Quote from N I K I :

You need to take all weight from XRR to be fast as FZR on AS nat. But when you look at other tracks 30kg should be perfect. I believe XRR can be faster then FZR on some tracks, but it's natural that XRR is slower by lap times , but that doesnt means that it's slower in enduro race

You might be right but no-one has yet produced any evidence to show this is true. Even if it is true, why shouldn't the XRR have a slight advantage during endurance racing? There is no reason why the FZR should be the fastest car in all situations is there?

Quote from JohnPenn :

Hi Maxim
Great report thanks
Its my car of choice because I love the the way It corners and handles, what I did notice after a break from Conedodgers Is the big increase of FXR drivers since patch X was released and how I can't seem to catch them.. or anyone.

They could add 50kg and I would still drive it, Try the XRR next time, you may like it

Thanks. Check my stats though and you will see I am a one car man (and it doesn't have an F in its name)

Quote from jayhawk :

I would use the XRR more, if it was not for the horrendous turbo lag! You pretty much have to left foot brake every slow corner, and that can be difficult to do sitting in office chair.

I know what you mean but I also find this is one of the things that make it a challenge and keep us hardcore XRR drivers behind the wheel. Removing this completely would just make it closer to being an FZR in a different skin.

At the moment if all cars had the same physical appearance an experienced GTR driver could tell what car they were in after just one lap and that is how it should be. What would be nice is giving all cars a chance relative to their difficulty. The FXR FZR balance is close (on AS3 at least) as both cars make up about an equal number on the grid. The FZR is more difficult to master but rewards the effort with a faster possible lap time. The XRR still needs a tweak to get it to make a regular appearance as the extra effort to tame it and its short-comings just put you level with an FXR.

Maxim (completely biased XRR driver but trying to be impartial)

It has been a few weeks since the online ballast on the XRR was reduced by 10kg and as srdsprinter asked I thought I would look at what affect this has had using the stats collected on Conedodgers’ CD2 server.
I am afraid I didn’t keep a record of total usage last time but for the record the current number of drivers (many will have drive more than one) recording a time under 1:44.00 is:
No. %
FZR 476 41%
FXR 547 47%
XRR 137 12%

This shows that the difference between the best car (the FZR) and the worst (FXR?) is not enough to put people off driving the FXR. From driving the server my feeling is that half the field is usually using the FXR.

Looking at the change to the fastest laps gives the result:
Old Now Diff
FZR 100.95 100.90 0.05
FXR 101.70 101.68 0.02
XRR 101.75 101.28 0.47

I terms of the fastest lap nothing much has changed for the 2 F cars but would suggest that the XRR has been given a big boost by the loss of the 10kg. I think this is a bit misleading however as the top XRR is an amazing lap that leaves even the second fastest even way behind. Using 2nd fastest would show 0.00, 0.08 and 0.26.

Looking at the changes to the averages for the top 15 gives the results:
Old Now Diff
FZR 101.38 101.18 0.20
FXR 102.26 101.97 0.29
XRR 102.40 101.82 0.58
This shows all cars improving even though for two of them there has been no change to their ballast levels. This is presumably because setups are getting refined and people used to the ballast.

As a result of the change the XRR has gone from just behind to just ahead of the FXR. The usage level of the XRR has in my opinion gone up very slightly but is still a very rare sight on the server. I have done some graphs this time is it is easier to see the relationship of the various stats that way.

As it stands the XRR has a very slight advantage over the FXR and shares the big deficit to the FZR. The XRR is still a difficult car to drive compared to the FXR and any advantage is may enjoy is not enough to warrant the learning curve for new drivers who may as well use their time to learn the FZR as the rewards are far greater.
My own opinion is that 10kg is about 0.2-0.3 seconds and that we need to try taking another 10kg off the ballast to move it closer to halfway between the two F cars and see if the useage level can be moved closer to 25%. Actually that’s not true as I think it needs 0kg ballast to put it even with the FZR but I know that isn’t going to happen.

I am not going to repeat all the stuff about this being one circuit one length etc as that got discussed last time and no-one could produce stats from anywhere else to back up their arguments and give my analysis some balance which I would like to be able to do.

