No need for sarcasm. The discussion moved to other topics and I was afraid the comment went "lost". A simple noted/thanks/etc would have been enough.
Scawen, did you notice this post?
At the mirror (tested virtual only), in "Name over cars" mode, the name only appears when the car behind is really REALLY close. On earlier versions I was able to tell the names over at least several cars behind (all LOD settings at max details), but now it's practically useless... as in.. 98% of the time the cars are just too far behind (I'd say 0.5s is enough) to display it..
E8400/4G/GTX260+/XP32-SP3. Res is 1920x1200, FSAA ranging from none to 16xQ (via the driver), AF from none to x16, both high and low res shadows, FPS changes from ~58 to ~75 over these different settings with max AI cars, paused, external view from the front backwards that enables seeing most of the grid (Fern Bay Green).
It's a bit hard to see the exact FPS since it changes so fast, but I'd say about 5% slower with the test shadows.
Interesting FPS observation: When looking from the last car forward (either cockpit view or behind-view), FPS is almost double (70-90% increase) than when looking from the front backwards toward the grid... even though on both cases the LOD setting is such that I can see all cars...
It's a bit hard to see the exact FPS since it changes so fast, but I'd say about 5% slower with the test shadows.
Interesting FPS observation: When looking from the last car forward (either cockpit view or behind-view), FPS is almost double (70-90% increase) than when looking from the front backwards toward the grid... even though on both cases the LOD setting is such that I can see all cars...
Last edited by avih, .
I don't think so. At least not to that degree. On idle my cpu is 1-2%. Same goes when listening to music etc. Not a chance the extra windows stuff takes 20-30%. My guess is that if not LFS directly, it's the GFX driver which is multi core aware. Also, on the processes list, except for LFS, all other processes together don't consume more than 2% (as in most 0%, 1 or 2 at max 1%).
I thought I should revive this thread for the sake of more info for people who might need it.
The time has come, I've replaced my system. The new one is based on E8400 Core2Duo 3GHz (P43), GTX260+ and 4G@800MHz RAM (3.3G actual, still on XP32/SP3). Even though it has a nice OC potential, the system runs at stock everything, cool and very quiet.
Using the same LFS settings as my previous post, I can now get 150-300 FPS, and usually more than 200.
When I max all the settings (16xQ FSAA via the driver, X16 AF, best LOD- had to check the manual for that, no MP speedup, high res shadows, paths, etc), I get around 80-90 at the back of a large grid and 150-200+ most of the time during the rest of the game.
When not testing, I use vsync and limit the FPS to 70 and I get a consistent 60 FPS without ever dropping a frame. Nice
LFS' CPU usage is 20%-35%, and about 25% most of the time. It does seem to use both cores though (Using Z15 patch and latest NVidia drivers), although not completely symmetrically. One core is about 20%-30% and the other about 20%-45%. I'd say it's good enough.
As suggested by my previous and current benchmarks, LFS would cope way more than enough with much weaker CPU than E8400 (X3 weaker probably), and I think that a decent GPU is contributing more. Still, at its current state, I guess GF8800 (or equivalent) would be enough for very high visual settings at very good frame rates.
Only problem with my current system is that I now started playing Grid once in a while (great fun! very dumbed down physics yet still feels great), and it completely destroyed my LFS skills... My FZR just keeps spinning and overheating the tires!
Die Grid Die!
The time has come, I've replaced my system. The new one is based on E8400 Core2Duo 3GHz (P43), GTX260+ and 4G@800MHz RAM (3.3G actual, still on XP32/SP3). Even though it has a nice OC potential, the system runs at stock everything, cool and very quiet.
Using the same LFS settings as my previous post, I can now get 150-300 FPS, and usually more than 200.
When I max all the settings (16xQ FSAA via the driver, X16 AF, best LOD- had to check the manual for that, no MP speedup, high res shadows, paths, etc), I get around 80-90 at the back of a large grid and 150-200+ most of the time during the rest of the game.
