Friday Facts #176 - Belts optimization for 0.15

[Klonan]

another dose of FFF goodness: https://www.factorio.com/blog/post/fff-176

[Rhamphoryncus]

It's impressive how much engineering has to go into something as conceptually simple as inserters just so you never have to think about them!

[dauphin]

I assume one of the big caveats for the performance increase on the belts is that, in order to render, you still have to calculate the absolute position of each item on the belt?

This has been one of the best recent FFFs. Good show!

[Cyberboss_JHCB]

Will underground belts still be any faster than regular belts, even if it's just miniscule?

[keyboardhack]

This is some impressive optimization work. At this point is it still true that bots are better for performance than belts in 0.15?

[Optera]

I can hardly wait to have my base currently crawling at ~20ups back up to playable 30-40ups again.
Will the changes in belt mechanic break the unintended splitter feature know as magic splitters?

[kinnom]

This is some impressive optimization work. At this point is it still true that bots are better for performance than belts in 0.15?

since when are bots better than belts?

[Tankh]

whenever a belt compresses - it will stay that way forever. It means that once two items are stuck close together - away from inserters they will stay stuck close forever. This property allows us to cache the index of the last positive gap location, and update it on the fly because that index can never increase, only decrease.

I'm guessing you also tested this for items going from a slow belt to fast belt. It's probably treated like a separate belt section.

[posila]

whenever a belt compresses - it will stay that way forever. It means that once two items are stuck close together - away from inserters they will stay stuck close forever. This property allows us to cache the index of the last positive gap location, and update it on the fly because that index can never increase, only decrease.

I'm guessing you also tested this for items going from a slow belt to fast belt. It's probably treated like a separate belt section.

That is correct, only belts of the same type are merged to the same section.

[SaintFlow]

Nice ideas for belt optimization!

I am very sad that its Feb now, first Feb FFF is out and you still didn't say a word about 0.15 planning. You guys know that people are looking forward to it quite a lot and I guess many of those looking forward expected at least some information in this fff today. I understand that stating something and then delaying might piss some people off, but not communicating at all and just plain ignoring such questions doesnt make anybody happy either. Smart Belts --> Coolbeans! No info about 0.15 --> sad.

[sparr]

I assume one of the big caveats for the performance increase on the belts is that, in order to render, you still have to calculate the absolute position of each item on the belt?

you don't have to separately calculate each one. you calculate the first one, then just increment it in the appropriate direction for the distance to the next item, with some logic to handle corners.

[Zeblote]

There's a bug in your first gif

[Xeanoa]

The first release in months, and it fixes what, 3 of the dozens of reported bugs?

[Loewchen]

whenever a belt compresses - it will stay that way forever. It means that once two items are stuck close together - away from inserters they will stay stuck close forever. This property allows us to cache the index of the last positive gap location, and update it on the fly because that index can never increase, only decrease.

I'm guessing you also tested this for items going from a slow belt to fast belt. It's probably treated like a separate belt section.

This got me alarmed as well...

[Sigma1]

There's a bug in your first gif

I noticed it too

[Neemys]

The first release in months, and it fixes what, 3 of the dozens of reported bugs?

Lots of bug were fixed but only in 0.15 (no ETA yet) (see here for list of resolved bug), this release exist only to address critical bug

[MrGrim]

Nice ideas for belt optimization!

I am very sad that its Feb now, first Feb FFF is out and you still didn't say a word about 0.15 planning. You guys know that people are looking forward to it quite a lot and I guess many of those looking forward expected at least some information in this fff today. I understand that stating something and then delaying might piss some people off, but not communicating at all and just plain ignoring such questions doesnt make anybody happy either. Smart Belts --> Coolbeans! No info about 0.15 --> sad.

No need for anger. Wube is following a long and grand tradition of "when it's done" release schedules, from the man who popularized the concept (John Carmack when he was at id Software) to other top tier game devs such as Blizzard and Valve. I suppose the youngsters around here haven't had time to appreciate that tradition.

It will be out when it's done, and it'll be better because it wasn't rushed to meet some arbitrary deadline.

[aober93]

news sounds cool

[SaintFlow]

Nice ideas for belt optimization!

I am very sad that its Feb now, first Feb FFF is out and you still didn't say a word about 0.15 planning. You guys know that people are looking forward to it quite a lot and I guess many of those looking forward expected at least some information in this fff today. I understand that stating something and then delaying might piss some people off, but not communicating at all and just plain ignoring such questions doesnt make anybody happy either. Smart Belts --> Coolbeans! No info about 0.15 --> sad.

No need for anger. Wube is following a long and grand tradition of "when it's done" release schedules, from the man who popularized the concept (John Carmack when he was at id Software) to other top tier game devs such as Blizzard and Valve. I suppose the youngsters around here haven't had time to appreciate that tradition.

It will be out when it's done, and it'll be better because it wasn't rushed to meet some arbitrary deadline.

Well, I suppose you don't refer to me as a youngster. As that would be wrong. However, yes, I know who John Carmack is. And that all studios you referred to are AAA Studios with many, many more people working on games and much bigger shareholders. Again, it is not about rushing, but about communication. The last minor update upto the bugfix today was months ago. If 0.15 is not on track for the Feb timeframe they anticipated, I am pretty sure they know by now. What's wrong in communicating that? I think people are way more relaxed the earlier they get to know such things.

[commandoby]

Скорей бы 0.15. =)