Re: CODERS: New cross-platform refresh rate / raster estimator useful to emulator authors (sync emuHz to realHz better)

Posted on 2023-08-10, 21:14

by

by mdrejhon

in Console & Arcade Emulation

in Console & Arcade Emulation

Video of Kefrens Bars in Javascript via VSYNC OFF frameslice beam racing: https://www.youtube.com/watch?v=Bk20l7akRUk It's doing 2000 frames/sec, so that's 2000 pixels rows/sec, works best at low refresh rates. Also unscreenshottable (only captures one pixel row as a stretched full-height frame), so …

---

Re: CODERS: New cross-platform refresh rate / raster estimator useful to emulator authors (sync emuHz to realHz better)

Posted on 2023-07-25, 01:08

by

by mdrejhon

in Console & Arcade Emulation

in Console & Arcade Emulation

Still discussing behind the scenes, how to publicize browser-based raster beam racing. Practical prototype confirmed on several Windows machines. Works in multiple chromium forks, although not in Edge or FireFox. P.S. I wish a HTML5 API existed to turn VSYNC ON/OFF, but perhaps publicizing this will …

---

Re: HOWTO: Possible Lagless VSYNC for Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC OFF

Posted on 2023-07-15, 10:02

by

by mdrejhon

in PC Emulation

in PC Emulation

One additional important requisite: New Emulator Latency/Jitter Rule: Prevent GPU Power Management Between Frames Try not to let a GPU idle for a full millisecond between graphics API calls (drawing or presentation) This is due to power management. GPUs can go into idle mode (0Mhz-like operation) …

---

Re: CODERS: New cross-platform refresh rate / raster estimator useful to emulator authors (sync emuHz to realHz better)

Posted on 2023-07-14, 22:48

by

by mdrejhon

in Console & Arcade Emulation

in Console & Arcade Emulation

SUCCESS! (shh). Just PM'd VileR who I'm letting beta test the world's first browser-based rasterdemo. EDIT: Confirms that it also works on VileR's GPU too, so he can vouch I actually put "web browser" and "beam racing" in the same sentence! Keeping details under wraps, trying to decide how to …

---

Re: CODERS: New cross-platform refresh rate / raster estimator useful to emulator authors (sync emuHz to realHz better)

Posted on 2023-07-14, 02:30

by

by mdrejhon

in Console & Arcade Emulation

in Console & Arcade Emulation

...This won't be useful for rasterdemos in a web browser since they're permanently VSYNC ON and do not generate tearlines (no way to listen to VSYNC ON tick-tocks while running in VSYNC OF mode) -- but will work with high-framerate VSYNC OFF standalone software, for precise tearline steering, but …

---

CODERS: New cross-platform refresh rate / raster estimator useful to emulator authors (sync emuHz to realHz better)

Posted on 2023-07-13, 02:16

by

by mdrejhon

in Console & Arcade Emulation

in Console & Arcade Emulation

I'm the founder of Blur Busters and creator of TestUFO, and want to make an announcement of one open source code module (used in TestUFO and VSYNCtester) that will be useful for some emulator authors. I happen to have experimented with the world's first cross platform Kefrens Bars (works on PC and …

---

Re: HOWTO: Possible Lagless VSYNC for Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC OFF

Posted on 2023-07-12, 22:14

by

by mdrejhon

in PC Emulation

in PC Emulation

Update! I got one of the modules relicensed to Apache 2.0. Source code now released in this thread: CODERS: New cross-platform VSYNC timestamp estimator useful to emulator authors (sync emuHz to realHz) Crosspost here, hidden in quotes: Excellent news: Relicensing vsync estimator to open source …

---

Re: Emulating CGA Snow

Posted on 2023-07-12, 03:47

by

by mdrejhon

in PC Emulation

in PC Emulation

The green shows up when the character byte on the bus gets read as the attribute: that test fills the screen with the sequence "20 0F", i.e. character 20h (space) with attribute 0Fh (white on black). If the 20h shows up when the CGA fetches the odd (attribute) byte, it'll set the color to black on …

---

Re: HOWTO: Possible Lagless VSYNC for Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC OFF

Posted on 2023-07-11, 22:13

by

by mdrejhon

in PC Emulation

in PC Emulation

Candidate Source Code For De-Jittered VSYNC Estimators To save y'all time on VSYNC dejitterer existing code, not all are flywheel style algorithms, and approach vsync dejittering differently, but existing works of art: Open source VSYNC djitterer used by TestUFO and VSyncTester https://github.com/ …