When not testing, I use vsync and limit the FPS to 70 and I get a consistent 60 FPS without ever dropping a frame. Nice
LFS' CPU usage is 20%-35%, and about 25% most of the time. It does seem to use both cores though (Using Z15 patch and latest NVidia drivers), although not completely symmetrically. One core is about 20%-30% and the other about 20%-45%. I'd say it's good enough.
As suggested by my previous and current benchmarks, LFS would cope way more than enough with much weaker CPU than E8400 (X3 weaker probably), and I think that a decent GPU is contributing more. Still, at its current state, I guess GF8800 (or equivalent) would be enough for very high visual settings at very good frame rates.
Only problem with my current system is that I now started playing Grid once in a while (great fun! very dumbed down physics yet still feels great), and it completely destroyed my LFS skills... My FZR just keeps spinning and overheating the tires!
Die Grid Die!
Last update on this subject: I've raced enough (many hours) with Z13 at above-normal priority and didn't experience the slowdown even once. So it's most definitely another process (probably windows explorer) consuming CPU cycles once in a while. I'm now always running it in a higher priority via a modified shortcut.
Is there any reason why LFS doesn't run by default in above-normal priority? After all, it does have some real-time requirements which might require a higher priority?
Raced some during the weekend on z13 trying to see if another process is getting those cpu cycles. Unfortunately, the bug didn't happen. On some of the times I upped LFS to above-normal priority, and in one of those I did hear the HD grinding noise but the framerate kept nice and steady. The process monitor was off at the time.
So, I don't know what to say. My guess would still be that it's windows explorer using CPU occasionally, and that it doesn't happen with z but does with z13. Sorry I'm unable to give better info right now. If I'll get something more concrete in the future, I'll post it.
Some more information while testing with replays only:
- Both z and z13 use about 270M of virtual memory although I once saw 400M usage (iirc with z, but it didn't happen again so I'm not sure it really was z). IMHO that's not enough to initiate swap usage on my 1G machine when all other apps closed.
- The full-screen/window change which results in very similar slowdown also occasionally exhibits the HD grinding noise, both z and z13, which might suggest that the issues are of similar origin. The symptoms really seem identical.
- Monitoring full-screen/window change using process explorer shows that windows explorer (the desktop manager process) consumes many cpu cycles for few seconds after changing the screen mode, possibly refreshing the desktop, which might also explain the HD usage while refreshing the icons. This coincide with the slowdown of LFS and it clearly shows on the graphs that LFS is going to near 0% cpu usage while explorer consumes the cpu.
- Elevating LFS process to above-normal priority (it's normal priority by default, and so is the explorer priority) solves the issue almost completely. Explorer still does it's stuff but since it's lower priority now than LFS, it takes longer to complete and LFS keeps using cpu cycles. Slight slowdown in LFS is noticable but far from the effect it exhibited in normal priority. Also, LFS is MUCH smoother in general with much less glitches (frame drops) when it runs in above-normal priority.
I still wanna play online for a considerable amount of time to see if I can come up with more concrete conclusions regarding the differences of z/z13 and the relation to explorer resource usage, specifically, if the slowdown in z13 normal-priority during play coincide with explorer activity burst.
Right now my guess would be that something triggers explorer activity, maybe more in z13 than in z. Regardless, elevating the LFS process priority shows promising results.
I'll keep this thread updated.
PS.
Trying to relate a desktop refresh to LFS, I have a shortcut to the sets directory of LFS on my desktop. This might be related, I'll try to relate to that parameter as well.
- Both z and z13 use about 270M of virtual memory although I once saw 400M usage (iirc with z, but it didn't happen again so I'm not sure it really was z). IMHO that's not enough to initiate swap usage on my 1G machine when all other apps closed.
- The full-screen/window change which results in very similar slowdown also occasionally exhibits the HD grinding noise, both z and z13, which might suggest that the issues are of similar origin. The symptoms really seem identical.
