For QA testing this is what I do.
example Card E1000
I download the 720p version.
When it is done, I go to the Models folder and Rename the Folder by adding a dash 720p
For card folder E1000 I rename to E1000-720p
Then I select card E1000 and change the resolution from 720p to 1080p
that triggers it to download again, but in 1080p
When that is done, I again rename the folder E1000 to E1000-1080p
Then I select card E1000 and change the resolution to 3K
that triggers it to download the 3K version.
When that is Done I again rename the folder E1000 to E1000-3K
and last, I change the resolution to 4K

This time I leave the folder as is, E1000
this is the 4K version.

I end up with 4 sets of folders
E1000-720p
E1000-1080p
E1000-3k
E1000
and in some cases
E1000-mixed ( see about my old laptop below )


I test this Version, the 4k Resolution.
when I'm done, I rename E1000 to E1000-4K
and I then can test one of the other versions.
to Test the 720p Version, I rename E1000-720p to E1000

Now I can test the 720p animations.

I don't have to make any changes in the App.
Not even do a rebuild.
I just rename the Folder, and play the Animations.

The App, knows what version the Animations are by the Animation Header.

On My Old Laptop.

I could Play the 3K version of Standing clips without any issues.
But the 3K table clips played slow..
The 1080p Standing clips were fine..

So I would create a Hybrid Folder that was made up of 1080p Table clips and 3K standing clips.

The Software Played all of this just fine, in the correct resolution because it got that information from the Animation files header.


It doen't matter which resolution you download 1st.
As long as you keep track and rename the folder so you know what resolution is in that folder
you can do the resolutions in any order you want.

The software will only use the files in the folder with a normal Name
so which ever folder you Name E1000
that is the animations the software will play.

More advanced Hybrid Card

Lets say I want to test the same Clip in All 4 resolutions.

I can copy the clip from one of the resolution folders and rename that clip.

The Last two digits of a clip, are just a 2 digit number.
So you can change these last two digits to make a custom clip in the hybrid folder

for a 720p clip I change the last two digits to 70
for a 1080p clip I make the last two digits 10
for a 3K clip I make them 30
for a 4K clip I make them 40

I often Keep the Last Digit the same...

So I will have 4 copies of the same clip name, one from each resolution, but the last 2 digits will be changed.

I can then Play any resolution of that clip, just be selection the correct one by looking at the last two digits

example card e1000 clip e1000_2052201.vghd

I make 4 copies, each from a different resolution of the card.
e1000_2052271.vghd - the 720p version
e1000_2052211.vghd - the 1080p version
e1000_2052231.vghd - the 3K version
e1000_2052241.vghd - the 4K version - there is no 4K resolution of this card though..

A rebuild collection is Needed after making any clip name changes.

https://www.istripper.com/forum/thread/29408/75?post=749653

Всё о iStripper / Share your FullScreen - Member Created Scenes here

multi Frame Delay https://drive.google.com/file/d/10YWKrvBlQ3atVdypwGxYBHTkpq-FjFBk/view?usp=sharing may not work on low vram GPU's as it uses a large texture (3840x21600) as a permenant frame buffer....

1 minute ago
@Z22

That's awesome. Sorta ***** vision. Especially effective with fast movers like Milkavi 😄 👍 😎

https://www.istripper.com/forum/thread/29408/75?post=7496531

You are not allowed to see this topic or access data relative to this topic

minute ago
@Z22

That's awesome. Sorta ***** vision. Especially effective with fast movers like Milkavi 😄 👍 😎

yeh i've been testing with saphira, especially the loop versions. arms and legs everywhere.

Anyone have these old halloween scenes?

https://www.istripper.com/forum/thread/29408/54?post=683259

Всё о iStripper / Share your FullScreen - Member Created Scenes here

I collected, made, midified aboout 47 Halloween scenes for fun this month. You can download them here. You will have seen some before & others will be new. Happy Halloween. Hope you Enjoy. https://www...

