Framerate Tanks When Any FF11 Instance Is In Focus

Operating System: Windows 10 Pro
Desktop PC
CPU: Intel I9-9900k
GPU: RTX 3090
Corsair Vengeance 16gb x2 DDR4 ram @ 3200Mhz
Final Fantasy 11 installed on an NVME
Screen @ 3840x2160 120hz through HDMI 2.1 (Final Fantasy 11 instances @ 1920x1080)

The issue: Framerates on all FF11 instances drop a lot if any FF11 instance has been clicked on/is in focus.
All instances run at full speed if they are out of focus/in the background.


Things tested: (As many as I can remember, the things I have tested over the years have been on impulse so I did not record them at the times).

• The framerates fluctuate when the problem happens, they do not sit at set FPS, and the fluctuations are random, not between divisors of the screens refresh rate
• The issue occurs in both windowed mode and borderless windowed mode
• Removal of VRR and Gsync (both game specific, and system wide)
• Setting framerate cap at 60 (both game specific, and system wide)
• Vsync on and off (both game specific, and system wide)
• Windows game mode on and off
• Hardware-Accelerated GPU Scheduling on and off
• Prefer maximum performance set in Nividia Control Panel
• Background Application Max framerate set to off and also tested set to 60 in Nvidia Control Panel
• Systematic process of testing the mentioned Nvidia Control Panel settings along with other Nvidia Control panel options
such as threaded optimisation, for all these things I go through a process of testing each 1 at a time, and also every
combination of the settings that I am changing. (If there are any specific settings you want to know if I tried from the Nvidia Control
panel that I have not mentioned, please just ask and I'll be happy to answer.)
• Disabled IGPU in BIOS
• Windows advanced system properties/Processor Scheduling tested both "Adjust best performance for Programs" and "Background Services"
• Tested with Hyper-V disabled
• Tested disabling hyper threading
• Systematic testing of CPU core and thread combinations, and putting FF11 instances onto them 1 by 1 singly, and also
in sets of 2, 4 and 8.
• Tested all CPU priorities other than real time
• 2 cores/threads max out at 100% in the case of my current 2 FF11 instances test, 4 max when using 4 instances, they max out when both windows are out of focus
and performing at high FPS. When I bring an FF11 window into focus, and the FPS dies on all FF11 instances, the cores that were maxed drop to around 80%
utilization. (In the case of the 2 instance test, it was one core and one thread that were maxing out, CPU2 and CPU15.)
• The issue is not present when running only one instance of FF11.
• Any changes that required to be pointed to an FF11 exe I have done across all of pol.exe, polboot, and all windower.exe's and also each 1 at a time.
• Used CPU and GPU monitoring software to confirm there are very minor changes in GPU utilisation, and CPU utilisation overall.
• The GPU monitoring software shows clocks are not being throttled.
• The CPU frequencies stay at 5.1 GHz without change when the issue is present (focused FF11 window) and when it is not (FF11 windows out of focus)
• CPU and GPU temperatures way below throttling thresholds, both in the 50's.
• Tested with and without a controller plugged in, this includes the bluetooth dongle I use for the wireless controller.
(The controller being used is a wireless Dualsense.)
• Most backround apps, internet browsers, discord, etc. not even intitiated, I run them all on a seperate tablet when doing these
tests on my desktop. (again, if there are specific programs that people want me to check are running in the background etc.
please just ask and I'll be happy to answer.)
• Tested clean install of my graphics driver.
• Using no programs with overlays during testing
• Can confirm my clean installs of windower have no residual files, I keep backups of clean installs before using them
and have been using the portable installs, so when testing I am using untouched installs
• Most of, if not all of these tests have been done on my main windower installs prior, but all have been done with clean installs also
• Tested it on 2 instances of Ashita
• The issue is present on 2 other computers I have tested, with completely different hardware and one on Windows 11
• Dgvoodoo is pointed to the correct folder with the pol.exe
• I have tested with and without Dgvoodoo, I use the preset settings and the only one I change is the VRAM in the DirectX tab
• I have tried more than one version of Dgvoodoo and other DirectX wrappers
• I have run with Dgvoodoo on, off, at 4096mb VRAM and 1024mb VRAM with the other settings at their default values
except for removing the watermark after confirming Dgvoodoo is running and using "disable and passthru to real DirectX"
(if there are other settings anyone wants me to change/try in dgvoodoo please just ask.)
• Tested the blockmouse addon, one of my thoughts at the time was that the problem could be related to the mouse
This addon destroyed my framerate when active, 3-4 fps, but another user said it worked fine for them
• I have made sure I disabled/revert all my tests after finishing