- Monitoring full-screen/window change using process explorer shows that windows explorer (the desktop manager process) consumes many cpu cycles for few seconds after changing the screen mode, possibly refreshing the desktop, which might also explain the HD usage while refreshing the icons. This coincide with the slowdown of LFS and it clearly shows on the graphs that LFS is going to near 0% cpu usage while explorer consumes the cpu.
- Elevating LFS process to above-normal priority (it's normal priority by default, and so is the explorer priority) solves the issue almost completely. Explorer still does it's stuff but since it's lower priority now than LFS, it takes longer to complete and LFS keeps using cpu cycles. Slight slowdown in LFS is noticable but far from the effect it exhibited in normal priority. Also, LFS is MUCH smoother in general with much less glitches (frame drops) when it runs in above-normal priority.
I still wanna play online for a considerable amount of time to see if I can come up with more concrete conclusions regarding the differences of z/z13 and the relation to explorer resource usage, specifically, if the slowdown in z13 normal-priority during play coincide with explorer activity burst.
Right now my guess would be that something triggers explorer activity, maybe more in z13 than in z. Regardless, elevating the LFS process priority shows promising results.
I'll keep this thread updated.
PS.
Trying to relate a desktop refresh to LFS, I have a shortcut to the sets directory of LFS on my desktop. This might be related, I'll try to relate to that parameter as well.
I usually have plenty of RAM to spare, especially with all other apps closed and LFS isn't particularily heavy on resources. It's definitely not a CPU heat issue, it's on 37-41 (c) (both diod and the other cpu sensor) and I just recently cleaned and remounted the fan and it's well greased. The CPU temp is rock solid. The GFX card might be an issue, I'm not monitoring it (I did clean the fan recently and I don't use OC), I hate ntune. Might install it though for the test. However, the symptom is not of constant lower performance. It goes from 30-100 fps to 1-2 for few secs, and then back to normal (the sound skips and HDD working also seem unrelated to the gfx card). I'd expect a heated gfx card to run somewhat (but not as drastic) lower performance for considerably longer duration. It also seem unrelated to gfx settings (resolution/AA/AF/etc).
Also, it never happened with any other game, including HL2/e1/e2 which I recently played, and is much more demanding from the gfx card and system in general IMO.
I'll try to run process explorer in the background and log all processes cpu and memory usage, see if something interesting comes up. If there's any command line args to LFS to output a verbose events log, I can use them and save the log after I encounter the issue. Are there such args?
The important thing is though, it doesn't seem to happen in z, but it does on z10/13. As I said, I'll play some on Saturday with z and see if that is indeed the case.
Last edited by avih, .
Perliminary update:
After more than an hour of online racing, no issues with Z. Now that I think of it, Z exhibits sort of start of this issue momentarily, like -0.1s-0.2s "hangs" (happened twice or so, but barely noticable), and it feels to me as if on z10/13 it would have developed to this slideshow/HD-noise thing, but it just recovers immidiatly and everything continues as normal. As much as I think of it, I can not assosiate it with a specific event (like user joining/disconnecting/lagging/etc). To me it just seems to happen rather randomly.
I'll try to race some more during the weekend and update further, but taking all info into account, I'd say that Z doesn't have this issue, or at least not in a disturbing way.
BTW, when I change Full screen/window (both Z and Z10/13), the frame rate drops to a crawl for few seconds and then recovers (tried on a replay but iirc it also happens online), occasionally with sound skips. It feels to me very similar to the effect I'm talking about in Z10/13, but it doesn't have the HD "grinding" noise. If you need I can post a short video taken with a cell phone of how it looks/sounds (the win/fullscn thing), it looks identical to the issue I'm talking about which happens during racing. It won't be today though.
When I try to think of reasons for this issue, garbage collection is something that might cause such apparently random resource usage. Though I used few graphics drivers, so I don't think it's it. Maybe GC within LFS?
After more than an hour of online racing, no issues with Z. Now that I think of it, Z exhibits sort of start of this issue momentarily, like -0.1s-0.2s "hangs" (happened twice or so, but barely noticable), and it feels to me as if on z10/13 it would have developed to this slideshow/HD-noise thing, but it just recovers immidiatly and everything continues as normal. As much as I think of it, I can not assosiate it with a specific event (like user joining/disconnecting/lagging/etc). To me it just seems to happen rather randomly.
