The Solution to TrackIR screen tears & jaggies!

[UK_Widowmaker 571]

Can we get this formalised?...and a proper Step by step guide done as a sticky?

[Tamper 9]

Can we get this formalised?...and a proper Step by step guide done as a sticky?

 

+1 - I'm sure I speak for everyone in saying the work being done is most appreciated. But I'm also sure that at least some of us are getting a bit lost.

 

Also, I had another question: If this is something caused/worsened by FPS being 'locked' or limited (by FRAPS or by other settings), then wouldn't I be able to duplicate that without necessarily having TIR?

 

The question was first raised about TIR, but it seems to me that, if this is a function of resolution/refresh/FPS, then anyone would be able to 'duplicate' the problem by setting the appropriate values to perhaps undesired settings.

 

Why would anyone do that? Well, I for one don't have the problems anymore, but that doesn't mean I wouldn't be interested in "optimizing" my settings. If there are 'good' settings, then there must be 'bad' settings as well. And if I'm following, a good part of this is finding a certain combination of aforementioned settings.

 

Can anyone clarify?

[Hellshade 110]

The TrackIR tearing is a totally separate issue from (for example) non-vsync screen tear or stutters. Simply having a faster rig in every way (for example) won't get rid of them as it might for screen stutters, as evidenced by the folks who have even modest setups yet had no TrackIR screen tears or jaggies. Yet some people with high end gear still have the issue. In other words, it's a gear independant issue, with the exception of course of the TrackIR gear itself. Solving screen stuttering problems, unfortunately, are not a guarantee that you wouldn't have this issue if you did add a TrackIR.

 

As for trying to duplicate the problem, the only way to do that is by turning your head left and right (yaw) quickly while using TrackIR in OFF over a nice patch of land. It then causes white jaggies or screen tears on the far right or left side of the screen only(depending on which way you are turning your head) where the land is visable. Settings and gear that work perfectly well for you without TrackIR are no guarantee that you wouldn't have the problem if you added one. Duplicating someones settings who is having the problem probably won't be much help to you either, since without the TrackIR there's no way to experience the effect and it's entirely possible that they are having no other display issues besides TrackIR screen tears. So basically, there's no way to duplicate the issue without the TrackIR gear because the effect only shows up when using it.

 

Best guess as to why it happens and how FPS or Refresh Rate can have an impact on it but not show up elsewhere is because it has to do with how TrackIR was programmed to interact with CFS3, which didn't happen natively. Pretty sure TrackIR support wasn't added to CFS3 until well after the game was out and the game engine itself may not lend itself well to the new technology. Old software can sometimes be very picky about how it works with new gizmos. That would explain why people can use TrackIR on more modern (and demanding) games and experience no jaggies at all with it there, but they immediately show up when running the CFS3 engine.

 

As for TrackIR itself, I couldn't possibly recommend it enough. In fact once I got used to using it (there is a learning curve), I can't imagine myself flying any flight sim without it. Just too constrictive without the ability to look almost anywhere at will. It adds so much to the immersion factor that going without it now would be almost like going to a silent, black and white movie at the theater today. Cute, but just not the same.

 

Hellshade

Edited May 30, 2010 by Hellshade

[Parky 8]

Also, I had another question: If this is something caused/worsened by FPS being 'locked' or limited (by FRAPS or by other settings), then wouldn't I be able to duplicate that without necessarily having TIR?

 

The question was first raised about TIR, but it seems to me that, if this is a function of resolution/refresh/FPS, then anyone would be able to 'duplicate' the problem by setting the appropriate values to perhaps undesired settings.

 

 

That makes sense to me, and you could certainly duplicate the situation where FPS are "locked" or limited (without even having TrackIR) just to see what the results are. I for one would be curious to see what kind of impact capturing a clip in FRAPS has on your system. That would kill two birds with one stone actually. It'll limit your in-game framerates to whatever capture rate you choose (one test at 30FPS and the second at 60FPS (fullscreen) would be ideal) and it also might give us an idea of how SSD's might improve performance in that scenario. I suspect the fact that there'll be a great deal of HD activity involved as the clips are being recorded to storage in real time will yield positive results in favour of the newer technology.

 

It's unfortunate you're unable to incorporate TrackIR into the final analysis, as on my system, the only time I run into tearing issues is with TrackIR in use. I think I mentioned earlier that if I kill TIR and use another method of panning, the problem goes away completely.....go figure. Surely TIR can't be all that resource intensive. I will swear on a stack of bibles though, turning off TIR is like losing your best friend. Seriously, the only time I can bring myself to fly without it is for testing purposes. Any chance somebody'll put one under your tree this year?? We'll keep our fingers crossed for you...

 

Cheers,

 

Parky

[RAF_Louvert 101]

.

