The online racing simulator
TEST PATCH 0.6H2 (now H11)
(391 posts, closed, started )
Quote from Inouva :i dont know what you mean, i search LFS and the first result is lfs main webpage, also seccond and third link show me wiki and lfswiki

Quote from RC-Maus :without word lfs

Obviously he ment when searching for race sim or with similar phrase.
Quote from Inouva :
Quote from RC-Maus :Hi Scawen,

Maybe offtopic
I have been searching for lfs on google without word lfs and I found out that lfs is impossible too find on internet.

I only find other race games and sims.

Maybe thats a thing too work on? Get on google Thumbs up

Anyway I thought I mention it too you Smile

Greets Storm

i dont know what you mean, i search LFS and the first result is lfs main webpage, also seccond and third link show me wiki and lfswiki

Without word lfs or live for speed
race sim online, position 3 (4 on .com). Race sim free demo. position 5 (6, ...). Quite off topic for a test patch this discussion. Wink

Although with a higher rate of patch releases you get more often discussed on other websites, more links, higher ranking.
Quote from RC-Maus :
Quote from Inouva :
Quote from RC-Maus :Hi Scawen,

Maybe offtopic
I have been searching for lfs on google without word lfs and I found out that lfs is impossible too find on internet.

I only find other race games and sims.

Maybe thats a thing too work on? Get on google Thumbs up

Anyway I thought I mention it too you Smile

Greets Storm

i dont know what you mean, i search LFS and the first result is lfs main webpage, also seccond and third link show me wiki and lfswiki

Without word lfs or live for speed

typed online racing simulators, first link simraceway.com, seccond link lfs main webpage, Third link iRacing
If you're looking for real-world results, do your test searches in a private browser window - google will put LFS near the top if you've searched for it before.



Back on topic, I've noticed a bug: when placing chalk, it's invisible until I optimise the layout objects.

This is currently happening in H4 on Westhill; I haven't tested H2/3 or different tracks yet, but it's showing up fine to people on H.
okay okay Smile I didn't use online or free demo Smile
Quote from Degats :This is currently happening in H4 on Westhill; I haven't tested H2/3 or different tracks yet, but it's showing up fine to people on H.

Thanks - that is the case with chalk objects placed on any track, in H4 only.
-
(valiugera) DELETED by valiugera : Spam.
-
(Flame CZE) DELETED by Scawen : reply to a post that is now deleted
Any way to increase maximum autocross objects number more than 1800 ? Smile Of course in last version of patch Smile If it doesn't ruin anything.
Quote from nacim : but I also think that a 5% improvement optimisation can wait for lasts test patches (before stable version 0.6I). I guess it's more logical for me to add new content and features and then optimise it,

Eehh no. There are already enough people who leave the server when Westhill gets voted because they have lower framerate as on the other tracks. I still don't get why the CPU (/thread) consumption is about 150-200% higher as on the other tracks. Seems the reason is this higher texture quality... But isn't this why they invented GPU's to take care of that..
If you would read Scawen's responses, you would know it's because of much more objects placed in Westhill environment. So there's much more to dig trough on the CPU to consider what has to be drawn and where and upload that to GPU. Even when textures and meshes are usually preloaded, there's still some serious upload each frame just to give the GPU processing commands which/how meshes should be transformed and drawn, can easily reach MBs of memory copied between CPU vs GPU. You may reason the 3D engine of LFS is not that optimal, than again, do other 3D games with similar possibilities run that much better? I'm not aware of any PC race sim with better graphics and FPS performance.
Quote from Ped7g : can easily reach MBs of memory copied between CPU vs GPU.

Still don't understand why this needs to happen. I have 4 Gb of VRAM right next to the GPU. CPU shouldn't be busy with 3D stuff and certainly not with transferring loads of data back and forth between CPU and GPU. But hey I am just a regular guy, zero experience with 3D engine gaming designing but which notices there is quite a big problem concerning CPU usage. I could never run LFS with an O.Rift in this situation and I thought this was the whole theory behind this. High texture graphics to suit the O.Rift well..