I'll try to race some more during the weekend and update further, but taking all info into account, I'd say that Z doesn't have this issue, or at least not in a disturbing way.
BTW, when I change Full screen/window (both Z and Z10/13), the frame rate drops to a crawl for few seconds and then recovers (tried on a replay but iirc it also happens online), occasionally with sound skips. It feels to me very similar to the effect I'm talking about in Z10/13, but it doesn't have the HD "grinding" noise. If you need I can post a short video taken with a cell phone of how it looks/sounds (the win/fullscn thing), it looks identical to the issue I'm talking about which happens during racing. It won't be today though.
When I try to think of reasons for this issue, garbage collection is something that might cause such apparently random resource usage. Though I used few graphics drivers, so I don't think it's it. Maybe GC within LFS?
Last edited by avih, .
Yes, just found a local copy of the 132M Z, extracted only the exe, and just checked and it works. I'll update with the results after a while of playing online.
No. I'm willing to though. How should I get back to z? just download z and extract it to the LFS directory overwriting z13? (it's online compatible, right?)
[edit]
since It's been a while since I changed to z10, I don't recall the version I had prior to that. Would z3 be a reasonably major version before z10?
I was thinking of downloading from here: http://www.lfsforum.net/showthread.php?t=51349
or should I download the latest Y and let it update to z?
[/edit]
[edit2]
Tried downloading Y32 and I get "An invalid or no filename given." error for both 31 to 32 and Y to y32 versions.
So, where do I download Z from?
Last edited by avih, .
I have an issue with z13 which I think is similar or same as I had with z.
Here's what I wrote on the z thread:
and
However, the symptoms are much more severe now. I'm getting these hiccups rarely (not more than once an hour or two of playing) but they do happen badly. The hard disc gets working, FPS drops to 1fps or less (slideshow), sound stutters, and that's usually race over, hopefully without taking others with me. It looks like something non LFS is consuming (many!) CPU cycles, but there's nothing that I could think of installed or running.
I've disabled everything I could think of that might consume cpu/network: peer guiardian, antivurus which I've changed since z and it happens with both, windows update, etc, and it still happens once in a while.
I use resolution 1680x1050 or 1440x900, but this issue seem to be completely unrelated to the resolution...
Am I the only one experiencing this?
Here's what I wrote on the z thread:
and
However, the symptoms are much more severe now. I'm getting these hiccups rarely (not more than once an hour or two of playing) but they do happen badly. The hard disc gets working, FPS drops to 1fps or less (slideshow), sound stutters, and that's usually race over, hopefully without taking others with me. It looks like something non LFS is consuming (many!) CPU cycles, but there's nothing that I could think of installed or running.
I've disabled everything I could think of that might consume cpu/network: peer guiardian, antivurus which I've changed since z and it happens with both, windows update, etc, and it still happens once in a while.
I use resolution 1680x1050 or 1440x900, but this issue seem to be completely unrelated to the resolution...
Am I the only one experiencing this?
It's not a "normal" FPS drop which lower rez might cure (and doesn't, as I've mentioned). It's not just generally lower FPS but rather sort of "hiccups" in smoothness. The closest, even if to a lesser degree, symptom that I can think of is of vsync on with a slow PC (as mine is). It's more than that though, like framedrops, or "hangs" or bad time sync with players/server or well.. hiccups.. and it's accompanied with hard disc usage noise that wasn't there on Z (maybe mem leak that triggers more swap?). If I could give a better description, I would. Sometimes, describing symptoms can help the devs even if the description is not absolutely specific. I'm sure that if Scawen needs more details or thinks the description wasn't good enough, he could say so himself. Also, this kind of symptom wasn't described on any earlier post on this thread so I think It does fit here.
Last edited by avih, .
It's not a resolution issue, at least it wasn't on Z. I did change rez occasionally but I've been running Z on this max rez for a very long while to have a solid reference. Don't troll.