@1VGconvert
Try Here

https://virtuastripper.net/ARoomWithAView/

@Wyldanimal

Thanks!!

@Z22

I made a reduced resolution version to work on my laptop.
hope you don't mind.
I used -S for Small

https://virtuastripper.net/Z22/Z22_Loops-S.zip

I added a Zig-Zag motion
Unzip to the same folder
Dash -ZZ for ZigZag

https://virtuastripper.net/Z22/Z22_Loops-S-ZZ.zip

13 minute demo video of the ZigZag Mod
https://virtuastripper.net/Z22/Z22-Loops-Modified-ZigZag.mp4

Thanks @WA ..got to see it at last...
I was getting no frame retentions on a black background before
Great effect @Z22 👍👌

@Z22

I made a reduced resolution version to work on my laptop.
hope you don't mind.
I used -S for Small

https://virtuastripper.net/Z22/Z22_Loops-S.zip

I don't mind at all dood.

Thanks @WA ..got to see it at last...
I was getting no frame retentions on a black background before
Great effect @Z22 👍👌

I have done my usual and not renamed diferent versions of the shaders so don't lob them in the same "shaders" directory without relabeling.

@Wyldanimal

Cool mod. 😎

I've had a lot of help from the people on this forum so I wanted to share a little scene I've been working on. I write lots of scenes but most of them only reach 90% completion before I start something different 😕. I've been experimenting with circular shapes built up from many rectangular shapes.

The file is a bit large so I'd be interested to find out if it runs ok for people.

Special thanks to @Wyldanimal and @TheEmu

Flying Saucers

https://drive.google.com/file/d/109vY73MnGWOZmLnMFFpahn_JuvV9kDoo/view?usp=sharing

Wow very nice scene @Calgon. 😊

The only problem comes from my graphics card, which shows slight jerks in the girls' movements. Yet it is not used at 100%.

Damn graphics card, nothing to understand.


Wow ! Très belle scène @Calgon. 😊

Le seul problème vient de ma carte graphique, qui démontre de légères saccades dans les mouvements des filles. Pourtant elle n'est pas utilisée à 100%

Maudite carte graphique, Rien à y comprendre.

re: Flying Saucers
@Calgon this last of yours is magnificent. It compiles just fine for me.!
The colors and the camera zooming in on the center model
periodically, goes particularly well with music like P.Floyd. For a
scene deploying five 3Kclipsprites the stuttering was negligible 👍

The main thangs I changed at my end were to halve opacity of
da reflections and comment out the unused shaders & some
other stuff that might have been slowing it down some.
I believe within a node like a sprite node, only the last thus
listed (declared) shader in that node is used.

@Calgon

It's running at 15fps on a rtx 3090.

Disk Intersection.fsh appears to be costing 6fps. (am sequentially removing shaders to see where the bottleeneck is)

Fake Azimuthal EquidistantMod01.fsh cost 8fps

By ***** those 2 shaders i can get it to run at 30fps. It's only using ~30% of my GPU so i'm guessing the bottleneck is in the .scn. Have fun figuring out where ...

If i ***** all the shaders (delete) i get 33fps. Most of the girls seem ok except the girl in the centre, she is jerky (like 7fps jerky). Might be a place to start looking.

@Z22 I'm using an RTX 3080, I get about 28fps.

@EverthangForever - I put reflections behind an opaque mask and leave the clips normal when I can so that if one reflection passes in front of another you cannot see through it. The mask is at line 8145 of the published version. Changing this one line will change the appearance of all of the reflections. I've changed this to 0.5 at my end, looks better like that, they were a bit bright.

@Z22 I'm using an RTX 3080, I get about 28fps.

@EverthangForever - I put reflections behind an opaque mask and leave the clips normal when I can so that if one reflection passes in front of another you cannot see through it. The mask is at line 8145 of the published version. Changing this one line will change the appearance of all of the reflections. I've changed this to 0.5 at my end, looks better like that, they were a bit bright.

