Well, I do know what page faults are, but in case someone is a real tech expert, here's the scoop of my system: A) Pretty close to state of the art, reflected in the LatencyMon excellent results (except for hard page faults) B) Initially I had both a large number of hard page faults and at least some with excess latency--unfortunately LatencyMon doesn't let you distinguish the ones that are the "big" problems as in slowest C) Once I got rid of the stupid MSI Update, which I would have already if had a simple "do not run with Windows launch" type of option, the number of hard page faults dropped to the good range. MSI update should have to check for an update for 10 seconds maybe once a week. It was running constantly!!! D) At idle, LatencyMon now reported that I have a killer machine. However, as soon as any disk activity was required, I still had page fault issues with high latency. Next step, temporarily get of my Windows Pagefile. Made no difference. I reset it back to a smaller range, but may eventually just reset it back to "managed by Windows." E) Double check that all mainboard drivers and BIOS are latest. Yes. F) Await new rF2 build and/or new AMD drivers with fingers crossed. Or, await some genius advice here
Marc, if closing the MSI Live Update process made a difference according to LatencyMon, then I would press Ctrl-Alt-Del and check the running processes and shutdown/kill any that shouldn't be there. I think a good system without much running should have around 40 processes last time I checked. That's a bit over-generalizing things, but just trying to give you a rough idea. The bottom should tell you the total amount of processes running.
Said this quite a few times....... I use HyperOS,works a bit similar to true image. Which means I can separate out all my programs into separate installs. My installs are minimal and specific to the job at hand. Example photoshop won't ever interfere with rFactor because only rFactor is on that system,photoshop is on it's own system. Backing up and reinstalling a system is a simple drag and drop affair. Each system is in it's own partition. It's like having 10 computers,all the exact same spec. If 1 system crashes for whatever reason and becomes unusable it's a simple case of format,drag backed up system,drop into free 'space'. All done and dusted within minutes (dependant upon components of the computer). You can recover from ANY software crash/malfunction or infection. here's what it looks like.......
Because I also suffering with stutters, I did some testing with the suggest settings here. I used minimal graphic settings to detect wheater its GPU related or not. My Specs: CPU: AMD FX8320 @ 3.8GHz MB: Asus M5A97 Evo R2.0 2x4 GByte Ram @ 1600 GPU: Gigabyte HD 7850 2GByte OC @ 1050/1300 MHz Running processes after Win7 SP1 64bit startup: 38 (shutdown CCC.exe and MOM.EXE) rF2 graphic settings: - NO AA and multisampling, AF biliniar - lowest possible ingame graphic sets except HDR on and 12 visible cars - limit FPS to 60, vsync off (without limitation the FPS are over 100. So the hardware is capable of providing constant 60FPS) First I checked my System with the Latency monitor (screenshot - latency_mon.jpg): - it reports an DPC issue with a high latency of the USBPORT.sys (usb driver 1.1/2.0) All drivers (usb/chipset/gpu etc.) and bios are up to date. Win7 SP1 has already included a fix for high latency with USBPORT.sys, but it doesn't seem to help here. I haven't found a solution for this on the www so far. HW monitoring with MSI Afterburner: I tested with the Megane Mod 1.44 and Nordschleife 2.01 with 29 AI in a offline race. As you can see on screenshot - MSI_hw1/2.jpg, the GPU usage is around 50% and GPU memory usage around 950MB. When the stutters appear, the framerate (ms) is rasing very high from normally constant 16ms to values of 100ms and even more. When this happens I got at the ingame CPU usage monitor green dots on the right side of the histogramm (see screenshot CPU_usage_rf2.jpg). I hope it is noticable...I also added a red circle. GPU fan's speed = 100%, temps with lower than 40°C are no problem at all. CPU usage is always one core very high (80-95%) and the others are > 50% or even only a few more than 0%. It is not always core 6 with is used most. Another test showed CPU1 with high level. The minimal power CPU setting in Win is set to 100%. With these lowest graphic settings rF2 is running relative smooth but it looks aweful. But the framerate limiter in rf2 player file to my monitor refresh rate is causing tearing. But with no limitation are the stutters even worse (like already mentioned, it doesn't matter weather the FPS are dropping or gaining). Bigger fps steps leading to stutters. Vsync option for rF2 make the stutters even worse. At least with this settings the stutters appear only at the race start and it is definitely running like it should be. After the race start the stutters are very rare. If I restart the race, even the stutters and the start disappear. So are the stutters caused by some ram/ssd memory loadings? But when I gain only the shadow and special fx from off to low, the stutters taking the fun. With medium settings the Nordschleife is undriveable even in practice with no visible AI. I know this system and especially the GPU is not enough anymore, but the abnormal CPU usage of rF2 and probably the worse AMD driver prevent a better result. But imo the difference to iracing, were I can use nearly full graphic settings with 4xAF/AA and it runs constantly over 100FPS even in a grid of more than 40 cars, showing that rF2 needs some improvements. But for sure ISI can't be blamed for the worse AMD driver.
I've been running ATI with Intel CPU's for a few years now and never had any stutterings. Just recently gone to an i7 from core2duo. Maybe I've been lucky or maybe it's because I've only ever had rFactor2 installed (with drivers of course) on a system? HD4870-HD6970-R9 290 Do you stutterers have your pagefile on the same partition? I've had mine on a different partition since using HyperOS. What about hybernation/hybrid sleep? They're long shots but that's the road we're down now.
The only thing that stutters is rF2. And it is very minor. If the page faults are happening consistently, they don't seem to affect other programs that you can notice. Turning pagefile off completely made no difference, so not sure moving it to another partition is of any use.
Attention !!! Since the current build this doesn't works for me anymore for whatever reason. It would result in reducing the FPS aprox to the half. I don't know wether it has to do with the problems the recently popped up because of the multithread issue the ISI is working on or something else, but anyway i like to let you know there is something off atm. I'll test this again when there will be the next update on this subject. Meanwhile i suggest not to use this hack ! Further i like to suggest AMD user to test the last beta driver of AMD 15.4.1 ( http://www.computerbase.de/downloads/treiber/grafikkarten/amd-catalyst-komplettpaket/ ) , the seems to perform very good, at least on my machine where i have the subjective impression that the stutter is reduced with, the i had here and there. I'm really surprised how good this old machine is perfoming with triples ( no multiview ). What means, i have some more time till i need to upgrade anything.