 

(posted this in the other related thread and thought it best to drop it in here as well)

 

 

I discovered another tweak this AM in regards to the tearing and/or stuttering issues associated with TrackIR and OFF. Since CFS3 was not designed to support multi-core technology it runs as good, if not better, on a single core as it does on two or more. I remembered an article many moons ago that touched on this very issue and the solution at that time was to go into your program processes via Task Manager and set CFS3's affinity for only one of the CPU's, as this would cure the stutters and shakes numerous flyers were experiencing at that time. So, I decided to give this a try myself with my OFF/TIR set-up. I launched TIR and OFF, then brought up Task Manager and set program priorities to "high" for each, then set TIR affinity for "CPU 0" and OFF affinity for "CPU 1", (so that basically they each have their own CPU to work with). Lo and behold I got even smoother performance from my TrackIR. I encourage others to test this out and see if they gain an advantage in their situation as well. Good luck.

 

Cheers!

 

Lou

 

.

[Morris 2]

Thanks Lou.

 

m

[Von Paulus 8]

Since CFS3 was not designed to support multi-core technology it runs as good, if not better, on a single core as it does on two or more.

It's not complete accurate, Lou. CFS3, as any legacy application, wants more clock speed (Ghz) than several cores. But if for example you've single core that runs @ 3.0Ghz and you've multicore system that runs @ 2.4Ghz, there might be a chance that it will run better with the single core. But if the clock speed of the multicore is equal or even bigger, then it's almost sure that the multicore system will win. And I'm not counting with Intel Turbo Boost or other technologies that might help the multicore systems.

 

However your mileage will still may vary. A lot of factors contribute to this. Operative system, background applications, system tuning, are only some.

What you suggest, Lou, it might work with some.

[RAF_Louvert 101]

.

 

Thanks for the clarification VP, clock speed is for sure more important to CFS3 than multiple cores. I simply put this info out there as it may be yet another solution for some folks. In my case I am running a dual core E8400 3.0 which I have OC'd to 3.8, and that gives my system more than enough juice to push OFF quite well on a single CPU. So by assigning OFF to one of the CPU's and TIR to the other I am giving both all they want and more in terms of proprietary processing power, (how's that for an alliteration). This will likely not work for most, but it may work for some, and, as it's become blatantly clear in this discussion, there is no one magic bullet that will kill all the OFF/TIR ills.

 

Cheers!

 

Lou

 

.

[Tamper 9]

That makes sense to me, and you could certainly duplicate the situation where FPS are "locked" or limited (without even having TrackIR) just to see what the results are. I for one would be curious to see what kind of impact capturing a clip in FRAPS has on your system. That would kill two birds with one stone actually. It'll limit your in-game framerates to whatever capture rate you choose (one test at 30FPS and the second at 60FPS (fullscreen) would be ideal) and it also might give us an idea of how SSD's might improve performance in that scenario. I suspect the fact that there'll be a great deal of HD activity involved as the clips are being recorded to storage in real time will yield positive results in favour of the newer technology.

 

It's unfortunate you're unable to incorporate TrackIR into the final analysis, as on my system, the only time I run into tearing issues is with TrackIR in use. I think I mentioned earlier that if I kill TIR and use another method of panning, the problem goes away completely.....go figure. Surely TIR can't be all that resource intensive. I will swear on a stack of bibles though, turning off TIR is like losing your best friend. Seriously, the only time I can bring myself to fly without it is for testing purposes. Any chance somebody'll put one under your tree this year?? We'll keep our fingers crossed for you...

 

Cheers,

 

Parky

 

Parky, I genuinely appreciate your response. I wouldn't mind having TIR, and while I'm always on a budget, the money isn't what stops me thus far. If I can be honest, it's a concern about motion sickness. I have trouble with many FPS games...I wear these silly wrist band accupressure things, and it seems to help a lot (all in my head? Perhaps - but better there than all over the floor ) My wife has always marvelled that the flightsims don't have that effect - can't explain why. But my concern is that if I could TIR in the flightsim mix, well, cleaning barf from keyboard...yuck.

 

Anyway, I'm glad you can see the point in my question.

 

I did actually do testing as you suggested, here are my notes:

 

All tests daytime/bright blue sky, 3 Dr1 v 6 DH2. In teh following, F=FRAPS setting for locked framerate; M=OFF MaxFPS setting.

 

F30_M60 - No visual tearing, no stuttering*. FRAPS showed FPS being 'capped' @30FPS

 

F60_M60 - No visual tearing, no stuttering*. FRAPS showed FPS being 60 initially, but starting vid cap lowers them substantially, they settle around 40 during 'action' while recording.

 

F60_M30 - as above. FRAPS showed FPS being 'capped' @30FPS.

 

I can't see any difference in the changes.

