'Pause' Integration...
 
Share:

APP 1.075 has been released !

Forum support from Vincent will be limited between now and  Oktober the 30th, 2019. Mabula will be available for support on the forum though.

2019 September: Astro Pixel Processor and iTelescope.net celebrate a new Partnership!

'Pause' Integration button  

  RSS

(@cheetah)
Main Sequence Star Customer
Joined: 2 years ago
Posts: 32
August 9, 2018 20:35  

I'd love to be able to Pause the processing.  I'm 11 hours into an integration and I need my laptop to run my scope this evening.  It may work better by reducing the number of threads APP is using, but it's less than ideal.  Cancelling is an option, but I'd prefer to 'PAUSE'.

 

 

 


ReplyQuote
(@mabula-admin)
Quasar Admin
Joined: 2 years ago
Posts: 2148
August 19, 2018 22:57  

Hi @cheetah,

I will put this on my list.

May I ask what you are doing that takes so long? are you integrating several 100s of frames ? Or a mosaic perhaps?

Mabula

 

Main developer of Astro Pixel Processor and owner of Aries Productions


ReplyQuote
(@cheetah)
Main Sequence Star Customer
Joined: 2 years ago
Posts: 32
August 20, 2018 06:29  

Thanks!

The biggest limiting factor is probably that I don't have hard drive space on my laptop.  The working directory and storage for the images is on a remote drive connected via wifi.  The 

The last job I did was 60 or 70 images @ 16mp each.  With 1st level LNC and 5% MBB, it took over 12 hours.  The (LNC in particular seems to take forever...)

My machine is a 2015 Macbook Pro, with a 2.7Ghz Intel i5 & 8gb ram.


ReplyQuote
(@adamj)
Molecular Cloud Customer
Joined: 2 years ago
Posts: 4
August 22, 2018 18:33  
Posted by: Mabula Haverkamp - Admin

Hi @cheetah,

I will put this on my list.

May I ask what you are doing that takes so long? are you integrating several 100s of frames ? Or a mosaic perhaps?

Mabula

 

I have done a 700 image run with 30 second subs on my ASI1600mm pro and it took maybe 8 hours. So I would be also be very happy with pause button. 

Incidentally how much would going from 8gb ram to 16gb ram help me in this sort of situation?


ReplyQuote
(@mabula-admin)
Quasar Admin
Joined: 2 years ago
Posts: 2148
September 25, 2018 20:35  

@cheetah and @adamj, Hard drive IO speed or the IO speed to the hardrive with the work directory, is the main limiting factor here.

Cheetah, you are severly limited by the WIFI connection if you use a work directory on the external drive through WIFI.

LNC is a technology that will alter the layers (images) in your stack, before integration. So if the integration consists of 20GB of data, then APP will need to alter 20GB per LNC iteration of data. If you perfrom that over a Wifi connection, the speed of the Wifi connection will most likely be the main bottle neck and thus determine how long it will take.

When you start to use a SSD drive for your work directory, then you will see very fast integrations and very fast LNC iterations, in that situation, more processors and more memory will start to prove it's worth as well. I know on my machine, with a M2 SSD, integrating with 16Gbs of Ram is about 15-20% faster than with only 4GBs of ram.

And processor threads scale nicely. 2x more processor threads make a lot of calculations in APP twice as fast.

Mabula

Main developer of Astro Pixel Processor and owner of Aries Productions


ReplyQuote
(@cheetah)
Main Sequence Star Customer
Joined: 2 years ago
Posts: 32
August 16, 2019 18:36  

It's time I brought this up again.

I've upgraded since the previous posts.  I'm now on a system that has a huge & fast SSD, but my processor is still slow and I don't have much RAM (8GB).  These are not upgradable, so I'm stuck with it.

Doing a 'normal' integration can take an hour or so.  I can live with this.  But...  I have recently started a process with 450 images.  Today is day 2...still processing.  My system is not usable while this is going on.  I would LOVE to be able to pause for a few minutes while I do something else, and then resume.

*Fingers crossed* this makes it into the next update.


ReplyQuote
(@vincent-mod)
Galaxy Moderator
Joined: 2 years ago
Posts: 845
August 16, 2019 19:38  

I don't think this is high up the list, but should be possible. Usually, when working with these huge datasets, I split them up in batches of say 120 and combine those later. This will make APP use less time in between, but ofcourse is still not as handy as a pause at will.

This post was modified 2 months ago by Vincent Groenewold - Moderator

ReplyQuote
(@cheetah)
Main Sequence Star Customer
Joined: 2 years ago
Posts: 32
August 16, 2019 19:48  

How exactly do you do that?  Does it affect the result?


ReplyQuote
(@vincent-mod)
Galaxy Moderator
Joined: 2 years ago
Posts: 845
August 16, 2019 19:55  

Just split them up in groups, process each group completely, make them look nice and then load those integrated group-stacks and register and process them as well together. I don't think it makes it look worse, you can always compare once, see it for yourself and then decide. 🙂


ReplyQuote
Share: