It's a good question. If I can get the graphics ready to a condition that is suitable for this release and could then get back on the physics so that is suitable in time for Eric to have finished the tracks, then this opens the possibility of working on separate physics and graphics threads. That would be a great help and the plan would be to combine this with some means of producing visual frames at the correct visual frequency for the monitor. That would either be by using a high frequency physics update that can be interrupted at any point for a graphical frame, or by using a lower frequency physics update that can be interpolated to produce graphical frames.
Obviously the graphics needs to be done then that would make it releasable, even with the current physics and it's possible this may happen. But I really hope I can sort out the new physics so we can move on.
Higher frequency physics plus an improved sidewall model sounds like nirvana
I have often thought that a "small" increase of physics clock to 180Hz would work well for VR users at half that as well as dividing well for the rest of us are 60Hz but maybe this is an over-simplification!
Probably a stupid question, but I have no experience in programming threads:
More threads means not multi-core support. Threads are (only) sub processes of the main process? Correct?
Will multi-core be supported by LFS in the future? What are your plans for multi-core or is this already one of the items on your work list?
The background of this question is that currently my CPU or their clock limits the performance with a big AI (>20) field. Not really critical at the moment but perhaps after the graphic/shadow system update? No idea if that could be related. Do I just have to be patient when you release the next version or should I look for a higher-clocked, newer CPU (anyway)? ;-) Currently I'm using an (old) i7-2600K with 3.4 GHz.
BTW: Thanks for the two screenshots. I am very much looking forward to further pictures.
One more thing regarding upcoming graphics update. Version 0.6K brings a new reflections system - nearby objects are visible in new dynamic reflections on cars, which is amazing, but i've noticed that not all nearby objects are visible - the road surface, road markings, kerbs and the grass. Is it possible these 'objects' to be added to the environment map of the dynamic reflections? An example - two screenshots from CarX Drift Racing Online:
And a little suggestion about the Layout Editor - is it possible to increase the stored views from 10 to 20, and the maximum transition time from 4.0s to 10.0s?
It would be absolutely amazing if that is likely to happen in any of the future updates. Thank you again!
Scawen, I applaud all three of you for your dedication to LFS. I haven't played for many years now, but I occasionally come back to check on progress. And progressed it has. Can't believe it's been 14 years since I first downloaded it. I can't wait to see what S3 brings. May it continue to prosper. And perhaps one of these days i'll get back into it...
On a side note, I remember in S1 as you went down the straight on Blackwood, there was a small cloud of dust that got kicked up behind the car just before the bridge. I always thought that was a nice touch and wondered why it got omitted in later versions?
If your pc is capable of playing any other game released later than 2014 then you should be fine IF the LFS update is released today......
It won't be, nor will it be released for 'a while', (enter time period here.)
As LFS has spent a lot of time developing VR, which does not run on a potato, then when the graphics, followed by physics update is 'actually' released you will have had enough time to enter the 21st century and actually have a pc that is capable of dealing with 2014 minimum graphics requirements.
Seriously, if you want to play 'current' games on a pc, spend some money on putting together a gaming pc. Or, just play on your playstation or xbox.
LFS has run on a potato for many, many, many years. It would be nice to actually see some product that has moved on from it's current state, which has not developed or changed for many years.
Sorry, there have been developments in DoS protection........
No further comment on this.......
EDIT: FFS, I can put together a decent gaming PC for NZ$600 that will run any current gen game at 60 FPS, it isn't hard to do. This will also run LFS with VR. Do try to get your **** together if you want to play 'current' games on pc.
News Flash - technology does in fact move on, and, at 'some' stage, so will LFS min requirements.
First, a thread is not a process. A process is what you would see in the Windows Task Manager. One of the hidden-by-default columns in Task Manager is the number of threads that a process is using.
It is very possible to have multi-core support by using multiple threads working in parallel, however just having multiple threads doesn't mean that you would get any benefit by running them across multiple cores. They must be able to perform tasks in parallel, somewhat independently.
One of the easiest ways of doing this for games is to split different tasks that do not depend on eachother (in the current tick at any rate) - splitting off AI, graphics and physics into their own threads is probably a good example, though there are some complications.
Alternatively, having a thread per car (or a small number of cars) would be another way to relieve the single-core burden of physics and AI, as - apart from car to car collision detection - the different cars are independent from eachother, which allows the different threads to easily work in parallel.
Depending on how the different threads are used and interact, there will be various limits on how many CPU cores can be effectively utilised.
Contrary to what many web browser developers would have you believe, you do not need multiple processes to have multi-core support. In fact, having multiple processes is usually less efficient than multiple threads in a single process.
I believe Scawen has said that multi-core support is already on the list of things to do, but I can't find the quote. I did find this one:
Also before 0.6T it was possible to place an object using IS_AXM with UCID of a player that is not an admin, in 0.6T we get "IS_AXM (PMO_ADD_OBJECTS) - UCID is not an admin" error.
Was this change intended?
I know we can set UCID of an admin or host to place an object but for example LFSLapper addon was using UCID from AXM info to detect which player placed the object. So this is no longer possible since AXM info would return "fake" UCID.
When you are playing online in LFS, there are two servers involved. One is "master" server of LFS devs, validating your license and collecting data about actual race-servers (their addresses, mode, configuration, available cars, etc), allowing you to search for some server. The other server is the actual game server, where the race is going on, and belongs to people who provide game servers.
Now if you attack the master server, and force it to choke on the traffic or crash, basically nobody can connect to any game any more. If you attack particular game server, then only the players trying to play on that server are affected.
The latest patch was about strengthening the master server, so there's no single-point-of-failure bringing down all games/players.