 

* videos show a few 'jumpy' pans, this is actually me struggling to fly, not get killed, and still pan around trying to 'force' visual anamolies - my thumb slips off the hatswitch or lets it center briefly. I can post videos if you want, but I'd need to compress them somehow. Right now they're 'native' AVIs and go 0.5-1G apiece.

 

Note that when recording I am writing to my C: drive (which is not the SSDs). As memory serves, SSDs are not known for impressive write speeds; rather they excel at reads and much more so random access. In fact, my arrangement is designed to minimize writes to the SSDs at all, because writing is where the biggest concerns with SSDs are.

 

Also, note that I probably have an "unfair advantage" when it comes to read/write performance. Recall I use a PCIe-based, "hardware" RAID controller. So, while onboard RAID controllers (esp. ICH7 controllers, if I recall) can experience bottlenecks, especially when moving data between multiple drives on the same controller, I believe hardware controllers are usually the choice to maximize 'throughput' across RAID arrays with multiple volumes.

 

Another point of interest: My "sliders" are set conservatively at 5-3-3-4-3; which seemed to do OK on my hardware, with no problems and plenty of visual detail for me ('course, Im blind as a damn bat anyway). Coud be that raising the sliders would cause the triangles or the stuttering; I really haven't played with it since initially setting them up after my last install.

 

Also, I have the process priority for OFF (CFS3) set to 'high' using a tool called PRIO.

 

So, your opinion? Anything I should try next? I can post my settings as well, but let's say they're pretty basic; not using Nhancer or other bolt-ons at the moment, just NVCP. AF 8x, AA 16xMS, texture filtering = 'performance', triple buffering off, VSYNC = per app (which would be disabled, I'm sure).

 

I would think I could introduce problems somehow, because I (and others) have reported some of this triangle 'jaggie' business, without even using TIR.

 

Do you think I'm wasting my time testing without having TIR?

 

I wonder if I should just turn the sliders up? I have a fairly conservative rig, so I wouldn't want to try "too hard", because I'm sure it will break *lol* I just don't know if it's a 'fair' test at that point; turning the settings up to what just isn't possible on my hardware. Whaddya think?

 

Appreciate your interest.

 

@Lou "...as it's become blatantly clear in this discussion, there is no one magic bullet..."

 

I'm going to go with Lou on this one.

[Hellshade 110]

 

I did actually do testing as you suggested, here are my notes:

 

All tests daytime/bright blue sky, 3 Dr1 v 6 DH2. In teh following, F=FRAPS setting for locked framerate; M=OFF MaxFPS setting.

 

F30_M60 - No visual tearing, no stuttering*. FRAPS showed FPS being 'capped' @30FPS

 

F60_M60 - No visual tearing, no stuttering*. FRAPS showed FPS being 60 initially, but starting vid cap lowers them substantially, they settle around 40 during 'action' while recording.

 

F60_M30 - as above. FRAPS showed FPS being 'capped' @30FPS.

 

I can't see any difference in the changes.

Do you think I'm wasting my time testing without having TIR?

 

 

As I stated above bud, you can't duplicate the TrackIR white triangle / jaggies (whatever we all wish to call them) without the TrackIR. You can duplicate the settings, but the effect won't take place unless you are turning your head while using TrackIR. I don't have stuttering or screen tearing at all on my system normally, except sometimes when trying to record a video and my non-SSD drives are struggling to keep up with recording all the action on the screen. The only place I was having jaggies show up was with TrackIR fast head turns. For me and at least one other person, settings MaxFPS=60 (which happened to also be my screens refresh rate) got rid of about 95% of them. As long as my system is showing that I'm running at 60FPS they rarely appear and then only in minor bits. BirdDogICTs solution also worked for me and himself. Adjusting the TrackIR speed to lower numbers slows your head turn rate on the screen and that also helps reduce the jaggies. The faster you turn your head, the more pronouced the jaggies are. I'm hoping Over50 will try and see what happens. At any rate, without the device, I'm not sure how you could go about troubleshooting the issue.

 

Hellshade

[Von Paulus 8]

there is no one magic bullet that will kill all the OFF/TIR ills.

Right as always, Lou.

[Tamper 9]

Hellshade,

 

Parky and I were discussing his curiosity, as he explained in his post. We both acknowledged that it's unfortunate I can't test with TIR. But, at least based on his response to my questions, he follows what I'm thinking about, and I'm interested in what he has to say.

 

Interestingly enough, I did have the traingles (whatever we're calling it now), with no TIR, after upgrading to my current machine - but before the SSD/RAID arrangement. I believe there are others who reported graphics anomolies without using TIR, as well. So, that would lead me to believe it's not exclusive to TIR.

