Great game. I have some feedback.

Describe your experience with the latest version of FreeOrion to help us improve it.

Moderators: Oberlus, Oberlus

Forum rules
Always mention the exact version of FreeOrion you are testing.

When reporting an issue regarding the AI, if possible provide the relevant AI log file and a save game file that demonstrates the issue.
Message
Author
User avatar
vincele
Space Dragon
Posts: 341
Joined: Sun Mar 23, 2014 6:10 pm

Re: Great game. I have some feedback.

#16 Post by vincele » Mon Jan 26, 2015 1:12 pm

Dilvish wrote:
Moriturus wrote:Other feedback: The interface is very sluggish. When I move the mouse, the pointer trails behind my movement so badly that it sometimes takes a full second to be sure I'm pointing at what I want to point at.
Hmm, the game used to get bogged down quite a bit, but that's been cleared up for well over a year now, and well before your version. I don't recall similar complaints for quite a while.
It still is for me on my underpowered linux laptop (my only coding computer) so much that I don't play on it, I can only run the game to test my code changes... I sometimes play on my mother's windows 7 computer, which is better HW-wise, and this is with prebuilt binaries. But still, I think there are quite a lot of performance problems in FO. And I know some of them will be very hard to fix.

Correct me if I'm wrong, but those are my perceived sources of slowness :

1 - screen is redrawn for each frame (bandaid : can be reduced by reducing FPS in settings, but would be really nice to fix for old HW)
2 - lots of computed things when opening views, changing system, etc... I think this could be cached so as to be only computed once per-turn (probably hard to do/fix without introducing new bugs)
3 - if you have a huge universe (lots of stars, ships, etc...) the game slows down a lot. This can be seen in late game, which runs a lot slower than at the start. Maybe we have some > O(N) algorithms hiding somewhere.
4 - AI is CPU-hungry (quite normal) and this is probably not the main point in felt sluggishness, as it happens between the turns, when the player knows he has to wait for the new turn.

I think the mouse pointer sluggishness is linked to item 1. This would be my main target if I had the required knowledge...

I even think some things are recomputed for each frame, so problems 1 and 2 may be linked somehow. But that's only suspicion, no actual evidence on that.
All the patches I'll provide for freeorion will be released under the GPL v2 or later license.
Let's unleash the dyson forest powa!

Mitten.O
Programmer
Posts: 255
Joined: Sun Apr 06, 2014 4:15 pm

Re: Great game. I have some feedback.

#17 Post by Mitten.O » Mon Jan 26, 2015 5:24 pm

1 - screen is redrawn for each frame (bandaid : can be reduced by reducing FPS in settings, but would be really nice to fix for old HW)
Redrawing the screen is not in itself a problem for hardware accelerated graphics. The problem is that all geometry is a) retransferred to the GPU every frame b) transferred using one api call per vertex instead of using buffer operations.

I think b) could be fixed with just a significant amount of work (by caching all calls in a frame to buffers and then sending them to the card at once), whereas fixing a) would require a fundamental change in the logic of GG, which would be almost equivalent to changing the graphics back-end entirely. (Windows would have to see themselves as collections of primitive shapes which the system draws, instead of having a draw function themselves).
Any code by me in this post is released under GPL 2.0 or later.

User avatar
Kassiopeija
Dyson Forest
Posts: 212
Joined: Wed Jul 16, 2014 6:14 pm
Location: Black Forest

Re: Great game. I have some feedback.

#18 Post by Kassiopeija » Mon Jan 26, 2015 5:45 pm

Vezzra wrote:The problem here is, we can't show habitability of a planet for "your species", as there isn't such thing as "your species". One of the fundamental design decisions for this game has been that empire != species. There is no "standard", "default", "main" (or whatever you want to call it) species for your empire, just a starting species. But this starting species is in no way different to the other species that get integrated into your empire during the course of a game, it's just the species of the planet you start with, that's all.

So for which of the species that belong to your empire should the habitability of a planet be shown on the sidepanel? There is no obvious preference, and showing it for all of your species is impossible. The only option we could maybe consider is showing habitability as long as you've only one species in your empire that is able to colonize.
IDK if it is planned that this right-click menu will show additional subjects in the future, but as of now, it only contains two items, and I assume not many people use the rename planet option alot. I also find the right-click to check on planet habitability tiresome. Actually I love micromanagment and most games start very interesting, but at around 100 planet count alot of micro makes turns take too long and repetetive. If I may suggest a few things:

The "rename planet" option could be toggled by simply left-clicking on the planet name.
The "habitable planet" popup could be toggled by left-clicking on the planet pic. Or just make two buttons for these in case someone if playing without planet pictures...

Once terraforming/gaia has been researched, I've got the hell lot to do to go to each planet individually and check his preferences and enqueue terraforming manually. A civ-wide option "automatically enqueue terraforming if available" would be really nice, maybe with a suboption like "and keep building it until gaia is reached" or "no terraform/gaia is available anymore" etc.

