Posted: 10/27/2023 5:35 PM
Quote
feos
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
NES TAS of 2011
We accept PCem Windows XP movies for submission now, and in addition to our guides we also need to have a place for all the discussion about this setup, results of tests and all sorts of useful advises, which we could add to the wiki guide.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Posted: 10/29/2023 3:09 PM
Post subject: Navigating bootup
Quote
InputEvelution
She/Her
Editor, Reviewer, Player (36)
Joined: 3/27/2018
Posts: 194
Location: Australia
The duration of time it takes for Windows XP to actually finish booting up and take you to the desktop is rather inconsistent between different time-and-dates in libTAS (with half of the outcomes being a full 17 seconds longer!). I've documented the duration and outcome of some of these possible starting times in a spreadsheet here: https://docs.google.com/spreadsheets/d/1_j1zkYPPHR4jLLW3bybvnjsN6K7sV_vWZHLm7X5SwaE Windows XP in general seems to be rather sensitive and unpredictable with lag. Simply moving the mouse around on screen seems to be able to affect the duration of loading times (often for the worse), and clicking on things the first frame they appear often seems to result in slower loading times than waiting a few frames beforehand. Without a better understanding of exactly what Windows does behind the scenes with managing tasks, the best solution here seems to be one of trial and error. Regarding avoidance of the mouse, I've found a rather comprehensive cheatsheet for keyboard shortcuts that may prove useful to others: http://www.shortcutmania.com/Windows-XP-Keyboard-Shortcuts-printable-cheatsheet.htm CD loading times in particular seem to present a difficult hazard for TASing, as loading earlier savestates after PCem has started reading from the CD image seems to cause desyncs, and even strange errors from XP itself on some occasions: Running the movie from the very start seems to give consistent timings without errors, but combined with XP's strange lag inconsistencies, it seems game installation in particular is likely to be a pain to optimise.
Posted: 10/29/2023 9:06 PM
Post subject: Re: Navigating bootup
Quote
Sand
He/Him
Player (143)
Joined: 6/26/2018
Posts: 175
InputEvelution wrote:
The duration of time it takes for Windows XP to actually finish booting up and take you to the desktop is rather inconsistent between different time-and-dates in libTAS (with half of the outcomes being a full 17 seconds longer!). I've documented the duration and outcome of some of these possible starting times in a spreadsheet here: https://docs.google.com/spreadsheets/d/1_j1zkYPPHR4jLLW3bybvnjsN6K7sV_vWZHLm7X5SwaE
Wow—super interesting! Just looking at the "Res increase" column, it seems to oscillate between 44.4 s and 53.42 s about every 10 years. (The difference between adjacent rows is 100000000 s, or about 3.17 years.) I wonder if the dynamics are as simple as: there are particular timestamps at which the time required to boot suddenly changes. For example, timestamp 1300000000 yields 44.4 s and timestamp 1400000000 yields 53.42 s: would a binary search find a pair of consecutive timestamps where one is 44.4 s and the other is 53.42 s?
Posted: 10/30/2023 7:48 AM
Quote
DigitalDuck
Active player (390)
Joined: 12/20/2022
Posts: 38
Location: UK
InputEvelution wrote:
The duration of time it takes for Windows XP to actually finish booting up and take you to the desktop is rather inconsistent between different time-and-dates in libTAS (with half of the outcomes being a full 17 seconds longer!). I've documented the duration and outcome of some of these possible starting times in a spreadsheet here: https://docs.google.com/spreadsheets/d/1_j1zkYPPHR4jLLW3bybvnjsN6K7sV_vWZHLm7X5SwaE
Every "long" load occurs during daylight savings time, and every "short" load occurs outside of it. I wonder if that's relevant.
Posted: 10/30/2023 1:00 PM
Quote
InputEvelution
She/Her
Editor, Reviewer, Player (36)
Joined: 3/27/2018
Posts: 194
Location: Australia
DigitalDuck wrote:
Every "long" load occurs during daylight savings time, and every "short" load occurs outside of it. I wonder if that's relevant.
Great observation, I hadn't noticed that at all!
Posted: 10/30/2023 3:10 PM
Quote
feos
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
NES TAS of 2011
Oh god so it makes a difference in the end D:
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Posted: 10/30/2023 3:29 PM
Quote
InputEvelution
She/Her
Editor, Reviewer, Player (36)
Joined: 3/27/2018
Posts: 194
Location: Australia
feos wrote:
Oh god so it makes a difference in the end D:
Well, I did a little bit of testing from the v2 verification movie before this, and the same starting times were longer there. So I suspect that either the removal of daylight savings doesn't actually get saved due to restarting during OOBE, or that it just takes longer during DST anyways for some reason.