 

Also interesting, is that during my tests I didn't seem to suffer any loss of video performance - that is, zero stuttering, triangles or anything else, regardless of my FPS settings or when recording with FRAPS. Even when FRAPS was set to lock a frame rate my machine couldn't keep up with (but was clearly trying to, because I saw the numbers go up and down as recording stopped/started).

 

This certainly seems like my video hardware, CPU, and memory are keeping up well enough at about 40FPS, and the mass media storage subsystem is efficiently moving things along without causing video anomolies. As I pointed out, I'm not recording to the SSDs, just playing OFF from them. My two plain old Seagate 7200 Barracudas were doing the FRAPS recording. (The write performance is better on the platter-based drives than the SSDs anyway).

 

So why did it not impact the video performance? Well, two reasons, I suspect:

 

1. The SSDs are doing precisely what I paid them to do: Reading data faster than anything else can, to make sure there is minimal delay in getting display data where it needs to be - the screen.

 

2. The hardware RAID controller I use is doing what I paid it to do: It's apparently capable of moving all the recorded data to the Barracudas without loading the system unreasonably, at the same time it delivers what the SSDs put out to the rest of the system (to be displayed).

 

I'll tell you what: I'll post the videos if you want, and you tell me which one's which. I'll bet you can't tell, even though there is an obvious difference in the work being done for one as opposed to another.

 

I may just buy TrackIR to find out how much impact there really is. But I am reminded of Morris, who uses TIR5 and has SSDs...and, well...you know.

 

All the above is particularly interesting when one considers that your rig and mine are substantially different (32b OS v 64b, 4G 800 RAM o/c to 933 or so v. 6G RAM o/c to 1443, a GTX260/216 w896M v. a GTX 280 w1G, a C2D e8400 o/c to 3.7 v Ci7 930 o/c to 3.8).

 

In fact, if I could dare at a cost comparison here, I'd venture that it may be possible I've tied up less on my system (including the 2 SSDs and maybe even the hardware RAID controller) than you've tied up in your system. Gotta be close, I bet.

 

But anyway, I digress...as I said, I was just responding to Parky's interest in my questions.

Edited May 31, 2010 by Tamper

[Hellshade 110]

I'll tell you what: I'll post the videos if you want, and you tell me which one's which. I'll bet you can't tell, even though there is an obvious difference in the work being done for one as opposed to another.

 

I may just buy TrackIR to find out how much impact there really is. But I am reminded of Morris, who uses TIR5 and has SSDs...and, well...you know.

 

All the above is particularly interesting when one considers that your rig and mine are substantially different (32b OS v 64b, 4G 800 RAM o/c to 933 or so v. 6G RAM o/c to 1443, a GTX260/216 w896M v. a GTX 280 w1G, a C2D e8400 o/c to 3.7 v Ci7 930 o/c to 3.8).

 

In fact, if I could dare at a cost comparison here, I'd venture that it may be possible I've tied up less on my system (including the 2 SSDs and maybe even the hardware RAID controller) than you've tied up in your system. Gotta be close, I bet.

 

But anyway, I digress...as I said, I was just responding to Parky's interest in my questions.

 

and I'm reminded of Uncleal and others who have nowhere near either of our system specs nor SSD drives and they claim to have no TrackIR jaggies at all. So, please explain to me what "you know" about Morris's TrackIR and SSD drive? Sounds like mighty flawed logic to me. Also, I fixed my TrackIR jaggies without an SSD drive...so really...what is your point? Other than beating a dead horse that is.

 

Second, you have no idea what I paid for my rig, since you don't know when I built it. More importantly, when did "how much did you pay for your rig?" become a part of the solution to getting rid of TrackIR jaggies? It's not. It's just part of your endless quest to prove to everybody how superior you and your computer are to everyone else. But it never gets old, really. Ever. I look forward to your condescending tone in every post. It makes you a colorful new addition to an established, well rounded and very enjoyable forum.

 

And what exactly would you be proving by posting videos? You don't have a TrackIR to replicate the problem. I didn't have jaggies or screen tears at all before I got a TrackIR either. They only showed up when moving my head from side to side quickly while using the TrackIR. How would you replicate that without a TrackIR? What exactly would the video be proving? That you don't have jaggies without a TrackIR? Neither do a lot of people. That's not news, sir. And it wouldn't prove that you wouldn't have them if you had the TrackIR.

 

Hellshade

Edited May 31, 2010 by Hellshade

[RAF_Louvert 101]

.

 

 

 

feel the love

 

 

 

.

[Over50 0]

Just tried your settings and it worked beautifully. TrackIR at 0.8 worked great for me. Excellent solution BirdDogICT!