Maybe we should look at this from another angle. At the moment the FZR enjoys an advantage over the FXR. This is not enough though to stop nearly half the field using the FXR due to its driving characteristics so that gap is probably about right although both sides would probably argue for a change. What reduction in the XRR ballast would it require to see it appearing in every race where say 8 people gather?

Maxim

Sorry to drag up old ground but could an iVibe packet be created? I am not sure what that would mean exactly but I understand they have made a specific request in detail previously.

I sort of get the impression form the various threads about ivibe support that 90% of what they need is available by mix and match from the other packets and they are being a bit lazy not even creating 90% support. However, while new packet are being created could this be considered? Maybe then add the surface under each wheel info which seems to be the one thing they can't get. I am not sure if they use that info for canned effects which goes against how LFS deals with things but letting someone else respond that way wouldn't be considered selling out would it?

Maxim

Since Patch X, admins can see other admins using the connection list and pressing shift-control as admins appear in a different colour.

Some people drive differently when they know there is an admin on the server. It was my understanding that it was felt that there shouldn't be an easy way for non-admins to know that there isn't an admin present so they don't abuse that fact. By not knowing they may find to their cost that one is present should they start doing things they shouldn't.

If I understand it correctly, the new conedodgers server system, when ready, will enable people to be given limited admin rights on the fly like you suggest, but I don't know how it achieves it.

Maxim

Ugly, unforgiving and laggy. Those are the reasons to use it.

Shame it is still a bit slow though.

Maxim

It is an accepted fact that the FZR uses more fuel. What has yet to be shown is that this does anything to reduce the current speed advantage this car has to the point it is of a similar speed to the other two.

The idea is the bring the car's performance closer together. Before the ballast a good driver in an FZR could beat an alien in one of the others. One of the cars will be the fastest and I don't think anyone believes this should no longer be the FZR, but that the size of this advantage needs to be pulled back so a mix of cars gets used.

Prior to the ballast you could regularly see a full field of FZR's on AS3 due to the size of the advantage this car enjoyed. Now we are seeing a good mix of FXR's and FZR's which suggests that the balance (around 1 second) is about right for these two cars on AS3. The ballast in the XRR has been reduced by 10kg and in a few weeks 1000's of more laps will have been done on ConeDodger 2 server and the relative speeds can be compared again to see where the XRR now slots in the mix.

I would really like to look at stats from other circuits as I am trying to compare using real results not just feelings. Unfortunately no-one has yet suggested a server where the number of laps being done is high enough and they are being recorded by lapper so for now I can only compare on AS3.

Endurance racing is a small % of the type of racing done in LFS. As such it is done in leagues and in those cases if the organiser thinks there is an imbalance then they can use the voluntary ballast to even things out. It would be wrong to balance the cars over 4 hour races if this meant they were out of balance over the distances used for the vast majority of LFS races.

Only Scawen knows how close he would like the cars to end up. If the ballast could be added to hotlaps then he could have used those to generate stats. However the ballast only affects online laps so I believe he needs the community to present real stats for him to consider. Unfortunately this is only AS3 so far which is far from ideal but apparently the best we can do.

Maxim

Quote from Gintoki :

What happens When we reach Patch Z S2, will it go to S3?

If more patches are required you usually add extra digits like going from 9 to 10 so the series continues

......X, Y, Z, AA, AB, AC......

Maxim

The idea behind the licence system is not to create elitist servers. If there was a way to filter drivers based on attitude I am sure we would use it as the licence is just an idea to offer cleaner racing. The time of 1:44.4 is just a guess at where once you have passed it you can probably lap fairly consistently. If you start causing trouble on CD2 (or CD4) it is easier for the admins as it is likely to be caused by attitude as you will have proved a reasonable level of driving for this combo.

Remember the current system is just a temporary one and will be replaced once the intended system has been rewritten for patch X. Don't worry too much about the current system or how it can be improved because it won't be with us forever.

Maxim

Quote from Scawen :

OK, I've reduced XRR handicap by 10 kg so it's just 30 kg handicap now. That should make it slightly more desirable I guess.

This decision was based on what I read here, considering the evidence presented and not wanting any of them to be the car that no-one uses.

Thanks Scawen. I am not sure 10kg will make much difference to putting XRR's on the grid but I will recheck how things are panning out (on AS3 at least) in a couple of weeks and make a new comparison.

