64hz fix?

Edit: Oh, drr nvm haha.

So far as I know, the 64 Hertz issue in Oblivion results in a slight unevenness or jerkiness to movement, turning, and everything else too, just easiest to see in turning from what I hear. dmxdex2020 describes "freeze for 2 seconds and fps will go from 75 my refresh rate down to 40s", which doesn't sound like the 64 Hertz issue to me. Possibly he was talking about 2 different problems somehow, but it didn't sound like it.

I get the same type of stutter, my fps is locked at 60 due to the refresh rate, every so often they will drop low then shoot back up, quite anoying, I'm hoping stutter remover will help with this at least a little.

I hope i can get it sorted i am one of these people who cannot play a game with such issues.

1.7 is the version to target, right?

I would say so

Yeah it is the latest that most people are appearing to use.

I'd avoid looking at 1.5 at all as I think many would prefer to forget that one.

For OSR I only ever supported 1.2.416 (hm... and I don't mention that in the documentation anywhere do I? should fix that)...

I've got Fallout3 v1.0 installed atm, never gone past the main menu. The fallout 1.7 patch is refusing to install without some sort of microsoft gaming live update or something, and that in turn insists on installing Microsoft .NET Framework 3.5 SP1 which it's busy downloading atm (so far as I can tell anyway, the progress bar is broken).
My CPU is busy trying to anolyze Fallout3 v1.0 meantime... in a few minutes I'll have at least a small clue what that one is like.

And I'm feeling a bit concerned about what to do in terms of a release format... FOSE 2 isn't out and doesn't seem to be imminent, despite the portions I need being finished, so a plugin dll is impossible. I could release modified FOSE dlls similar to what I've done briefly in the past, but that path is frought with hazards, and the hazards get worse the larger the user-base is. My impression is that the potential user-base for FSR is larger than anything I've done modified OBSE dlls for in the past... I think I need to talk to the FOSE devs again. Possibly I should do source-only releases of modified FOSE dlls, or persuade the FOSE team to issue me beta version numbers for modified FOSE dlls or something.

The standalone patch or a gfwl download? Also, good luck

I FINALLY got the 1.7 patch to work (lots of trouble with .NET Framework 3.5 SP1, a little trouble with the patch itself).

I talked to Ian Patt of OBSE/FOSE, he says that there will probably be some FOSE support for plugins next tuesday (the 8th of september). So there should be some form of FSR release then.

There may be alpha versions before then, but any such will probably be limited distribution only.

Great news, how are you finding FO3 otherwise?

Haven't gotten past the main menu. Well, I did reconfigure my keys, but that's it. I have managed to supposedly port most of OSR and add a fast-exit, but haven't managed to test anything yet. I've spent almost all of my time struggling with MSVC and FOSE/OBSE to get both OBSE and FOSE build targets working correctly at the same time.
Hopefully I'll get a chance to play in a day or so. Though I'll probably have to spend most of my early play time checking that everything ported over from OSR is functioning properly. The background art for the main menu is weird.

Shouldn't you play a little to gauge performance?
(This way you can have fun under the guise of research )

Hah, interesting enough, my stuttering went totally away after implementing Fellout! Maybe all the green/gray haze was slowing things done. Anyway, haven't stuttered since! Strange . . .

I've gotten a bit of play time in now. Not much aside from testing & debuging, but enough to get out of the starting dungeon, er, vault.

status of features:

FastExit: Working. Makes quitting faster.

64 Hertz fix: Working. In theory removes some microstuttering. I can't tell the difference.

FPS/time management: disabled atm - I could probably enable it with only a little work, but atm I don't have any method to do some of the things it likes to do, like distinguish between menus and game mode, or even safely tell whether or not the main menu has been reached yet

Critical Sections: I think the basic critical sections stuff is working, but the debuging & profiling mode for it is failing. More importantly, Bethesda no longer uses microsoft critical sections implementation for most of their stuff - they replaced most instances of critical sections with a custom lightweight equivalent. Bethesdas custom critical sections looks lower performance than microsfts - but higher fairness, so it will be slow, but it won't stutter much.
To really do fallout critical sections properly I'd have to write new code to handle the new thingies that I'm calling "LightCS"es (for Light-weight Critical Sections) in addition to existing code for regular critical sections (which fallout still has some of), and that would be... complicated... by the fact that they inlined some of the LightCS functions.

Heap: Currently heap replacement is disabled. Bethesda still uses their own custom heap manager, but they've changed the implementation - their new one suffers from contention slightly less and when it slows down it doesn't stutter, it just slows down, as it uses the new LightCSes for locking.
This means that a replacement heap on FO3 would probably increase speed, but have little effect on stuttering.

Also, FSR is currently not generating any console output, as I don't know how to safely do that on Fallout yet.

IIRC, the current FOSE build supports console logging. Any issues with calling Console_Print ( assuming FOSE uses the same identifier OBSE uses ) ?

I think that's "printc".

To the script compiler, yes. I was talking about FOSE's internals - Console_Print is the function that prints output to the console.

If I simply let OSR or FSR call Console_Print at will, either will crash.

In OSR I deal with this by first checking: (*(void**)0xB3A6FC) != NULL
In FSR I tryed to do the check like this: ConsoleManager::GetSingleton() != NULL
However, the check itself crashed. I might be able to deal with it by checking if FSR has been running for more than 30 seconds so far (IIRC, OSRs Console_Print crashes were due to trying to print before the console was initialized... which happens relatively early), but for now I'm just doing output to the log file only.

Current status:

Release Schedule: Actual release is still waiting on some kind of FOSE release to enable plugin support. There was some suggestion that that was going to happen monday, but it's now wednesday here so I think that's been delayed a little. If the delay turns in to a lot I'll start some kind of closed beta or something with a custom FOSE variant.

Features:
64 Hertz Fix appears to work.
Fast-Exit appears to work, but atm is not correctly saving configuration changes made in-game on exit, because kInterface_Console is required for that and not supported by my current FOSE version.
FPS management has been ported, and appears to work. It is currently unable to tell when the game is in menu mode, so it's effecting menu mode as well as game mode, but that's probably no big deal.
Critical Section stuff has been ported, and appears to work. Further, I've added support for the Fallout Critical Section substitutes that I'm calling Light Critical Sections. However, any form of profiling code for each seems to result in occasional instability for unknown reasons.
Heap replacement has not yet been ported, but that will get done sometime soon.

Performance Improvement:
Poor. The limiting factors for Fallout are mostly not the same the limiting factors for Oblivion, at least for the kinds of things that OSR concentrated on specifically. This may get better as knowledge of exactly what causes Fallouts performance problems improves.

Great work thanks for the update and I'm looking forward to your first release

I'm wondering exactly why a fast-exit thing is even needed. I've been in and out of the game more times than I care to count and anytime I've ever clicked to quit, it drops out about as fast as I could hope for with no modifications necessary. Probably about as fast as OB does with FastExit2.

With FastExit enabled it's noticably faster. But, yeah, it was already fast enough to not be a problem. I figured maybe some mods slowed it down? I've only been testing on vanilla so far.

I was up to about 130 mods at one point before I started using FWE and it quit out just fine. If the mods any difference I didn't notice it.

Initial alpha version:
ftp://71.115.192.77/sr_Fallout_Stutter_Remover.dll
No documentation or anything yet. Heap replacement is not yet supported. Performance isn't great atm. Also I think the defaults are configured more for measuring things than optimizing things atm.

You need FOSE 1.2 beta 1 or later:
http://fose.silverlock.org/beta/fose_v1_2_beta1.7z
That was released less than an hour ago, so you probably don't have it yet; install it first.