Judguing by gpu activity its not the gpu. I have a 3950x CPU. none of the cores are getting pegged full but there is a lot of spikes on 4 cores, which are on the same ccd so it's not a cross core latency problem. It's not the number of girls, i have made scenes in the past with 5 and used to be able to run 20 girls on the desktop (till they nerfed it).

As ET says, if it's only displaying the last shader in a node but is still calculating the others it could be the cause (worth investigating).

hmm..I haven't had time to look closely at the .scn code, however with a third of a million lines of code in a single .scn it seems like a good place to start
Sure, i'd expect the model in center gets jerky because 3D camera zoom rescaling up mostly goes there.
There's "0;" ending typos btw in every sprite node rot parameter...etc..etc.
@Z22 what resolution models are you using to test this ?
What happens to overall FPS when you // eliminate each model ?

hmm..I haven't had time to look closely at the .scn code, however with a third of a million lines of code in a single .scn it seems like a good place to start
Sure, i'd expect the model in center gets jerky because 3D camera zoom rescaling up mostly goes there.
There's "0;" ending typos btw in every sprite node rot parameter...etc..etc.
@Z22 what resolution models are you using to test this ?
What happens to overall FPS when you // eliminate each model ?

Am just using the default resolution

@calgon, you have a bunch of errors in the .scn for ***** things like "Sprite" should be "sprite" to sprites with multiple shders active in them, redundant things like allow and block (you only need one or the other). Am having a go at cleaning up the code so we can see whats going on. though i'm not sure whats meant to be goin on in some eg:-

sprite
{
source: Texture
size: 41.2682932795517, 1500
pos: 5000, 0, 0
hotspot: 1, 1
scale: 2, 2
rot: 0, 270, 0;
shader: fragment, Shaders/Generators Redux.fsh
shader: fragment, Shaders/Disk Intersection.fsh
shader: fragment, Shaders/calgon_black.fsh
}

That is calculating all three shhaders but only displayin the last one AFIK.

@Calgon , if you use Notepad++ you can do a search/replace for any wonky code. Can also remove
empty lines and justify lines left for ease of reading.
saves a lot of time. Humm, ..not sure why needs such huge decimal places in the size parameters (?)😉

@calgon, you have multiple textures with the same name.

texture
{
id: Texture2
source: Images/Black_1010.png
}

texture
{
id: Texture
source: Images/Black_1010.png
}

texture
{
id: Texture
source: Textures/Metal - Rusty - 01.jpg
}

texture
{
id: Texture2
source: Images/Black_1010.png
}

Three calls for the same texture file?

@EverthangForever - the code is generated from a separate program that calculates the size, rotation and maths of the placement of "stuff". I haven't worried about the tidiness of the final code at this point because I don't edit the scn file directly. It's more about seeing whether it works at all at this stage. I was shocked when I first found that a file of 200k lines would run at all.

@Z22 "multiple textures with the same name" - yarp, I'll fix that, though these are tiny files and unlikley to affect the performance.

Keep those "snags" coming because the program has many things that are set as a single parameter and can easily be changed and the full scn code regenerated.

Should I produce a "lite" version with less detail on the circles ??

Yeah, use Notepad++ Edit >Line Operations to tidy it up a tad first methinks

@EverthangForever

There's "0;" ending typos btw in every sprite node rot parameter

Gotcha - thanks

@Z22 re:

shader: fragment, Shaders/Generators Redux.fsh
shader: fragment, Shaders/Disk Intersection.fsh
shader: fragment, Shaders/calgon_black.fsh

Gotcha - changed this as well but it made no performance difference.

I have no idea how the core program works behind the scenes but I would have thought that when the code is compiled the final shader is the one that is calculated but it was worth a try and I "shaved off" about 50k lines of bloat from the code.