Maxim

Quote from DeKo :

but the XRR has the advantage of being able to do much longer runs than either of the other 2, which balances it out, and the FXR is much easier to drive. Its pretty spot on, the FZR should be the fastest, but hard to drive and has to take more stops than the others.

In the screen shots of the results of the 4 hour race all 3 top teams stopped 5 times so does that really support any real advantage for the "legendary" fuel and tyre efficiency of the XRR? The stats from the AD3 server I quoted are after 100's of individuals have raced there over a period of time. The results from teh endurance race are interesting but we can't rule out that the best team won rather than the car having any advantage.

Are there no other lapper equiped servers out there running GTRs so we can gather real data from other circuits?

Maxim

Quote from anttt69 :

Hmmm I am interested in doing this. What mobos will work for an nvidia config?
Also Maxim & Pudel could you post a pic of your setups running LFS, I want to see them .
Thanks

Not quite what I have now as I have replaced the top of the "table" with one capable of taking all 3 monitors and repalced the middle monitor. I now have 3x 19" screens. In the picture the left screen is a 17" screen although I didn't really notice when racing. Using SoftTHG I can now more my original 22" widescreen back into the mix but I haven't gotten around to it yet. The desk is also has a bit more mess on it

My motherboard is actually an AMD crossfire board running nVidia cards. You should be able to use any SLI or crossfire board ok. I never managed to get an ATI and nVidia card running together though but not sure if that is because it isn't possible or it was too much for my knowledge.

Maxim

Last edited by MaximUK, Fri, 22 Jun 2007 12:54.

It might be I remembered the order wrong but I think the important part was not trying to install both cards at the same time. I remember the frustration of endless blue screens though.

All 3 of my monitors are running using DVI.

Maxim

Quote from PudelHH :

thank you both....will check the dual mode and also maxim's hint.

@maxim....how did you "stopped the 8800 being used"?

When to are in the screen settings (after right clicking your desktop) you can ask for certain screens to form part of the desktop. You can also chose which card is the primary card. I just made the desktop use both cards then chose to make the primary card tthe other graphics card, then asked it not to stretch the desktop onto the screens driven by the 8800, leaving me with one screen driven by the 2nd graphics card. At that point my desktop is 640x480 in 4 bit colour but only lasts that long while you reboot.

Maxim

Maybe the make-up of the grid is the better judge of when the cars are balanced?

The fact that on AS3 the FXR and FZR are now about equal numbers on the grid might suggest the balance is correct between these two. To make the XRR compete with the FZR would require the removal of all of it's ballast although even then it would still be 0.5 seconds shy. I don't think even then you will see 1/3 of the grid in XRRs but it would help to get it back on the circuit.

What about it Scawen? Remove all the XRR ballast and see what it does to the grids?

And +1 to all cars getting stuck in the gravel although this will need to wait for a patch with physics changes.

Maxim

I had a lot of trouble getting 2 nVidia cards working together. The drivers I finally got to work with both were 57.?? as anything later caused blue screens.

What I did was only have my desktop on the 8800 one output. I then booted and installed drivers for the 8800. When it asked to install the other drivers I cancelled it. I then stretched the screen onto the other card which was running at 640x480 4 bit colour. I then made this the primary screen and stopped the 8800 being used. I then rebooted and installed the drivers for the other card. I got lots of blue screens which I believe is probably becasue the two cards share some of the files and not all drivers support both cards but that is just a theory.

You will need to read your motherboard book to find out about the single dual option. You need both slots runnign at x8 rather than one at x16 and the other at x1. I don't know which way around this is though on your motherboard.

If someone else knows a better way I would like to try later drivers if it is possible.

Maxim

Quote from Slopi :

Not sure if you participated in the test patches, but this was discussed quite a bit during testing. The final result seemed to have plenty of people pleased. I would suggest that what you're looking at is a very small sample of drivers on a single circuit, and that's it.

That's why I said it was one circuit and we need to analyse more.

I did run the test patches but the fields were very small prior to X as most seemed to stay with W. I have tried to present figures to support my views. The discussions during the test patches were unsupported feelings. No harm in looking at it using real figures. Like I said maybe nothing will change.