 

Hellshade

 

Well, I reset the MaxFPS to zero again and set my TIR to .8 speed and 25 smoothing. No Joy unfortunately. Then I tried various TIR speed settings trying to find the TIR/monitor refresh rate sweet spot .... but still no joy. With every thing tried so far, only slowing my head movement alleviates the tearing.

 

What's maddening is when just setting on the ground with engine running and with all sliders at 5 I can quick scan front to rear and back as fast as I can turn my head w/o any edge jaggies. Same for take off but holding it on the ground with the terrain rapidly changing. But once I start climbing out where the view range begins to increase the jaggies start showing up.

 

Sorry to rain on the parade guys but neither the MaxFPS or TIR speed setting below 1 circumstance has worked for me. Be it a quad core vs dual core or XP vs W7 thing, I'm still out in the cold on this issue. The only remaining fix that's worked for a couple folks is a SSD....which I've already said isn't worth the added hassle and cost just for one game.

[RAF_Louvert 101]

.

 

Over50, did you happen to try the idea I noted farther up in this thread today? Just curious.

 

Cheers!

 

Lou

 

.

[Tamper 9]

and I'm reminded of Uncleal and others who have nowhere near either of our system specs nor SSD drives and they claim to have no TrackIR jaggies at all. So, please explain to me what "you know" about Morris's TrackIR and SSD drive? Sounds like mighty flawed logic to me. Also, I fixed my TrackIR jaggies without an SSD drive...so really...what is your point? Other than beating a dead horse that is.

 

Second, you have no idea what I paid for my rig, since you don't know when I built it. More importantly, when did "how much did you pay for your rig?" become a part of the solution to getting rid of TrackIR jaggies? It's not. It's just part of your endless quest to prove to everybody how superior you and your computer are to everyone else. But it never gets old, really. Ever. I look forward to your condescending tone in every post. It makes you a colorful new addition to an established, well rounded and very enjoyable forum.

 

And what exactly would you be proving by posting videos? You don't have a TrackIR to replicate the problem. I didn't have jaggies or screen tears at all before I got a TrackIR either. They only showed up when moving my head from side to side quickly while using the TrackIR. How would you replicate that without a TrackIR? What exactly would the video be proving? That you don't have jaggies without a TrackIR? Neither do a lot of people. That's not news, sir. And it wouldn't prove that you wouldn't have them if you had the TrackIR.

 

Hellshade

 

"You know" was about Morris and the fact that he doesn't have the problem - and, according to his accounts, nothing worked until he got his SSDs. Take it for whatever you will.

 

I didn't say I knew how much you paid for your rig at all - only that I bet mine, with the SSDs, probably wasn't more than yours. There is a point, and that is that people shouldn't have to do all these high-end upgrades to get the most out of this sim - and obviously some people don't (Lou, for instance).

 

"How much you paid for your rig" became part of this discussion a while back - I believe it was you who promoted the idea that folks shouldn't have to (necessarily) spend money to fix the problem. And I've said all along that the fix for the issues I had didn't cost nearly what some had thrown at their machines. Cost vs. benefit has been a part of just about every problem-solving scenario I've ever had to do with.

 

All the rest of your comments about what I say and do here are...well, maybe more in your head than anywhere else. Seems you're a little touchy that your "solution" isn't the magic bullet. Look, I didn't tell you to go posting in three different threads you had the answer; you did that of your own accord.

 

All I've done is say what works for me. And yes, I'm recommending it to others...so? Why shouldn't I? Any law around here says you're the only one who can make suggestions?

 

The point behind the videos was to demonstrate that you couldn't see the difference between refresh rates, MaxFPS settings or recording FRAPs - which Parky had expressed curiosity about comparing the two (even without TIR - and, as I said, even though we both acknowledge I'm not testing w/TIR). He was interested in knowing whether the added load might make jaggies show up in my case, as was I. So I did some testing.

 

I understood his point, and I think he saw what I was getting at earlier: It shouldn't necessarily take TIR to duplicate the problem, and (again) some people who never even used TIR do have the problem. Along that same line, if there's a 'fix' that overcomes jaggies on a system that doesn't have TIR, then it just might illuminate a previously unnoticed relationship to the others who do have the problem with TIR. Sometimes to solve a problem, you look at what different cases have in common (like how they find serial killers...look for patterns, links, commonalities). In this case, the jaggies probably have something in common - just that no one's figured it out yet.

 

I understand the frustration, but there's also no law that says there's only one way to work on a problem. I'm not keeping you from doing whatever you want, and I'm sure the community welcomes your effort. At the same time, maybe you could try not getting bent out of shape just because I found something that works for me. Of course, I recommend it.

 

I am going about working the problem differently from you, that's all. I think a certain amount of 'diversity' is good in situations like this; makes sure all the bases are covered. Like I said the other day: Good work doesn't mind scrutiny.

[Over50 0]

.