I have been asked which DAT and graphics mods I am using, I will try to explain this and both my FFXI config and Windower settings as best I can in a
moment, the latter 2 I will post a link to screenshots. (feel free to skip to them past the upcoming text).

Everything I have posted here has been in good faith, I would never ignore or discount help that is offered, it would be counterproductive to the very reason I
created this thread, I am not a person who likes to bother or take up others time, so the layout of my posts were always with the intention to bother people as
little as possible by not writing a wall of text, I thought it might also be off putting and people would not read it. (along with me not having documented my previous tests).
My only thought was to not waste others time, which is why I thought I was doing a good thing when replying about things I had already tried.
I do think it was unfair to call me out on a public forum for something that was a preconceived notion, which could not have been farther from my intentions.
I did respond in a manner that was too blunt to Rubenator, for that I already apologised, for what it's worth, that exchange happened as I was getting ready
for bed after almost 2 days without sleep, and I maybe would have responded differently had I just went to bed first.
Like most people here, I highly respect what both Rubenator and Thorny do, so i was taken aback a bit when the conversation went to where it went.
I responded to what I took at the time to be antagonistic in Rubenator's third reply, which is no better than Thorny's assumption about my intentions.
So I do hope that was another misunderstanding, and hope I showed some humility in my apologies for my blunt response, and my assumption, and would hope
Thorny can also do the same for his assumptions and vocalizing them in a public forum.
Again, both of your helps were and are still appreciated, I'll leave it at that.

*DAT's and settings info continued*
This is where things get a little less straight forward with my info now lol
My DAT mods over the years i've tested in a specific way, and that was to only ever install one at a time, do a period of testing to make sure there are no
issues, then permanently add it to an overall installed DAT folder.
This DAT folder has grew to be very large, and I have not documented all of the mods inside of it.
Due to this it's almost impossible to list the DAT mods I have installed unfortunately.
I did answer before though, that a lot of my testing is with clean installations, so no DAT mods involved, I do hope that's a good enough answer
as they cannot be affecting the issue on those tests, so hopefully we can strike them off the list, but if I'm overlooking something, please let me know.

The reason I had so much confidence to say I'd already tested things for some of the suggestions, even if I did not remember till they mentioned
it, is simply because I always take the same approach when testing these things, and do a systematic run through settings, and setting combinations,
unfortunately OCD will not let me do it any other way, it was not to be dismissive, again, thought I was being helpful.
So even though someone has just made me remember I tested something, I can be certain when I did, my approach was systematic, because It is how I always
do it.
Something like the Windower settings for example, even if it was something like 2 years ago, I can be very confident that when I did test them, I will
have tried them at multiple resolutions, windowed, borderless windowed, fullscreen, and a combination of every setting 1 at a time, gradually building up,
2 settings at a time, 3, and so on and so on and every combination of the settings starting at 2 at a time and building until I've exausted all the combinations
of settings.
Some considered this dismissive, and I did apologise if it came off that way, but I thought I was streamlining it for those helping, and am always still open
to the help they were offering, me saying that wasn't a full stop on their recommendation, it was just me trying to convey that I had tried that suggestion.
But not being aware of the way I do my testing definitely caused a misunderstanding, because when troubleshooting of course you go through things one at a time
and be systematic about it, but they were not aware that, I can't do my tests any other way, and they were just doing what is the normal thing to do when
helping/troubleshooting.
Because I also know the troubleshooting process, I figured I was saving them a lot of time, hoepfully this explanation sheds light on that part.

