Thank you all for the feedback. I intend to answer various questions when I run through all the posts since H6 release, making notes of the remaining issues.
Just a could of questions about recent posts.
What method do you use to minimise your full screen LFS? I can get similar results to what you are saying by using ALT+TAB. It seems LFS doesn't get the usual messages from Windows that it has been minimised in that case. I'll look into how LFS can detect that it is no longer in full screen mode after ALT+TAB. But please let me know if you have a different method. The high CPU usage is presumably because the frame rate limitation is by waiting for vertical sync but DirectX is never waiting for the vertical sync, but somehow it's letting LFS do all the image rendering as if it was still full screen...
Which bug are you talking about? It looks as if you are replying to Flame's post, but I can't see how the bug he found could cost you a league race.
Sorry to hear that a bug cost you a league race! Better make sure I know what it was!
I frequently use CTRL-ESC, I think it's similar to this dedicated Windows start button on keyboards. When back in Windows environment, if the "minimize all" button is used in W7 right next to the time in the right lower corner then everything gets minimized to the taskbar but LFS still runs with the same CPU power.
Equally, when you leave LFS in windowed mode and ALT+TAB to another window, LFSs CPU usage rockets. I realised it doesn't just occur when minimised but I guess it's all related.
I do seem to have some significant FPS increase on my Pentium 4. I finally have enough FPS to drive around and enjoy the new Westhill . Thanks! Keep up the great work guys
I could replicate vitaly's bug right from the beginning of the replay. Whenever he's not in the lead, his dashboard disappeared. It appears to be based on proximity to other cars. If you spectate cars further back in the field, when they bunch up again their dashboards vanish at ~2-3 car lengths.
OK, I have been through the thread since the H5 release note and made notes of bugs that I need to fix and issues I need to investigate. Thank you for the feedback.
Some things I want to answer or ask more about...
I've noticed that my skins are visibly more smudged in newer patches, compared to pre-new Westhill releases.
This is now fixed in H5.
The graph works well. I get something in the Wait field only sometimes (and I can't reproduce this), or when VSync is enabled. What does it show exactly? Waiting for GPU?
The "Sleep" column is the sleep milliseconds that are used for frame rate limitation. This takes place at the start of the game loop, before the physics updates. Also if "Sleep every frame" is set and there is no other Sleep or Wait then there will be 1 ms Sleep shown in this column. If you use frame rate limiting then you will normally see something in this column. Sleep is time deliberately given up to the operating system each frame because LFS doesn't need to do anything.
The "Phys" column shows the number of physics steps calculated in the frame, before moving onto drawing the visible representation of the game state. These are the physics steps that represent 0.01 seconds of game time. So if you have frame rate limited to 100 Hz and you are not paused then you should see one physics step per frame, which will look like a solid vertical orange line. (This column does *not* show milliseconds)
The "Wait" column is the number of milliseconds of Sleep while LFS is waiting for the GPU to finish rendering. If you do not limit frame rate then you may see something in this column. In that case you are more likely to see something if you set "Maximum buffered frames" to zero. LFS is then waiting for the GPU to finish before starting to render a new frame. If you use buffered frames then it may be you do not see anything in the Wait column. You should not add more buffered frames if there is no waiting, because it that case you would just be adding lag. If you use full screen vertical sync and you are in full screen mode then you will see Wait. In this case you are waiting for the GPU to present the image to the screen, just in time for the monitor refresh.
Phys in graph isn't exactly just car physics as far as I figured, but it includes time to handle interface and stuff.
Just a count of game updates, as described above.
While messing with that graph I noticed that everytime I activate LFS window I get stutter almost a second long. What I mean is your LFS is on half on screen and on other half you have other program which is active, when click on LFS to activate it that stutter happen. Anything possible to do with it?
I don't get that stutter. It seems to seamlessly switch to LFS. I see a slight lack of Sleep in the graph but I don't get a visible stutter or lag. Could there be circumstances that produce the lag, like being online or having a lot of AI?
However i did find a "funny" texture missing error once at WE1X see attached.
I found it both in H and H5. i did try to reproduce in a second run without luck, but it is happening in the attached replay (time 2.51). I am not sure if this is a know bug i just wanted to report it.
Yes, that is a known problem when you take certain lines on Open Configurations, it doesn't always switch quickly enough to the most appropriate path that knows which objects you should see from your location.
Little offtopic though, but if we are talking about FPS optimizations and so, there is a little freeze every time when I switch to the Game page in the Settings. If I visited the Game page in the current session already and I open it up again, there is no freeze (probably because it's cached already).
This bug is an old one (as long as I know it) and not introduced with the recent (test) patches.
Maybe this one is easy to fix?
That is because of generating the Chinese, Japanese and Korean characters which must be cached on textures.
Upgraded from Win 7 to Win 10 and first time I started LFS on Win 10 it crashed.
Unhandled exception at 0x085DD54E in LFS H5.exe: 0xC0000005: Access violation writing location 0x0E06000C.
Looks like a crash address outside of LFS. Is there any way to reproduce this bug and get a call stack at the time of the crash, so I can track it back to LFS?
I have some temperature issues after applying the new patch. My laptop is normally running on pretty high temps, but this is extreme. My GPU usually goes up to 75-80°C, but with new patch I got a max temperature of 127.5°C which concerns me a bit. I will try the old patch later and report back with the max temp.
