’s just run faster and I’d go with a minimum of 32gigs of ram and of course an SSD drive.

Tareq Abdulla:
Andy Wray:
All I can say is that I have a core-i7 8700 with 32G RAM and dual NVme drives and it works fine on anything I can throw at it. My PC is not exactly state-of-the-art; I built it myself about 3 years ago.

Maybe you should give us some idea about what you have and what you can afford.

Ok cool, i will

For now my main computers for processing are the following three, all old models:

1. Asus RIVE + i7-3930K + 32GB RAM and GTX Titan

2. Apple Mac Mini late 2011 [or late 2012 i don't remember], it has 16GB RAM and whatever CPU or GPU i don't know

3. My last gaming laptop, HP Pavilion, i7-7700HQ, 32GB RAM, 1050 GPU

All above are installed with either 1TB or 512GB SSD, only laptop has NVMe M.2 512, Mac Mini 512GB 2.5" and Asus RIVE 1TB 2.5".

I am thinking about buying an extra mini PC but with either AMD Ryzen 9 5900HX or 9 5980HX and crank RAM/SSD to highest i can have, that will be like 2 in 1, or even 3 in 1, so it will replace my laptop HP and Asus RIVE.

To be honest, your i7-3930K is only 17% slower than my 8700, so should be absolutely fine with the 32G RAM you have

Dale Penkala:
I would like to add that if your looking for power regaurdless of what componants your going to use, I’d stay in the PC configuration instead of a laptop. PC’s just run faster and I’d go with a minimum of 32gigs of ram and of course an SSD drive.

I don't think it is about more RAM and SSD drive, because i can do that even with my another old desktops that has i7 or even i5 as all those desktops i built has either 4 or 8 slots for RAM so i can go up to 32GB or 64GB, and all of them are using SSDs, and yet i can't use them to be as super machines for processing comparing to many computers nowadays, unless if many people are just don't care to have super-fast or super powerful machines for processing and they are happy with whatever they have, this i can understand can live with.

Andy Wray:
Tareq Abdulla:
Andy Wray:
All I can say is that I have a core-i7 8700 with 32G RAM and dual NVme drives and it works fine on anything I can throw at it. My PC is not exactly state-of-the-art; I built it myself about 3 years ago.

Maybe you should give us some idea about what you have and what you can afford.

Ok cool, i will

For now my main computers for processing are the following three, all old models:

1. Asus RIVE + i7-3930K + 32GB RAM and GTX Titan

2. Apple Mac Mini late 2011 [or late 2012 i don't remember], it has 16GB RAM and whatever CPU or GPU i don't know

3. My last gaming laptop, HP Pavilion, i7-7700HQ, 32GB RAM, 1050 GPU

All above are installed with either 1TB or 512GB SSD, only laptop has NVMe M.2 512, Mac Mini 512GB 2.5" and Asus RIVE 1TB 2.5".

I am thinking about buying an extra mini PC but with either AMD Ryzen 9 5900HX or 9 5980HX and crank RAM/SSD to highest i can have, that will be like 2 in 1, or even 3 in 1, so it will replace my laptop HP and Asus RIVE.

To be honest, your i7-3930K is only 17% slower than my 8700, so should be absolutely fine with the 32G RAM you have

And it is really doing processing just fine for my DSO and planetary so far until now, not sure later when i go higher in data and captures, so i won't wait until bottlenecking my i7-3930K for that, and hopefully the new mini pc i will buy can do way way much better for me.

Related discussions