Here is the link to the screenshots of my FFXI Config, and Windower settings: https://github.com/KruelFantasy/FF11-Windower-Settings.

Hopefully that covers everything, please if there's any info needed I have not supplied, let me know, again, this is everything I could remember.
I have spent a lot of hours re-approaching this problem over multiple years, and it would have been much better had I documented it all, but I go on OCD
benders when I do these things, and I'm just jumping from one test to the next, and do not stop to document unfortunately.

Thanks everyone who take the time to read through all of this, this post has covered everything talked about in this thread so far, so hopefully saves
people time having to find all of it in the thread.

Not trying to be a *** since you wrote a giant post, but listing out everything you've tried over years is still not helpful.

As I stated yesterday, the correct troubleshooting method is to test one thing at a time, with no other changes in between, and document that it's been tested and exactly how. I understand you want to avoid repeating existing work, but the truth is you could've done all of the listed testing in an hour or so if you have even minimal computer proficiency.

Again, nobody who has ever done IT is going to take your word on this. You might luck into a correct solution from someone throwing ideas out, but the idea that you tested something years ago and couldn't possibly have done the test wrong so you won't redo it is incredibly offputting. I'm sorry if your disorganization and incompatibility with this method is due to your OCD, but it doesn't change that it creates a very bleak picture for someone who wants to help.

I am not going to get into a back and forth on this, but did you read the parts where I said I am only trying to save others time?

Yes trying to avoid repeated work is one thing, but my sole intention for that was me trying not to inconvenience others, not myself.

I stated multiple times I would have no problem testing those suggestions, and my only reason to reply the way I did was to try and save the person offering help time, how have you managed to again try to make me feel bad about that?

You have stated "you won't redo it is incredibly offputting", but where are you getting that from?

I explicitly said this, "and am always still open to the help they were offering, me saying that wasn't a full stop on their recommendation, it was just me trying to convey that I had tried that suggestion." and on multiple other occasions in this post made sure to make people aware that I am not discounting any advice, and am willing to still try.

Your final comments about my OCD which seem to be trying to be hurtful are not productive either, I said them in a context in my post to help explain the situation as best I could, you on the other hand, worded it in a way that comes across like you are trying to put me down, if I had written that, even just a second glance would have made me realise how insulting it could sound, and made my best efforts to fix that.

You have taken a long, carefully thought out post, in which I was trying my hardest to explain everything as best I could, and all of it is intended as starting points, and just taken and twisted snippets of it for a response that was completely unhelpful and insulting, shame on you.

What was wrong with collecting all of the things that have been went over in this thread and culminating it in a single post?

I would of thought that would be helpful, instead you decided to say, "but listing out everything you've tried over years is still not helpful".

I don't know how else I could've written anything to not set you off in some way it would seem.

I went over that long post for over 4 hours to make sure I wasn't offending or insulting anyone, I wish you could've given your post even 5 minutes to do the same.

Please do not respond again unless it is to apologise.

I was willing to do that for what I wrote yesterday, and admit when my response was not appropriate, I hope you are capable of the same.

There's a strong possibly that your GPU is just not ramping up the clocks and staying in a low power state as this game take nothing to run and uses pretty old DirectX, I would use something MSI Afterburner and manually set the clocks and see if this fixes your issue, not an uncommon issue with a powerful GPU and an old game.

Thanks for the suggestion, I will edit my post and add that to the list.

It is something I have tried already and forgot to add to the list, so I will edit and add it now, and will now try again as to make sure it doesn't look like I'm ignoring peoples advice.

The clocks were not getting throttled in my tests, but worth testing again.