 

Over50, did you happen to try the idea I noted farther up in this thread today? Just curious.

 

Cheers!

 

Lou

 

.

 

 

As a matter of fact, no. I'll give it a shot later this evening..

[Parky 8]

Tamper,

 

First of all, I'd like to say thanks a bunch for taking the time to run those FRAPS tests. I was almost hoping you'd report back that running at 30 FPS had induced the jaggy phenomenon regardless of the lack of TIR in the mix. Looking back at the Poll, I see only one respondant (eric) who claims to experienced video "distortion" even without the use of TIR, but unfortunately he doesn't expand on what he means by that. It may not even be jaggies/tearing at all. Hopefully if he reads this, he'll take the time to provide a more comprehensive description.

 

It would seem just from looking at the poll results that there is a very strong correlation between the tearing and the use of use of TIR. As indicated previously, I personally am unable to replicate the problem if I resort to other panning methods, even if I run my terrain and scenery sliders up to 5.

 

As far is it being a waste of time for you to have run these tests without TIR? Absolutely not! I was more interested in seeing wether or not we could induce the tearing phenomenon on your particular setup by doing the FRAPS trial. I was also interested in finding out just what kind of impact your hardware RAID controller and SSD's might have on your FRAPS performance, even though OFF is the only thing installed on the SSD's (a couple of questions about that stuff to follow). Your results have yielded valuable info as far as I'm concerned. It strengthens the hypothesis that TIR is indeed instrumentally responsible for what some of us are experiencing. The fact that some people with lower end hardware and TrackIR aren't seeing this condition at all is nothing short of downright baffling....to me anway! Aren't computers wonderful??

 

As a point of interest (or maybe it won't interest anyone at all, but I'm gonna' drone about it anyway), I've got four 1T WD's in RAID-10. These are controlled by onboard ICH10. I've got my paging file (and nothing else) for the RAID array on a 36Gig Raptor. I've experimented with page file size. I've even out of desperation, experimented with using no page file

at all in an attempt to force texture loads in and out of physical memory as opposed to the alternative. Yes, I know running without a paging area isn't recommended due to the potential for program or system instablility, but this was strictly an OFF experiment. I figured with 6 Gig of RAM I should be ok and oddly enough when I tried it, there seemed to be a slight improvement....although the jury is still out on that one. Could have been my overactive imagination....or too many Coors.

 

Allow me to ask you for a bit more info if it's not too much trouble. I'm curious to know if you have your paging file on the SSD's or if it's on the platter based drives. I recall reading somewhere about concerns over having the page file on the SSD's due to excessive wear? Any truth to that, or is it merely speculation? Also, I'm curious as to which storage solution upgrade on your system seemed to yield the most significant results in overall "smoothness" in OFF. Was the hardware RAID controller installed subsequent to the SSD's or was this part of the same project. If done seperately, which one of the two upgrades in your opinion proved to be the most valuable?

 

By the way, I'm sincerely sorry to hear about your motion sickness issues. I happen to know it's no laughing matter. A friend of mine can't even look at a 60hz LCD for any longer than about 30 minutes without getting excruciating headaches and sometimes getting violently ill as a result. He's gone back to good old cathode ray tube technology for obvious reasons. I suggested he might get away with a 120hz LCD, at which point I think he said something rather impolite about my ancestry. Hopefully your condition is nowhere near as severe as his and one day you'll be able to get away with using a TrackIR with no ill effects.

 

Thanks again for your efforts and info

 

I think I'm gonna' go try Louvert's affinity/priority trick and see what happens there.

 

Also......BirdDog.....messing with my TrackIR speed and smoothing has resulted in definite improvement. You're a star!!

 

 

Cheers,

 

Parky

[Over50 0]

As a matter of fact, no. I'll give it a shot later this evening..

 

Well, no soap. Tried various settings including real time priority as well assigning TIR and OFF separate cores plus adjusting TIR from .6 to over 1 speed ranges and smoothing from the default 10 up to 50.

The only other thought I've had has to do with caching, in other words, pre-rendering the image on the fly. For example, with FSX the first time I do a full left to right view sweep and back the image jerks but after that no more jerking. Then, too, I never do see the edge jaggies (with TIR) regardless which means it's probably unrelated to caching the image.

 

Congrats to those who've found a workaround with one of the suggestions here. Just not in the cards for me apparently.

Edited May 31, 2010 by Over50

[RAF_Louvert 101]

.

 

So Over50, your problem is the jerkiness you see when you first start to fly? Does it settle down after you have been in the air for a while? I have the same issue unless I go to an outside view of my kite and do a full pan around, up, and over before I roll. If I do that as part of my pre-flight it's smooth sailing for the rest of the mission. This of course goes right to your thought on caching.

 

Good Luck,

 

Lou

 

.

