Page 1 of 5

Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:27 pm
by kovarex

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:28 pm
by luc
kovarex wrote:soon
Can't wait :)

Edit: you're editing your post? So can I!

Will the replacement between belt and splitter also work with two belts? Since I'd use it most often when having a lot of belts next to each other, when doing some balancing between them.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:32 pm
by kovarex
luc wrote:Will the replacement between belt and splitter also work with two belts?
Yes, it would be sad if it didn't :)

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:35 pm
by d.kr
It turned out that it is most probably caused by slow SSD speed, which could be caused by writing a huge amount of data when working. From what I found on the internet, the typical SSD write capacity is something around 1TB of data, which is not so hard to approach if one recompilation cycle of Factorio generates 5GB of data.
Why are you compiling on a phyiscal disk and not a ram-disk? You might need more ram (but your beefy pc probably can handle that) but its faster and you only a few build artifacts of that data anyway.

Edit: Reading the everything before commenting might have helped:
and I'm planning to get enough memory to compile on ramdisk to prevent this problem in the future.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:36 pm
by Bilka
Nice new features! Some many QoL changes again :D

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:48 pm
by LD100
The white poles at the top front of the pumpjack look wired shouldn't there be a rope and not a pole like in this video: https://youtu.be/GQJWcp8EcrE?t=13s

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:51 pm
by Formal Rice Farmer
Image

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:53 pm
by beiju
The belt fast-replace looks like it'll be great! While you're there, can you also make it so rotating a belt by fast replacement doesn't put its items in your inventory? You can already achieve that by using R with an empty cursor, so it should be possible, and it would eliminate the last common case for getting random extras in your inventory now that splitter and underground fast replacement doesn't do that.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:55 pm
by meelock
The storage tank interior background really needs to be dark, otherwise, the fact that the interior isn't 3d is really glaring.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:55 pm
by peet1993
d.kr wrote:
It turned out that it is most probably caused by slow SSD speed, which could be caused by writing a huge amount of data when working. From what I found on the internet, the typical SSD write capacity is something around 1TB of data, which is not so hard to approach if one recompilation cycle of Factorio generates 5GB of data.
Why are you compiling on a phyiscal disk and not a ram-disk? You might need more ram (but your beefy pc probably can handle that) but its faster and you only a few build artifacts of that data anyway.
He literally says that he will do so in the FFF :D
I want to add that a cheap SSD like the Samsung 750 Evo (using TLC memory) has a lifetime of 35TBWritten. I don't know where you got the 1TB value from, but i guess you got the lifetime of a single cell. This is enhanced by so-called wear leveling. Of course you can reach the TBW cap pretty fast aswell if you recompile often :D

Nice article, and say thanks to your new guy for this awesome QoL change!

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:56 pm
by ohmusama
Formal Rice Farmer wrote:Image
Non joining pipes confirmed

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:59 pm
by kovarex
peet1993 wrote:I don't know where you got the 1TB value from, but i guess you got the lifetime of a single cell. This is enhanced by so-called wear leveling. But you can reach the TBW cap pretty fast aswell if you recompile often :D

Nice article, and say thanks to your new guy for this awesome QoL change!
Sorry, I meant 1000TB (I updated the post).

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 8:59 pm
by kovarex
ohmusama wrote:Non joining pipes confirmed
Things like that happen when the composition is created in the graphics editor, and not in the game.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:02 pm
by peet1993
kovarex wrote:
peet1993 wrote:I don't know where you got the 1TB value from, but i guess you got the lifetime of a single cell. This is enhanced by so-called wear leveling. But you can reach the TBW cap pretty fast aswell if you recompile often :D

Nice article, and say thanks to your new guy for this awesome QoL change!
Sorry, I meant 1000TB (I updated the post).
Well I guess that makes my explanation obsolete :)

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:02 pm
by Engimage
I am glad you found some time to make your own life easier.

I know compilation times just rob you of your time but they also give you some rest. And you just removed these 5 min breaks :) Now to get a break you will need to find another reason (hopefully not smoking).

I am really happy about belt/splitter replacements. Hopefully this will also work with blueprints (maybe in Shift [force] mode).

And yes SSD speed becomes lower when it is getting full. Especially if you have a lot of write operations while reaching max capacity.
So for dynamic data (like compilation) or databases it is best to use HDD (preferably SAS) in hardware RAID arrays (with a Strip in nature, RAID5 [fine] or RAID10 [best]). Such array will serve you for many years with near to no damage like SSD have. However this is far more expensive compared to common SSD drives.

And as always - love HR graphics. And waiting for 0.16 like there is no tomorrow.

I promise to make a set of cool video tutorials when it comes out.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:04 pm
by ledow
Where does the time spent on that compile optimisation appear on the classic XKCD chart?:

https://xkcd.com/1205/

And the more junk like boost that you throw in (and I imagine much of that is spent compiling code that's later discarded because it's not actually used) the worse it will be.

I'm more surprised you don't just have a cached/distributed compile and/or automatic compilation when a SCM update is pushed.

By the time you've pushed new code to a tree, and navigated to a shared drive of some type to run the new binary, it'll probably be ready.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:05 pm
by BHakluyt
The fast belt replacement looks fantastic. Can't wait for 0.16 for all the nice features and HR graphics. I hope with all your optimisations and speeding up processes that you have more time to make the spidertron. Keep up the great work and thank you all for one of the best games ever!

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:08 pm
by ohmusama
kovarex wrote:
ohmusama wrote:Non joining pipes confirmed
Things like that happen when the composition is created in the graphics editor, and not in the game.
Non joining pipes unconfirmed. :(

Although I could use to different kinds of pipes that don't connect (except to structures, like pumps, tanks, refineries, etc). Would be nice for a fluid bus.

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:13 pm
by Durabys
So Jai Programming Language? You mean this?

https://en.wikipedia.org/wiki/Jonathan_ ... I_language

Re: Friday Facts #206 - Workflow optimisation

Posted: Fri Sep 01, 2017 9:14 pm
by bman212121
Of the XKCD posts I would have picked this one first:

https://xkcd.com/303/