Rallyepoints would be nice, or another way to tell specific ships where they should automatically heading once they are build. No idea how that could look but I'm facing two "problems":

I don't play with starports but instead use engines+lighthouses among the lanes to the front so warships+troops always go the same route. In this case it would help if all ships build on a planet are sent automatically at a sytem. Or maybe the Rallyepoints could be attributed to a specific design so all my troops don't dive too far into (not yet) secured terrain....

Another thing is with midgame colonizations. When I expand militarily I encounter uninhabited planets or get a new race that now could be used to colonize planets already in my system. Whenever I encounter such a planet I enqueue a colony ship but it takes ~10 turns to build it and at that time forgotten the exact location of the planet it was ment for. It gets worse esp. when every second turn a colony ship is build and alot of colony ships are still en route, so basically I have to search for the planets yet over again and also consider if a colony ship is already en route to these systems. It even gets aggravated if in the meanwhile the maxpop changed due to techs like cyborgs or specials, because this can make another species more suitable (in terms of maxpop) and suddenly no perfect planets are available for the built colony ship.... so basically I need a feature to send a single ship to a specific system and be able to specifiy that in the buildqueue. Or maybe someone knows how to create duplicate colonyships albeit with another name, just as "Organic Colony for Formalhaut 1" etc but the game won't let me

User avatar
Vezzra
Release Manager, Design
Posts: 5186
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Great game. I have some feedback.

#19 Post by Vezzra » Mon Jan 26, 2015 5:54 pm

Kassiopeija wrote:Another thing is with midgame colonizations...
This is currently worked on. Revised colonization mechanics are already in the process of being implemented, you can follow discussion/progress here.

User avatar
Kassiopeija
Dyson Forest
Posts: 212
Joined: Wed Jul 16, 2014 6:14 pm
Location: Black Forest

Re: Great game. I have some feedback.

#20 Post by Kassiopeija » Mon Jan 26, 2015 6:17 pm

Vezzra wrote:If I play on a huge map, research costs for all techs will be much higher. But the advantages you get from each tech will be exactly the same.
But they will be attributed to many more planets so the sum of total prod/res will rise exponentially and not linearily. If I have 10 planets under my control researching Adaptive Automation & Nascent Artifical will bring me +50prod/+20RP while only 5/2 with my sole starting planet, although the cost of these techs stay exactly the same. The problem is that if research costs are weighed versus available planet count right from the start then even the first available techs will become quite expensive, and the game very slow.

Research cost could be based on the number of colonized planets, or, could be based to let techcosts inflate once certain breakpoints have been reached, ie. after 500 RP all remaining available techs become 5-10% more expensive, next step will be at 2500 RP etc. Crucial techs (such as new weapon-tiers etc) could see the max inflation while non-crucial like pre-requisites etc could see the minimum inflation.
Such a concept is interesting in a way that techs that are prioritized are more easy to get while techs which have been ignored are penalized in a way.

User avatar
Vezzra
Release Manager, Design
Posts: 5186
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Great game. I have some feedback.

#21 Post by Vezzra » Mon Jan 26, 2015 7:17 pm

Kassiopeija wrote:But they will be attributed to many more planets so the sum of total prod/res will rise exponentially and not linearily. If I have 10 planets under my control researching Adaptive Automation & Nascent Artifical will bring me +50prod/+20RP while only 5/2 with my sole starting planet, although the cost of these techs stay exactly the same.
That only applies to techs that grant boni per unit of pop or per planet. The majority of techs doesn't fall into that category - many, yes, but not most.

The other problem is that only techs get more expensive while yielding the same results. Production still works the normal way: you get the same thing for the same amount of PP, regardless of map size. PP cost don't scale, RP cost do. Which makes balancing much more difficult, or even impossible (if tech costs scale too much).
The problem is that if research costs are weighed versus available planet count right from the start then even the first available techs will become quite expensive, and the game very slow.
That's one more issue with that approach. Complicates things still further.
Research cost could be based on the number of colonized planets
That's an extremely problematic approach, because it results in a dynamic where you have to pay more resources the more resources you produce (because more colonies -> higher resource output but also higher tech costs), thus making increase resource output by making new colonies pointless (at least to a certain degree), which contradicts the basic idea of a 4X game.
or, could be based to let techcosts inflate once certain breakpoints have been reached, ie. after 500 RP all remaining available techs become 5-10% more expensive, next step will be at 2500 RP etc. Crucial techs (such as new weapon-tiers etc) could see the max inflation while non-crucial like pre-requisites etc could see the minimum inflation.
Such a concept is interesting in a way that techs that are prioritized are more easy to get while techs which have been ignored are penalized in a way.
That's basically something similar to our current build-cost-increase-mechanic for ships. Generally speaking, the more you have of them, the more expensive they get. As much as I dislike this approach for building things, with research that's an approach really worth considering. Research costs increase based on already completed research, which would be something like abstracted "maintenance costs" for techs. Will help a lot that the tech tree can't be completely researched too quickly and easily. I definitely like that approach a lot more than the other suggestions so far.

Hm, what do the others think?

Post Reply