Thanks for the idea.

edit. the GPU clock is still staying stable and not throttling unfortunately, thanks again for the idea.
It is MSI Afterburner I use, and did also try some manual clocks.

I'm not trying to make you feel bad. I am pointing out that as someone who has done FFXI troubleshooting for literally hundreds of people for free, it is very common for someone to claim they have already tried something that later turned out to still be their solution. That doesn't mean they were lying, they may have felt they tested it in the past and either forgotten, or made a mistake. If I were solving an issue like this for myself, I would still start by going over the most likely causes and documenting them, even if I thought I had checked sometime in the past.

Someone who understands the issue and is trying to help you work through it may want you to do things you've already done, and they're not going to take your word for it when it's presented in the scattered fashion that you've prevented it. This isn't because they think you're lying, it's because there is a finite amount of possible causes for an issue and the symptoms give indications of which are more likely. Those indications point toward things you claim to have already tested, and it makes more sense to thoroughly go over them and make sure you've tested them properly before going into the weeds on more obscure causes.

I meant no offense to your disability, and clearly my input is not welcome here, so I'll leave you to it.

I thank you for this response.

I had no preconception that people might think I am lying, honestly I've said it quite a few times now, my thoughts were always only about making it easier on the people helping me, after taking onboard your advice from your posts last night, this is a big part of why I created the large post today, you made me realise that was the wrong way to do it.

It killed two birds with one stone, gathered together all of the info that had been discussed in this thread into an easier reference point, and also allowed me to make edits to it as a way of documenting things as we try them, and keeps anyone new coming to the thread up to date.

I am still thankful for your help offered, and your most recent post.

Does DGVoodoo actually show your GPU in the "Adapter(s) to use/enable" section?

Which powermode are you using? Ultra?

What's your performance options look like?
Advanced system settings > Advanced Tab > Under performance click 'Settings' > Click Advanced Tab

When it "drops FPS" what is the actual FPS show? a locked number? or fluctuates?

Yes, my 3090 is shown in the adaptors to use in Dgvoodoo.

Windows is set to its high performance plan, the highest power plan available, and all the games executables are set to prefer maximum performance, please let me know if you meant a different setting to set to ultra.

Advanced system settings > Advanced Tab > Under performance click 'Settings' > Click Advanced Tab is currently set to Programs, but tested it with that, and tested it being set to background services also.

edit. sorry forgot to add the Framerate fluctuates when it drops FPS.

It doesn't sit on numbers that are divisible by the screen refresh rate, if that extra bit of info helps.

I meant ultimate performance, but high performance should be good enough as that's what I used on my last PC.

Everything I've read points to gsync/Vsync/borderless window problems.

Might be worth your while to see if reformatting helps as you seem to have put a lot of manhours into this.

Thanks very much for the suggestion.

The main thing holding me back from just reformatting/re-installing windows is, the problem is replicated on my other 2 PC's, a Surface Pro 9 Tablet (windows 11), and a backup desktop PC I have running an Intel 11400F with an RTX 4060TI @ 1080p 144hz.

I do plan a complete rebuild of my main PC when the new RTX 5000 series comes out, and Intels upcoming new gen CPU, so was trying to avoid having to do a full re-install before that lol

Things just happen sometimes that make me re-visit this issue, this time it was playing FF11 as one character for the first time in a long time, and just simply enjoying the game again without all the FPS lag problems.

I was dreading going back to multiboxing again thanks to that, so I spontaneously went back down the rabbit hole which is this problem.

First off: wow the 'tism is loud in here.

Secondly it is not recommended to have your framerate uncapped, ever. All except the main instance should probably be capped at 30.

That being said on a cursory read it looks like one of these two possible scenarios:

1. The fact that it is doing it across multiple computers is likely telling you're running some software that is causing the issue.

B. you need to reinstall from scratch and do ONE THING AT A TIME like was said. May not even need to reinstall just don't use datmods and use a fresh windower copy and see how it runs on each computer.