Maxim

Patch X has now been out a while and I would think some meaningful statistics can now be drawn regarding the car balancing being attempted online. I thought I would try to start the ball rolling by presenting the results of the GTRs on Conedodgers 2 server. This server runs lapper and therefore keeps a record of the best lap times recorded by each racer.

This is one circuit and represents only the fastest laps driven on a 10 lap race. We need to collect stats for other circuits and possibly distances but we have to start somewhere. If you know of other circuits running GTRs and lapper then highlight them here even if you don't want to present the stats.

Ok, I have taken the top 16 times for the FZR and FXR as these cars are by far the most popular to have been used on the server. I have taken the top 8 times for the XRR on the basis that if it was driven by twice as many people the 8th position time would drop to 15th. This improves the XRR average by 0.35 seconds but is probably more representative.

The fastest recorded times on the server are:
FZR 1:40.95
FXR 1:41.70
XRR 1:41.75

The average times for the top 15 times (or 8 in the case of the XRR) are
FZR 1:41.38
FXR 1:42.26
XRR 1:42.40

For this one circuit the results show that the FZR is still much too fast compared to the others. If the aim is to balance the field then the FZR needs more ballast and the ballast in the XRR is too much. Personally I feel the XRR and FZR are generally harder cars to drive in that they require better throttle control and put a wheel on the grass and you are in trouble. The FXR on the other hand lets you get away with clipping the grass and seems to plough through loose tyres with little trouble which would end you race in the other two. For this reason I would rather see their ultimate speed slightly greater than that of the FXR as the FXR driver is likely to be able to drive closer to its limit for the average driver over the course of a race.

My personal experience is that the FZR has lost about a second as a result of the ballast and the XRR has lost about 0.5 seconds. I don't know what the current ballast levels are (80kg & 40kg?). Doubling the FZR ballast would reduce it to where the FXR is now. For the reasons above this would seem too much as everyone would just drive the FXR so maybe a 50% more is what is required to just close them up? The XRR seems to need to loose it's ballast, or at least most of it to return its slight advantage over the FXR and get it to a similar speed to the FZR.


This is one circuit though so we need more statistics from other layouts to make an informed judgement. There was some talk about the endurance advantages of the various cars. To me endurance racing is a tiny part of how these cars are raced and endurance racing is much more organised. Given it is organised then the voluntary balancing can be employed by the organisers if they feel there really is an imbalance over longer distances. In my opinion, the cars should be balanced over the most commonly raced distances and this seems to be 5 - 10 laps.

The current useage of the cars shows that even with the imbalance, people drive the FZR because it is clearly the fastest and they drive the FXR because it is the easiest to drive near the limit. The XRR just isn't used that much because it has the handling troubles of the FZR but not the speed advantage. If nothing else is changed it seems to me the XRR ballast needs to be reduced or removed so that the mix in the field is more even. I will declare my bias here as I am traditionally and XRR driver on the verge of jumping ship to one of the others.

Lets get some other circuits analysed so the guesswork can be removed from the current ballast levels. It might be that AS3 just favours the FZR and the current levels will be seen to be correct across the full range of circuits. We won't know until we start checking.

Maxim

A bit off topic I know but does anyone have any idea what the best beginner's RC model plane is?

And as a separate issue, does anyone know a good beginner's RC model plane?

Maxim

Quote from Becky Rose :

Nobody wants to drive a car designed by a 7 year old with a thick crayon, even if it is a real car.

Here kids, stick some of these boxes together and Daddy will take it to work add some wheels and call it a volvo.

Maxim

Quote from Burnzoire :

well it all adds up! Certainly though, if you play the demo for a year and still say you can't afford to buy it, I call BS.

Buying S2 is the only way of proving to everyone that you're serious about LFS. If a demo user can't save up enough for S2 or if they would rather spend money on other things, then it shows they just don't care enough about LFS. So why should their voices be heard if they obviously don't care enough in the first place?

Maybe Scawen and co just misjudged how much to give away for free? Maybe once the improved AI have been completed an automatic online ban after 3 months of demo use should be introduced? I can imagine the endless list a whining posts already:haha:

Maxim

Quote from DEVIL 007 :

PCI has theoretical bandwidth just 133MB/s !
In fact it will be like 100-110MB/s.Thats really not so much.

