[Canceled] spring peformance increase

[Canceled] spring peformance increase

A place for offers of Bounties and discussion of funded proposals.
Post Reply
Super Mario
Posts: 814
Joined: 21 Oct 2008, 02:54

[Canceled] spring peformance increase

Post by Super Mario » 20 Jan 2015, 05:09

Proposal:
The basic idea behind this is to increase performance by better usage of multithreading, and general optimization. Without having game developers required to change their code in order to benefit any increase in general performance.

Current status:
Developer yet to be determined nor specification on how to increases spring performance.

Current pledges: total 0$


I would like to ask that this thread only for active spring devs, people who had pledge, and people who are interested in supporting this by considering pledging, providing tests, and contributing to the code.

Mods, feel free to update this post with information. :-)

Yeah, this is never going to work out in the first place.
Last edited by Super Mario on 19 Jun 2018, 17:23, edited 2 times in total.
0 x

Super Mario
Posts: 814
Joined: 21 Oct 2008, 02:54

Re: [WIP] spring peformance increase

Post by Super Mario » 20 Jan 2015, 05:17

Also, Anarchid, googlefrog I hope you don't mind that I interpret your request of fixing the engine to be like it was when it's 0.91, to general performance increases. If you disagree about, please replay and I will remove you from the pledged list if you request me to.
0 x

gajop
Moderator
Posts: 2972
Joined: 05 Aug 2009, 20:42

Re: [WIP] spring peformance increase

Post by gajop » 20 Jan 2015, 05:18

This is a pretty bad proposal (it's way too ambiguous). It needs to be more specific in exactly how much and what kind of performance should be improved. This can be discussed here, but looking at how these topics usually work, I have some doubts on that.

Also note that there has been certain discoveries regarding the effect of locks, which should put development engine on the same level as 91.x for windows people and thus I'm not sure those offers by GoogleFrog and Anarchid are still valid.
0 x

Super Mario
Posts: 814
Joined: 21 Oct 2008, 02:54

Re: [WIP] spring peformance increase

Post by Super Mario » 20 Jan 2015, 05:36

gajop wrote:This is a pretty bad proposal (it's way too ambiguous). It needs to be more specific in exactly how much and what kind of performance should be improved. This can be discussed here, but looking at how these topics usually work, I have some doubts on that.

Also note that there has been certain discoveries regarding the effect of locks, which should put development engine on the same level as 91.x for windows people and thus I'm not sure those offers by GoogleFrog and Anarchid are still valid.
I already understand that when I wrote that post.I was planing to use this thread for discussing the specific details on this use. My knowledge of the engine is still rookish/noobish/fragile.

For me, having the spring use more than two core processors is my general basic idea of performance improvements.
0 x

hokomoko
Spring Developer
Posts: 530
Joined: 02 Jun 2014, 00:46

Re: [WIP] spring peformance increase

Post by hokomoko » 20 Jan 2015, 06:45

(EDIT to make thread make sense):
I hope this won't be relevant because the "recent discoveries" will indeed be integrated into the engine *nudge* *nudge* :P .
Last edited by hokomoko on 20 Jan 2015, 18:23, edited 2 times in total.
0 x

Super Mario
Posts: 814
Joined: 21 Oct 2008, 02:54

Re: [WIP] spring peformance increase

Post by Super Mario » 20 Jan 2015, 07:48

*deleted*
Last edited by Super Mario on 20 Jan 2015, 15:18, edited 2 times in total.
0 x

User avatar
Anarchid
Posts: 1372
Joined: 30 Nov 2008, 04:31

Re: [WIP] spring peformance increase

Post by Anarchid » 20 Jan 2015, 10:47

I would consider this obsolete as the LockFreeLua branch achieves technical performance specs on par or better than 91.0 has for ZK benchmarks, even while those are using an older version of ZK that doesn't have many important optimizations such as raw move for UnitAI.
0 x

Super Mario
Posts: 814
Joined: 21 Oct 2008, 02:54

Re: [WIP] spring peformance increase

Post by Super Mario » 20 Jan 2015, 15:16

Anarchid wrote:I would consider this obsolete as the LockFreeLua branch achieves technical performance specs on par or better than 91.0 has for ZK benchmarks, even while those are using an older version of ZK that doesn't have many important optimizations such as raw move for UnitAI.
OK, will edit OP.
0 x

luckywaldo7
Posts: 1397
Joined: 17 Sep 2008, 04:36

Re: [WIP] spring peformance increase

Post by luckywaldo7 » 20 Jan 2015, 18:55

I will pledge double whatever Super Mario pledges to the bounty.
0 x

User avatar
PicassoCT
Journeywar Developer & Mapper
Posts: 9950
Joined: 24 Jan 2006, 21:12

Re: [WIP] spring peformance increase

Post by PicassoCT » 20 Jan 2015, 22:46

The future Engine bottleneck are physics, in particular unithpyhsics and particle physix. Lups has taken some particles and sfx to the gfx card allready.

So what would raise performance is to get all of unit animation and drawing allmost completely to the gfx card, which is partially problematic cause it would have to happen in sync.
The idea is to not look for what other software does for optimization (server software lives and breaths from good multithreading- which we got via threadpools by the way) but to look on the bottlenecks.

Even when this is more difficult.

Image
You do not have the required permissions to view the files attached to this post.
0 x

Post Reply

Return to “Bounty Offers”