One thing I think you didn't mention - what is the frame rate in the new patch compared with the old? As some people said, LFS can draw using less CPU now and this means your graphics card can work harder as it's not just sitting around waiting for instructions. I guess you have a higher frame rate and this is why your card is getting hot. With the latest changes, I now recommend a limited frame rate. Better you decide what the frame rate should be and let the card have a little time off each frame, instead of just having the highest frame rate possible at all times (which will always vary depending on what you can see).
Exactly that, ALT+TAB.
Equally, when you leave LFS in windowed mode and ALT+TAB to another window, LFSs CPU usage rockets. I realised it doesn't just occur when minimised but I guess it's all related.
As mentioned, I can reproduce high CPU usage when minimised, after ALT+TAB from full screen vertical sync.
But this last one you have described is mysterious. If LFS remains in a window and you move to another window, LFS CPU usage really should not go up. I wonder if this is a Windows 10 thing? Is this visible in any way using the frame info graph in Misc Options?
Basically CPU usage is always there unless the minimize to system tray button is used. If LFS is being alt-tabbed and not on the foreground anymore it still does the CPU cycling. (W7).
Comparing H5 to H this has changed. If LFS_H1.exe is being alt tabbed from fullscreen to, for example, this browser... CPU usage of LFS_H1.exe drops to near zero.
Equally, when you leave LFS in windowed mode and ALT+TAB to another window, LFSs CPU usage rockets. I realised it doesn't just occur when minimised but I guess it's all related.
As mentioned, I can reproduce high CPU usage when minimised, after ALT+TAB from full screen vertical sync.
But this last one you have described is mysterious. If LFS remains in a window and you move to another window, LFS CPU usage really should not go up. I wonder if this is a Windows 10 thing? Is this visible in any way using the frame info graph in Misc Options?
I ran the LFS benchmark replay on Max settings in both H and H5 versions, in Windows 7 and Windows 10 and as other people reported somewhere, FPS is lower in the newer OS. I'm not sure why.
I wonder if D3D9 could be slow in Windows 10. As I understand it goes through a sort of translation layer (that started in Vista). Microsoft might not bother to make this any good in Windows 10, either because they feel it's a waste of time, or possibly a deliberate tactic to force developers to use the latest DirectX, in turn forcing consumers to upgrade, etc...
If it's easy for you to swap between Windows 7 and 10, maybe we can get a clue like this:
With frame rate unlimited, and game paused, go to a SHIFT+U mode position in Westhill where most of the world can be seen and that has a low frame rate. Save that position with SHIFT+NUMBER. The idea is that there are a lot of DirectX calls at this point. Now limit the frame rate to a level that will be sustainable at that position in Windows 7 or Windows 10.
Then get a task manager reading of LFS CPU usage at that same location, with that same frame rate, in both Windows 7 and Windows 10.
My thought is that it could be interesting to see how much CPU usage is required to produce the same result. Not that I can actually do anything about it!
I don't get that stutter. It seems to seamlessly switch to LFS. I see a slight lack of Sleep in the graph but I don't get a visible stutter or lag. Could there be circumstances that produce the lag, like being online or having a lot of AI?
Neither being online or AIs noticeable changed lag. Attached screenshot with my Graphs when that happens.
Looks like a crash address outside of LFS. Is there any way to reproduce this bug and get a call stack at the time of the crash, so I can track it back to LFS?
When I upgraded to Win 10 I left my GPU drivers intacted, now downloaded and installed GPU drivers for Win 10 and it seems to be fine.
I ran the LFS benchmark replay on Max settings in both H and H5 versions, in Windows 7 and Windows 10 and as other people reported somewhere, FPS is lower in the newer OS. I'm not sure why.
This is bugging me for years already...
Using same replay on Windows 7 I've got 156 FPS, while on Windows 10 I get 125 FPS. Performance seem to decrease with every new OS. I got significant increase in Windows XP, which seems like the best option if you are only about LFS.
Everyone on the Internet brag Windows 10 regarding performance, why is then performance in LFS going down with every new OS?
As mentioned, I can reproduce high CPU usage when minimised, after ALT+TAB from full screen vertical sync.
Perversely, my Win XP box now behaves better with vsync than it did with 0.6H. If you recall, I saw CPU usage rise to a full core's worth any time I used vsync, despite the rate-limited 60 fps CPU load being really low. NOW, with H5, I'm seeing "normal" CPU load with vsync while driving, and it also doesn't rise when I minimise with alt-tab. (I forgot to try windowed mode - my testing so far is all full-screen.) Perhaps this will help you to understand what's going on!
Thanks for explaining the sleep/phys/wait columns btw - that helps a lot. I had a somewhat confused idea of what they meant Will look at them again tonight with my new understanding.
Equally, when you leave LFS in windowed mode and ALT+TAB to another window, LFSs CPU usage rockets. I realised it doesn't just occur when minimised but I guess it's all related.
As mentioned, I can reproduce high CPU usage when minimised, after ALT+TAB from full screen vertical sync.
But this last one you have described is mysterious. If LFS remains in a window and you move to another window, LFS CPU usage really should not go up. I wonder if this is a Windows 10 thing? Is this visible in any way using the frame info graph in Misc Options?
problem! i play lfs on a laptop, sony vaio windows 8. works perfectly fine. but the thing is that i get 250 fps if i put the laptop in battery mode and i lose 100 fps if i put back AC adapter. i noticed something changing in this thing, again the game is only textures mod i guess its fine. look at this image. in the first image, i see a lot of activity in "wait" but zero activity in "sleep", in the second pic its the other way around. i first suspected that the difference comes from power plan but i checked that and its not that.
in battery mode my cpu fan is louder, strangely. if anyone who had this problem solved it help me please.