When you have 2 cards in PCI-E slot they can output up to 4 screens if SLI is not enabled.
PCI-E bus has more then bandwidth for this purpose.

Also maybe the CPU might be really limiting for 3 screens as I guess with softTH its the CPU who dispatch the graphical data for each of the card.I really dont know exactly how this program works.Maybe you could dd a test just to overclock the CPU by 200Mhz and see if there is any FPS difference.Might be interesting to test.

You're right, that would be an easy way to prove it one way or another. I am going away for a few days but I will give it a go when I get back. It did seem strange to me that the PCIe bus wouldn't be enough but it still seemed a more likely cause than CPU limiting it. I will let you know what happens. In the set-up you mentioned each card would output it own data so wouldn't need to transfer data to the other card. In SoftTHG one card does the work then transfers it to the other card over the PCIe bus to be output by that other card. Tests have shown AGP isn't enough so the transfer rate must be quite high. I guess kegetys would know the answer but hopefully he is spending any spare time he has redoing pitspotter for patch X rather than reading this as I really miss that routine in a 32 car race.

Maxim

Maxim

Quote from DEVIL 007 :

I am curious if 2 graphic card in 2x PCI-E slots could help to increase a performance as the old PCI slot is very limited with the bandwith.

If you dont enable SLI it should work I think but I have really no idea.Also the AMD 4000+ might be another limiting factor.For example E6600 should increase the performance.

Both cards are in PCIe x8. It is a crossfire board so no option for SLI but softTGH doesn't like SLI (or crossfire) as I understand it. If I use one screen at 1280x1024 then I get well over 200fps so I don't think CPU load is a limiting factor but I don't really know so you might be right. I am surprised that PCIe bus speed would be a factor as it is only one screens data being transferred so it might be driver issues.

Maxim

I thought I would post about my upgrade experiences. My games pc is based on an AMD 64-4000 with 2Gig of RAM.

With an ATI1800XT and 1300 combo I could run LFS at 3 x 1280x1024. I needed to turn down a lot of the settings and max out the dynamic LOD settings to achieve 19fps in T1 and 30-35fps when alone.

I have switched this out and replaced it with an nVidia 8800GTS + ATI1300 combo. This worked fine as a desktop but softTHG couldn't create the ATI screen. There also seems to be issues with the ATI 3D drivers working.

I have now replaced the 1300 with an nVidia 7100GS. I had lots of trouble getting a set of drivers that worked with both without blue screening during bootup but did eventually get it working. Getting the ATI1800+1300 working together was easy.

I maxed everything out ingame and got 29fps! I dropped the settings back to the ones I used with the ATI combo and it made no difference and it remained at around 29fps. Adding 16xAF and 16xAA also made no difference to the fps although when combined with maxed out game settings did improve the graphics a lot and no more square wheels. What I did find was running the side screens at 1024x768 did increase the fps (which it didn't on the ATI combo). Dropping it to 800x600 increased the fps to 55fps with about 48fps in T1 which is much better than I was putting up with. Using 800x600 with 16AF and 16AA still looks great and you don't usually look directly at the side screens.

These results suggest that it is the PCI-e bus speed limiting the fps. It also suggests an 8800GTS is overkill as the fps don't change with maxed out settings. It seems from my limited testing that a graphics card in the ATI1800XT class is about the minimum you can use to achieve 3 x 1280x1024. With that you will need to compromise your graphics settings to enable you to get through T1 safely at the back of a grid. Once you get to the next level with ATI1950's(?) and equivalent then it looks like you should be fine. The drivers I had to use are only 57's so it might be that I could get better performance with newer drivers?

I went for the 640Mb version of the 8800GTS. Memory useage with 20 AI drivers at 3x1280x1024x32bit is 285Mb so I would expect that with the new bigger fields with patch X the 320Mb version would run out of memory. If you are willing to run at 16bit colour though the 320Mb version should be ok and is much cheaper.

Maxim

Quote from Bob Smith :

Dell P1130s are 21" CRTs. I think they can handle that resolution and refresh rate.

Oops :ashamed:

Do your monitors support 85Hz?
I would start by adding the extra mode 1280x1024x60 in the config file.

You need to post your config file contents as that is probably where the problem lies.

Stick with it as it is worth it.

Maxim