Been trying Z10 this evening on line. I disabled shadows because I didn't like the new look (low rez) and I prefer higher FPS. Vsync off, no FPS limit, same as with Z.
I get weird slow downs. Can't pinpoint it. It looks a bit as if vsync is on, even though it isn't, and also, I hear HD swapping/access much more than usual (usual is practically none in LFS Z, all other apps are closed, as usual). At least few times it got down to crawling (1-5 FPS) and it seems to me that it's either due to excessive CPU or HD usage or swap.
Overall.. something is making Z10 not smooth as Z is.
[edit]
My system:
Athlon XP 2500+, 1G RAM, 7600GS 512M (latest drivers), Win XP SP3, 1920x1200 and I get with Z 30-100 FPS, pretty smoothly.
I get weird slow downs. Can't pinpoint it. It looks a bit as if vsync is on, even though it isn't, and also, I hear HD swapping/access much more than usual (usual is practically none in LFS Z, all other apps are closed, as usual). At least few times it got down to crawling (1-5 FPS) and it seems to me that it's either due to excessive CPU or HD usage or swap.
Overall.. something is making Z10 not smooth as Z is.
[edit]
My system:
Athlon XP 2500+, 1G RAM, 7600GS 512M (latest drivers), Win XP SP3, 1920x1200 and I get with Z 30-100 FPS, pretty smoothly.
Last edited by avih, .
yah
AXP 2500+, no OC or other tweaks. An A7N8X-X ASUS bord and Kingmax 1G ram which wasn't state of the art even when I bought it few years ago. Seems to do the trick.I thought so too. But (i.e.) at the aston national backstraight I get consistent ~115 FPS on every lap.
Talking about breathing some fresh air to an older system
Ok, a short update.
I've replaced my aged FX5600 Ultra 128M with a 7600GS 512M AGP. There's a VERY noticable improvement. On the back of the grid (multiplayer) I can now get 30-40 FPS and in race I've seen more than 110 fps and if I limit it to 60 (my refresh rate), it almost never gets lower than 60, except on the grid and other crowded situations. That's with 1920x1200 resolution but with very little filtering.
So there you go. Athlon XP 2500+ can cope with LFS very well as long as it's accopmanied with a reasonably modern GPU.
Hope it helps
I've replaced my aged FX5600 Ultra 128M with a 7600GS 512M AGP. There's a VERY noticable improvement. On the back of the grid (multiplayer) I can now get 30-40 FPS and in race I've seen more than 110 fps and if I limit it to 60 (my refresh rate), it almost never gets lower than 60, except on the grid and other crowded situations. That's with 1920x1200 resolution but with very little filtering.
So there you go. Athlon XP 2500+ can cope with LFS very well as long as it's accopmanied with a reasonably modern GPU.
Hope it helps
If that's true, then anyone gaining noticeable FPS from this should turn off these filters IMHO, because it's the GPU that limits the frame rate on such cases.
Or bad initialization on LFS part...
Heck, if that trick really does work, LFS can automatically switch window/full-screen on start and improve the perforamnce..?
Thanks for your replies guys, it has become an interesting and informative discussion, at least for me
By what I understand from your data and descriptions, It seems that everyone's oppinion/conclusion about the lower framerate on the grid is that it's NOT more physics but rather the graphics subsystem. That includes CPU cycles for "preprocessing" and GPU cycles for rendering. It could be that the "preprocessing" also includes more implied physics CPU cycles, but I guess that only Scawen can answer that.
Which brings me to a conclusion that separating the graphics subsusyem into it's own thread/core could be relatively easy to achieve (relatively because.. well.. software usually is never easy to modify much) and might give a nice gain on multicore systems.
Is there a way to understand on a specific system whether the limit is the GPU or the CPU, per situation in the game? I was thinking of a tool that would record the following 3 values, say, once a sec:
1. current FPS
2. theoretical FPS with current CPU and blazingly fast GPU system
3. theoretical FPS with current GPU and blazingly fast CPU
I'm quite sure that near 100% of the graph, graph #1 will either equal graph #2 or equal graph #3. If graphs 2 and 3 are quite similar, then the system is well balanced.