That all being said: you're better off looking for assistance in a discord.

you don't need an addon for this, it is literally a feature in windower...

If you have a second monitor you can test this but also, you should be able to alt-tab between a windowed fullscreened instance and windowed instance on one monitor.

Your CPU being 100% is very suspect as well. To me that suggest your CPU is emulating something for FFXI. To put that in perspective, my 7800x3d (32GB ram and a 7900xt) only use ~15% cpu when playing FFXI. Try using a different Sound device, like a USB headset. In googling your problem I saw a few instances where a sound device was the culprit for low foreground FPS (which can make sense if sound only plays when the game is in focus). You could also try turning off sound in the windower profile.

Yeah, you're correct, I can alt tab between a windowed fullscreen instance and a windowed instance on one monitor, both windowed fullscreen and windowed are both affected by the issue, even when using the combination of two like you suggested.

My actual CPU overall is only hitting around 20%-30% usage while playing FF11 multibox, the 100% usage was just on specific cores and threads.

Your computer, is it using 15% overall, or was it 15% on the cores/threads FF11 has been designated?
If it is only using 15% on the cores/threads FF11 is using, it will be something else I'll look into.

I am going to refocus on the sound problem you mentioned, and work that angle as my current testing, thanks very much for the suggestion.

w/ 3 XI instances running at 2425x1358 each, 60fps supersampled HD textures all that trash. Lots of background applications, web browser, etc.; my CPU is sitting at 9~11% on the cores I have it bound to, and GPU is at around 25% sitting at the minimum possible clock it can be in 3D rendering mode. As drakefs said, definitely something wrong if you're at 100% CPU. 9900k is old but it isn't that old.

Thanks for the response Prothescar, was still addressing that when you responded, so only managed to post it just before your post, the info is in the post directly above your post.

edit. Very interesting to know yours has 9~11% on the cores you have it bound to, confuses me all the more lol, because this issue is on all 3 of my computers.

One of the computers is very barebones as far as installed programs etc. goes, compared to the other two, they all have different hardware, and one has pretty much zero in common with the others as far as extra programs installed that could be a good direction to look in for possible conflicts.

Really appreciate that info, now scratching my head even more lol

It's late here now, and gonna try a few more things tonight then try and catch up on sleep.

I've decided to try the nuclear option tomorrow if I don't get lucky with the things I'm trying tonight.

But not going nuclear on my main PC, I plan to reformat/re-install windows on my backup desktop PC, and only install FF11, Windower, and only activate the config plugin.

The backup PC is still total overkill for FF11

Windows 10
1080p 144hz monitor
Intel 11400F
RTX 4060TI
32gb RAM
GEN 4 NVME

This is gonna be as barebones a fresh install as I can test, so if the issue is still there, I give up.

Once you see the light and go back to Balanced power mode (disable sleep timer), you never go back to High Perf anything

tbf im running a super overkill CPU (7950X3D) so low utilization is to be expected. I have enough cores to assign 2 to each XI instance without negatively impacting any other programs. (Project Lasso is great for this cause it doesn't forget what you want it to do)

I feel like this is somehow related to dgvoodoo. I have a similar issue as OP for a while now and its strange also. Currently my CPU and memory keep spiking near or past 100%(and it closes instances). I have a lot of crap open, but the games themselves don't lag until I alt tab out and into other stuff... Like I could have an XIV instance open, it will play perfectly until I alt tab out. Then the whole PC lags until I alt/tab into an XI instance and it goes away. The weirdest things.. I thought maybe I had a setting in dgvoodoo too high, but I am not sure what it is. The PC itself is brand new with all of the up to date specs for the most part.

the ~15% is the amount of processor utilization that the FFXI process is using.

CPU2 is being loaded 50%~70%, so still weird your CPU core is hitting 100%.

Well I have just finished my biggest test.