Anyway... There seems to be no fast fix to lower CPU usage or make use of more CPU threads. So why not make two Westhill or even three Westhills? Low, mid and high quality texture mode?

If you want to compare it to other games, I think other games/sims do it like this as well.

And it's not for me, I could throw more money at it to get a newer CPU which performs better on single threads with ease, I am talking in favor of the people which cannot easily do that or, don't want to do. And then we also have the group of people which doesn't even realize that their CPU thread gets maxxed out.
Because it's not problem with texture quality, but amount of triangles in view, and internal representation of those. If you have 4GB of VRAM, lowering textures quality will gain you probably just 1-5% of GPU performance, and will do exactly zero to elevate CPU burden.

"There seems to be no fast fix to lower CPU usage or make use of more CPU threads"

Exactly. No easy fix. Except remodelling the Westhill in some low poly mode, like SEGA arcades from ~1995... Wink Or reprogramming whole LFS engine to be multi-thread. Actually remodelling content in low poly would be easier task, than rewriting the engine (just to give you idea, what kind of task is ahead of Scawen, because unless he drops LFS development, one day he will have to do that anyway, probably much later, than sooner). But then again, people already moan the LFS doesn't look up to date enough, so the low poly version wouldn't satisfy them?
I did some experimentation, seeing how long some calls to DirectX were taking and where there was long waiting. I found that a good frame rate increase was possible in many places at Westhill by allowing the program to go ahead and start rendering one more frame before the last one is completed. Really, it's the way it should be done and is basically standard practice. But unfortunately these big frame rate increases were in places where the frame rate was already high. At the slowest places, there was no improvement.

I noticed the graphics card fan was spinning up at the new high frame rate places but slowing down again at the low frame rate places.

This shows that it is the CPU that we are waiting for at the low frame rate locations, and the graphics card is taking a bit of a break there. It seems wrong that the card should be working less hard at the highest detail places, so I have been looking at ways to improve this.

I have done quite a lot of experimenting, with varying results. The best improvements have been in the last few days. I have done some sorting of the objects in the static vertex buffers and index buffers. This results in less switching between textures while rendering the world objects and there have been noticeable increases in most places.

That job is still not finished. I know of more ways to decrease the switching between textures (calls to SetTexture) so should do them. There are also a lot of draw calls (DrawIndexedPrimitive) at the high detail locations. The number of draw calls could be reduced by better sorting of the objects, so that the triangles of the objects that are likely to be seen at the same time, are positioned consecutively in the buffers. When they are positioned consecutively they are drawn with a single call instead of multiple calls. This sorting is done quite roughly for some objects at the moment. Actually some types of objects are sorted very well but some others are not well sorted. This could also be improved.
Good stuff, sounds promising!
Meanwhile, is there any mileage in the band-aid that Dave proposes, i.e. releasing a decimated version of Westhill?
(Only worth doing if it's semi-trivial for you guys, which I'm hoping it would be.)
I don't think we can do that kind of thing. Eric's busy on the S3 tracks. I'd like to just get this improved performance + Rift support patch finished and released so I can get on with some other things that need to be done for the S3 update.
No worries. (In fact, music to my ears!)
Quote from Scawen :I don't think we can do that kind of thing. Eric's busy on the S3 tracks. I'd like to just get this improved performance + Rift support patch finished and released so I can get on with some other things that need to be done for the S3 update.

Wow that is good news happy with it SmileThumbs up
I'm trying to think of a valid reason to keep Minimum Sleep option. What I mean is there really isn't any benefit from it, other than maybe reducing produced heat, but that is already possible by limiting FPS which is far better option.