If 1 equals 2 most of the time, then the bottleneck is definatly the CPU
If 1 equals 3 most of the time, then the bottleneck is definatly the GPU system
If 1 equals 2 most of the time but on the grid it equals 3, then a better GPU system will benefit mostly the grid scenario and so on...
Should be an interesting tool
btw, now that I think of it, is there a tool that can show a graph similar to 1/2/3 but that displays poligon count? if excluding FSAA, AF etc, it might give a pretty nice info for that kind of benchmark...
By what I understand from your data and descriptions, It seems that everyone's oppinion/conclusion about the lower framerate on the grid is that it's NOT more physics but rather the graphics subsystem. That includes CPU cycles for "preprocessing" and GPU cycles for rendering. It could be that the "preprocessing" also includes more implied physics CPU cycles, but I guess that only Scawen can answer that.
Which brings me to a conclusion that separating the graphics subsusyem into it's own thread/core could be relatively easy to achieve (relatively because.. well.. software usually is never easy to modify much) and might give a nice gain on multicore systems.
Is there a way to understand on a specific system whether the limit is the GPU or the CPU, per situation in the game? I was thinking of a tool that would record the following 3 values, say, once a sec:
1. current FPS
2. theoretical FPS with current CPU and blazingly fast GPU system
3. theoretical FPS with current GPU and blazingly fast CPU
I'm quite sure that near 100% of the graph, graph #1 will either equal graph #2 or equal graph #3. If graphs 2 and 3 are quite similar, then the system is well balanced.
If 1 equals 2 most of the time, then the bottleneck is definatly the CPU
If 1 equals 3 most of the time, then the bottleneck is definatly the GPU system
If 1 equals 2 most of the time but on the grid it equals 3, then a better GPU system will benefit mostly the grid scenario and so on...
Should be an interesting tool
btw, now that I think of it, is there a tool that can show a graph similar to 1/2/3 but that displays poligon count? if excluding FSAA, AF etc, it might give a pretty nice info for that kind of benchmark...
The specs are as described on my 1st post. Game settings can be modified as needed to balance between visual quality/details/frame-rate.
I see your point KROM and I understand that on a fast and balanced system, the CPU, or rather, the inability of LFS to use multi-core, is the main bottleneck.
However, my system is few years old and I'm not sure how balanced it is. Other than LFS' slightly lower frame rate at the grid, it suits all my needs, including video processing etc. My CPU is more or less the maximum I can use on my system/board without completely replacing it (to SATA drives, to PCI-E GFX card, etc). I can still put a better GFX card though (i.e. a used 7600GS AGP), for few $$. Do you think it'll improve LFS performance on my system or would it be neglectable?
thanks again.
However, my system is few years old and I'm not sure how balanced it is. Other than LFS' slightly lower frame rate at the grid, it suits all my needs, including video processing etc. My CPU is more or less the maximum I can use on my system/board without completely replacing it (to SATA drives, to PCI-E GFX card, etc). I can still put a better GFX card though (i.e. a used 7600GS AGP), for few $$. Do you think it'll improve LFS performance on my system or would it be neglectable?
thanks again.
Thanks for your reply
Not sure how you concluded it's the CPU though..
Let me rephrase my question:
It's a known fact that LFS has a lower framerate when starting the race than in most in-race cases (I'd relate to online races without getting into the CPU of the AI cars for our discussion).
The question is: What changes on the grid compared to in-race?
- Is it much more CPU to handle more physics of cars which are near?
- Is it much more GFX processing due to higher polygon count of the many near cars or of the pits area? If it's the GFX processing, how much is the CPU involved in that increased processing, if at all?
If anyone got data as for the approximated percentage load increase of the CPU and of the GFX card on a full grid compared to a "normal" in-race scenario, I'd like to hear your comments.
Thanks again.
Last edited by avih, .
FGED GREDG RDFGDR GSFDG