Posted: 10/31/2023 2:54 PM
Quote
Dacicus
Editor, Player (69)
Joined: 6/22/2005
Posts: 1050
This could be legally bad for Microsoft if their OSes intentionally hinder productivity due to DST. Quick: Someone leak this to the press!
Current Projects: TAS: Wizards & Warriors III.
Posted: 11/4/2023 5:02 PM
Quote
feos
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
NES TAS of 2011
InputEvelution wrote:
CD loading times in particular seem to present a difficult hazard for TASing, as loading earlier savestates after PCem has started reading from the CD image seems to cause desyncs, and even strange errors from XP itself on some occasions: Running the movie from the very start seems to give consistent timings without errors, but combined with XP's strange lag inconsistencies, it seems game installation in particular is likely to be a pain to optimise.
I think this proves that game installation deserves to be in a separate movie if a CD is used. Do you know if floppy disk installation has the same problem?
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Posted: 11/5/2023 3:22 PM
Quote
InputEvelution
She/Her
Editor, Reviewer, Player (36)
Joined: 3/27/2018
Posts: 194
Location: Australia
feos wrote:
I think this proves that game installation deserves to be in a separate movie if a CD is used. Do you know if floppy disk installation has the same problem?
I don't, unfortunately. I'm not even sure if I currently have a working floppy disk reader...
Posted: 11/5/2023 4:10 PM
Quote
feos
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
NES TAS of 2011
Here's what I came up with https://tasvideos.org/Wiki/PageHistory?path=EmulatorResources%2FPCem&fromRevision=80&toRevision=81
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Posted: 3/11/2024 7:45 PM
Quote
feos
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
NES TAS of 2011
Link to video
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Posted: 5/30/2024 3:32 PM
Quote
chuky
Joined: 5/30/2024
Posts: 1
For some directx 8 and 9 games, it can be useful to use a wrapper to render in software, because directx 8 and 9 are too modern for the Voodoo 3 which is currently the best graphics card emulated in PCem. Some games can look wrong or not even start in hardware rendering, but with software rendering it can be fine. Be aware that it's very heavy on the emulated cpu. With Pixomatic in Windows XP, after upgrading to directx 8.1 I'm able to play Hamsterball. The main menu still has a black picture but without the wrapper many polygons would be black and textures would have no transparency. I was also able to play Hotline Miami in Windows XP, with a different version of Pixomatic, but the game seems to be too slow for the emulated cpu with software rendering. It's also a Directx 8 game. You just take dlls and files from the shader directory and put them in the same directory as your game executable. For Hamsterball I used Pixomatic version 2 from here https://www.vogonsdrivers.com/wrappers/files/Unsorted/rast.7z For Hotline Miami I used Pixomatic version 2 from here https://www.vogonsdrivers.com/wrappers/files/Direct3D/Software/Pixomatic/pixomatic_softogl.7z There's also an opengl wrapper in the zip. For other games, SwiftShader version 2 may work better but I didn't find good examples. There's a version of SwiftShader version 2 in the first link. Here's another version of SwiftShader 2 https://www.vogonsdrivers.com/wrappers/files/Direct3D/Software/SwiftShader/SwiftShader%202.x/ There are other versions of SwiftShader so if you're interested you can search for them. People compiled versions to run in Reactos so you can try them as well.
Posted: 5/30/2024 6:16 PM
Quote
feos
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
NES TAS of 2011
Amazing find!!!
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Posted: 10/6/2024 11:04 AM
Quote
InputEvelution
She/Her
Editor, Reviewer, Player (36)
Joined: 3/27/2018
Posts: 194
Location: Australia
For some reason, PCem only seems to respond to mouse inputs from the player every second frame. (e.g. mouse button inputs won't be registered if performed on odd-numbered frames, and mouse movement will only be detected if it carries over to an even-numbered frame). At PCem's default framerate, this equates to a mouse input rate of 50Hz, while keyboard inputs can be performed at 100Hz. If PCem st-2 is emulated at say, 200fps, then the mouse input rate will instead be 100Hz. It may be best to run PCem at at least double the framerate of your game (or otherwise wait for a fix) if you're planning on making a TAS.