[Von Paulus 8]

Allow me to ask you for a bit more info if it's not too much trouble. I'm curious to know if you have your paging file on the SSD's or if it's on the platter based drives. I recall reading somewhere about concerns over having the page file on the SSD's due to excessive wear? Any truth to that, or is it merely speculation? Also, I'm curious as to which storage solution upgrade on your system seemed to yield the most significant results in overall "smoothness" in OFF. Was the hardware RAID controller installed subsequent to the SSD's or was this part of the same project. If done seperately, which one of the two upgrades in your opinion proved to be the most valuable?

Interesting. I too would like to know.

[Over50 0]

.

 

So Over50, your problem is the jerkiness you see when you first start to fly? Does it settle down after you have been in the air for a while? I have the same issue unless I go to an outside view of my kite and do a full pan around, up, and over before I roll. If I do that as part of my pre-flight it's smooth sailing for the rest of the mission. This of course goes right to your thought on caching.

 

Good Luck,

 

Lou

 

.

 

Yup, I get the jerkiness initially, and like you, it stops after a 360 degree outside view scan a couple of times while still setting on the ground, and from then on I'm good to go from roll out to airborne. Of note, the initial jerkiness occurs with the game default view panning and with TIR. This is what makes me think it's cache related, where the jerkiness ceases once the images have been cached. And of note and as I mentioned prior there is NO edge tearing either while I'm on the ground stopped or rolling out but not airborne yet. But once I get above the trees and the terrain view expands the jaggies begin and remain a constant to varying degrees depending on TIR view rotation speed.

 

I've seen some terraincache.xml files in the OFF directory but none of the values are intuitive enough to understand - but I've tried changing values anyway just to see what happens but none had any effect.

Edited June 1, 2010 by Over50

[Tamper 9]

I was almost hoping you'd report back that running at 30 FPS had induced the jaggy phenomenon regardless of the lack of TIR in the mix.

 

I know exactly what you're saying - although I can't say I "hoped" for it *lol*; in a perverse technical fashion, I had anticipated that outcome as well.

 

Looking back at the Poll, I see only one respondant (eric) who claims to experienced video "distortion" even without the use of TIR

 

Right - which I do follow you on. But here's the thing (and maybe this is just being overlooked) I, myself - after upgrading to my current setup (but before SSDs) did have the light blue/white "jaggie" triangles (as well as some stuttering). With no TIR. So, even though I couldn't truthfully answer the poll that way now, it's only because I no longer have that issue. If I were asked about 6 months ago or so, I woud've been right there with 'Eric". One bad thing about polls is you get to vote - but the 'answer' you're voting on is as worded by the poll's author.

 

Point is, for me, that some people who have experienced 'jaggies' w/o TIR. Hardly scientific, I know...but I've seen it myself. So it tells me that the 'jaggie' problem, while I agree it's aggravated by TIR, cannot be strictly caused by TIR. Make sense?

 

I was more interested in seeing wether or not we could induce the tearing phenomenon on your particular setup by doing the FRAPS trial.

 

Yup; I was following you, and was just as curious. Not as a substitute for the TIR issue, per se - but a possible different angle on the same set of symptoms.

 

I was also interested in finding out just what kind of impact your hardware RAID controller and SSD's might have on your FRAPS performance, even though OFF is the only thing installed on the SSD's (a couple of questions about that stuff to follow).

 

Also an interest of mine. Though I had used FRAPS before, I hadn't done any recording at all. And, it does appear that if I don't 'cap' frame rate at some level, my machine tops out at about 40-45 FPS. I seem to remember seeing 60+/- when not recording, so I guess all told you could say my particular machine takes about a 30% FPS hit when recording with FRAPs.

 

I was very encouraged, though, that there's no apparent loss of visual quality even with the recording 'hit', and I do not rely strictly on FPS to make that determination. FPS, according to what I've learned, can mislead one into thinking all is well in video land. But, it's very possible to have >35 FPS and still have 'lag' or 'stutters' that are quite noticeable. Even 120FPS isn't (of itself) a guarantee of smooth graphics display; it depends entirely on the 'rate of distribution' of the frames, rather than a strict measurement of frames-per-second.

 

Although I could be mistaken, I believe firmly that performance of the drive(s) has a lot to do with 'smoothness', and the lack of jaggies and stutter. But, as we know, this doesn't take TIR into account, and that appears to be *the* question.

 

I've got four 1T WD's in RAID-10. These are controlled by onboard ICH10.

 

Now I'm curious as to what kind of performance you get from that setup. It is my understanding that RAID10 is second only to RAID0 in performance, but that assumes same number of drives in both arrays. However, since RAID10 is "striped mirrors", I'd anticipate the performance of a RAID0 array with two drives (plus the redundancy afforded by the two other disks, which I don't have). Mind you, your platter-based drives are likely way faster than mine. And the ICH10 controllers are said to be way better than the ICH7s were.