A completely fresh install of windows, with only the graphics driver, FF11, Windower, and Dgvoodoo installed.

Initial tests were without Dgvoodoo, used a completely clean FF11 install.

Nothing else activated on Windower except the config plugin.

The problem still happens...

The issue has now been present on 3 different PC's, two on windows 10, one on windows 11, and all are running Intel CPU's.
All have different hardware, and none have any usb peripherals in common.

This should really narrow down the troubleshooting, but crazy that this test did not fix it.

My thoughts lead me towards possible BIOS settings, Intel CPU's, or specifically, how my Intel CPU's are handled by Windows in the FF11 multi instance situation.

Really wish I had an AMD CPU to test.

Can anyone else confirm they multibox on FF11 at 60 fps with an Intel CPU without the issue?

How exactly are you testing the FPS loss? You aren't just logging all your characters into Western Adoulin are you? Western Adoulin seems to have a driver issue related to the smoke effects. Having 1 character in Western Adoulin will negatively affect all my FFXI instances regardless of where the others are.

Other things you could try if you haven't:
Test with C-states disabled
Lower mouse polling rate if using a high polling rate mouse.

Can you confirm that nothing - zero files - came from or was copied or transferred from the other computers - no install files, dats, config files, everything was downloaded fresh - including windower and xi files?

Edit: I am on a 13700k Win 11 system - I wasnt paying attention but had no issues multi boxing 2 chars during the free login campaign. I might resub the mule just to retest for you.

It's not your cpu. It is your settings. The fact you are able to replicate on different computers with the same install should tell you something.

You need a completely clean install, do not install dgvoodoo, do not change any settings. Just log into windower. Once you confirm, in a stable zone while staring at a wall, eg the mog house, that performance is ok, then you start adding dgvoodoo and changing things one at a time.

One of the things you are changing is doing something. You won't know which until you stop doing more than one thing at a time.

You are skipping steps B C D E F and G and wondering why you can't go from A to F.

Yeah, I'm aware of the Western Adoulin shader issues, been using there, Nashmau and Batallia Downs (S) as my main testing areas, as they are the easiest areas to get low FPS in.

The other ways to drop my FPS involve instanced areas with 20+ hour lockouts unfortunately, but can confirm when I get massive FPS drops in places like Sheol C, the moment I put all my FF11 windows out of focus, all the enemies, battle effects etc which were dropping the FPS do so no longer, all windows run at 57 FPS solidly.

Re-tested Disabling C-states last night.
Tried 4 different mice, and several different polling rates.
Also tested without a mouse enabled in game.
Thanks for the ideas.

Yeah, I can confirm everything tested in last nights re-install were freshly downloaded.

It is much harder to find spots to get my FPS to go lower than 57 only dual boxing, even with whatever the issue is doing, my specs can overpower the issue in most places, so maybe your specs are high enough to have not seen the issue dualbox?

It is still confirmable by checking CPU cores/threads and unlocking FPS though, the issue will still happen at like 100 FPS, so is still testable if you struggle to find spots that go lower than 57.

If the issue wasn't most problematic in places like sortie, odyssey, etc. I wouldn't care about it lol.

I really appreciate you offering to re sub a character to test, but please do not go out of your way, and especially do something that costs you money to test on my behalf.

I am very thankful for all the help you've offered, and really appreciate that you would go to that effort.

That was exactly what I did lol

First test was with changing absolutely nothing.

These are your words.

First you say you installed ff11 windower and dgvoodoo. Then you say "initial tests were without dgvoodoo".

Both of these can't be true. You either installed dgvoodoo or you tested with default settings. Installing dgvoodoo is not using default settings.

Again, change one thing at a time. The back and forth with "I did this but before that I did that" tells me you are not taking it one step at a time which will lead you nowhere.

If you do manage to not change any settings, that means not installing dgvoodoo or changing any config files what so ever, you can try disabling all effects in game to see if that changes anything.