As far as I know most of other games do not have similar option and to my experience that option often confuses people. Some pump it up (logic: more is better) and then cry about low FPS.
In some old thread, I came across Scawen explaining why it's there - to allow Windows to do a spot of message handling etc... Well, that explains the presence of the sleep, not the fact that it's a user-settable parameter. Indeed, other games don't expose such an option.
(Last time I programmed a time-critical Win32 app, there was only one CPU to deal with so I'm not sure about this but you'd really think that most systems today could handle LFS never yielding at all, provided they have multiple cores... But maybe that's now how Windows works.)
Like Neilser, I don't know if a minimum sleep is needed any more. Do dual core computers or later versions of Windows neglect their vital duties? Can you still buy a PC with single core CPU? The problem that we sometimes saw in the past looked a bit like forgetting to beat your heart or breathe because you are busy thinking about something else. But it's a vague memory now...

Daniel, I like the idea but what frame rate limits are actually good? Trying them I only found the 100 fps option to be any good. The current granularity of the setting comes from the 1 ms timeGetTime and Sleep calls. Maybe it should be allowed over 100 fps now?

Now that my latest version can get much higher frame rates in places and the graphics card fan spins up loudly, I'm becoming a bit concerned about overheating cards. Maybe the frame rate should be limited by default?
Good work, test patch looking good at this moment. Hope all the necessary things get done with ease so we can look forward to a new patch.
Quote from DANIEL-CRO :I'm trying to think of a valid reason to keep Minimum Sleep option. What I mean is there really isn't any benefit from it, other than maybe reducing produced heat, but that is already possible by limiting FPS which is far better option.

As far as I know most of other games do not have similar option and to my experience that option often confuses people. Some pump it up (logic: more is better) and then cry about low FPS.

I got considerable improvements with Minimum Sleep option (0 in case), FPS limiter not made much difference to me. Maybe you have a better computer and this makes it difficult to evaluate between your and weaker computer. I think Minimum Sleep should remain, also because many users have a similar system.
Quote from Scawen :Maybe the frame rate should be limited by default?

That sounds good, I very rarely let any game go fully berserk on my GPU without some kind of framerate limiter to save my ears.

Are you thinking about doing it by V-sync or manual framerate limiter by default? I was wondering how the 144Hz G-sync/Freesync monitors would behave.
Quote from Abone :I got considerable improvements with Minimum Sleep option (0 in case), FPS limiter not made much difference to me. Maybe you have a better computer and this makes it difficult to evaluate between your and weaker computer. I think Minimum Sleep should remain, also because many users have a similar system.

Daniel's suggestion is that the minimum sleep option should be removed, and this would be the same as setting it to 0.

So I don't think it is really a performance issue. The question is more like "Does minimum sleep 0 cause problems on some computers?"

Quote from Matrixi :Are you thinking about doing it by V-sync or manual framerate limiter by default? I was wondering how the 144Hz G-sync/Freesync monitors would behave.

Maybe "full screen vertical sync" should be enabled by default. It's easy to turn off. And perhaps 100 Hz frame limitation enabled by default in windowed mode. Or maybe I should add a new option "windowed vertical sync" as well.

Is there a standard safe setting used by other games, limiting frame rate by default in some way?

I would be interested to hear from anyone who knows about G-sync and Freesync monitors and how they interact with the vertical sync option.
Quote from Scawen :
Quote from Abone :I got considerable improvements with Minimum Sleep option (0 in case), FPS limiter not made much difference to me. Maybe you have a better computer and this makes it difficult to evaluate between your and weaker computer. I think Minimum Sleep should remain, also because many users have a similar system.

Daniel's suggestion is that the minimum sleep option should be removed, and this would be the same as setting it to 0.

So I don't think it is really a performance issue. The question is more like "Does minimum sleep 0 cause problems on some computers?"


I have used 0.6H 4 every day and has been very positive in comparison to 0.6H. "Does minimum sleep 0 cause problems on some computers?" well ... the computer is like a car, if not take care of it, it will stop eventually. some people do computer dustbin, for this reason it is difficult to answer this question. I think no one will have problems, but i dont know.
This thread is closed

TEST PATCH 0.6H2 (now H11)
(391 posts, closed, started )
FGED GREDG RDFGDR GSFDG