 

What is your experience?

 

 

Allow me to ask you for a bit more info if it's not too much trouble. I'm curious to know if you have your paging file on the SSD's or if it's on the platter based drives. I recall reading somewhere about concerns over having the page file on the SSD's due to excessive wear? Any truth to that, or is it merely speculation? Also, I'm curious as to which storage solution upgrade on your system seemed to yield the most significant results in overall "smoothness" in OFF. Was the hardware RAID controller installed subsequent to the SSD's or was this part of the same project. If done seperately, which one of the two upgrades in your opinion proved to be the most valuable?

 

Long story, but I appreciate the question; it is perfectly relevant. I think what you read about paging files on SSDs is true for the most part, and not just wear. Obviously, wear is a bad thing - but I recall some analysis done where they determined that it would take 83 years of writing to a drive enough to fill it up to the point it caused a failure (Lord, don't ask me to quote where...) So, although wear is bad, I think a far more immediate issue is that of 'wear-leveling' - what the do in the firmware now to offset the wear problems.

 

Of course, now wear-leveling comes with its own problems: Slowing performance over time, also caused/worsened by normal reading/writing to an SSD. This is all stuff that traditional drives didn't have to worry about, and the industry is only just now becoming familiar with this stuff. As I understand it, these write/erase/move problems are an inherent part of the NAND components used in SSDs. But, most recently, the drive manufacturers are releasing utilities that allow a kind of bulk erase for these drives that restores the original performance. Problem is, it blasts a drive and cannot be done on a drive that's part of a RAID volume.

 

You already know that RAID0 carries the biggest performance, but also the biggest risk - having no redundancy. And, I'll bet you know what a PITA it can be trying to viably image a RAID0 boot volume.

 

(*phew*, forgive me for the length here, just trying to be thorough) All this leads to why my layout is what it is:

 

2xSeagate Barracudas RAID0, boot volume, C: Performance near twice a typical platter drive, and with the hardware controller it really is good - especially at writing, which SSDs don't excel at. Cheaper per unit storage by far than SSDs; a terabyte volume costs around $100 (just the drives).

 

2x OCZ vertex 30G SSDs makes a 60G volume that, while expensive per unit, is fast. And, since it's not my boot volume, I just get the volume built, install whatever, then image the drive as normal in Windows. I can break the RAID array at any time, take the two drives out for their 'periodic blasting', and it won't keep my machine from booting. And since I store the image of that D: drive on my TB C: drive, it's only 60G max used space to have on 'on-board' recovery for the SSD's when they come back home after being brainwashed :)

 

So, anyway - back to your question (sorry)...I have tried to do everything I could to eliminate (or at least minimize) writing to the SSD volume; no paging file, only basically 'installed' games (so far, only OFF). By doing this, my intent is to minimize the rate of performance decay and thereby reduce the frequency of 'brainwashing' the SSDs. These two guys are left alone, to do what they do best: read, and even better if it's random. Burst speeds are phenomenal

 

The hardware controller adds to it as well. Far as I can tell, both volumes are performing near the theoretical bandwidth of SATA 3Gb/s (2 drives in RAID0 would be 600MB/s - see charts to follow). Flash drive technology, like SSDs, has caught up with the bus rate - which has prompted the industry to respond with SATA rev3 (SATA 6Gb/s)

 

I'm going to post some results below, and you can see the (empirical) answer to your question about which upgrade did more...from single drive (both SSD and platter), on-board RAID (both), and hardware controller (both).

 

They almost seem linear...

 

From my perspective, I would have a difficult time saying which upgrade was the best, anecdotally, because I either didn't test it long enough to determine, or really preferred to do it another way to fit my particular needs. For example, having the SSDs as the boot volume (with OFF) and the other drives as D: to install everything else...well, it worked just fine, booted fast as hell, etc. But precisely due to the concerns you mentioned about writing to the SSDs, and the obvious maintenance issues imposed by having to recover a boot image, I tested each stage, documented it, and then felt compelled to move toward the arrangement I have now.

 

 

I think I'm gonna' go try Louvert's affinity/priority trick and see what happens there.

 

Definitely look forward to your observations here.

 

(pics of HD benchmarks to follow...)

[Tamper 9]

OK, HDtach test graphs: on teh left, the two SSds on the 3Ware 9650SE 'hardware' RAID controller; right, the Nvidia 780iFTW's 'onboard' controller. Note the difference in both burst transfer and sequential reads. (Also of note, the 0.1ms random access time - this usually goes around 12 ms for an average single platter-based drive; that's a 100x improvement).

 

Edited June 1, 2010 by Tamper