---

Re: HOWTO: Possible Lagless VSYNC for Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC OFF

Posted on 2023-07-11, 21:37

by

by mdrejhon

in PC Emulation

in PC Emulation

I've been looking into how to implement some of this stuff in MartyPC using the library stack that I have, and I've been scratching my head a bit. Rust's standard timer is Instant: https://doc.rust-lang.org/std/time/struct.Instant.html It's back-end implementations are listed there. I am not sure …

---

Re: MartyPC

Posted on 2023-07-11, 06:27

by

by mdrejhon

in PC Emulation

in PC Emulation

On the related "video port idiosyncracies" talk such as RAMDAC/transceiver lag behaviors, there's another rabbit hole: GPU-vs-CPU clock slew effect (unrelated to CGA snow). So as another potential latency optimization for emulator authors, I added yet another lag-reducing tip/algorithm (much easier …

---

Re: HOWTO: Possible Lagless VSYNC for Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC OFF

Posted on 2023-07-11, 02:13

by

by mdrejhon

in PC Emulation

in PC Emulation

Updated my post with a few edits (potential CPU-vs-GPU clock sources). In short, stuff like RTDSC/QueryPerformanceCounter is a CPU clock source. - Good for best & consistently low latency performance on VRR (since VRR = software-initiated refresh cycles) And fixed-Hz VSYNC-monitoring (also …

---

Re: MartyPC

Posted on 2023-07-11, 01:26

by

by mdrejhon

in PC Emulation

in PC Emulation

Wouldn't that simply be filling a 1-byte FIFO buffer during each T3 state with the read/write value, cleared after T4 occurs (finishing cycle). Then simply replace CGA read ram when fetching for rendering with said value. Although UniPCemu still renders the EGA/VGA way, latching 4 bytes from all 4 …

---

Re: MartyPC

Posted on 2023-07-11, 00:57

by

by mdrejhon

in PC Emulation

in PC Emulation

We have snow! A nice good old fashioned Canadian snowstorm! (I'm Canadian, by the way. ATI and Matrox were Canadian companies by the way) I suppose my question was a fantastic one that triggered your decision to implement CGA snow -- because artifacts are "unexpectedly" a great debugging tool (much …

---

Re: HOWTO: Possible Lagless VSYNC for PC/DOS Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC

Posted on 2023-07-10, 02:20

by

by mdrejhon

in PC Emulation

in PC Emulation

This is because if the game is allowed to blindly have a frame rate exceeding VSYNC ON refresh rate (without a precision waitable swapchain in Full Screen Exclusive mode) For those curious about waitable swapchains, here's the Microsoft page: https://learn.microsoft.com/en-us/windows/uwp/gaming/ …

---

Re: MartyPC

Posted on 2023-07-08, 00:49

by

by mdrejhon

in PC Emulation

in PC Emulation

Question -- does your emulator have capability to optionally accurately emulate CGA snow? Would be great for visually debugging demos, so they don't create CGA snow on real machines! Also to show other people computer history, about how early computers glitched.

---

Re: HOWTO: Possible Lagless VSYNC for PC/DOS Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC

Posted on 2023-07-08, 00:20

by

by mdrejhon

in PC Emulation

in PC Emulation

You said you're using VRR, right? VRR is typically more common on higher-Hz monitors (e.g. 165Hz). So most of the lag reduction is caused by VRR's fast scanout. Those "70Hz" refresh cycles are scanned out onto your screen at max-Hz speed (e.g. 1/165sec for 165Hz monitor). 1/165sec is ~8.2ms less …

---

Re: HOWTO: Possible Lagless VSYNC for PC/DOS Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC

Posted on 2023-07-06, 00:44

by

by mdrejhon

in PC Emulation

in PC Emulation

The way surge-execution works is that an entire entire refresh cycle of ticks is run at full native machine performance. So multiple consecutive ticks are consecutively surge-executed, until a whole emulated refresh cycle is generated. If the machine uses 1ms ticks and the machine's refresh rate is …

---

Re: HOWTO: Possible Lagless VSYNC for PC/DOS Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC

Posted on 2023-07-05, 19:35

by

by mdrejhon

in PC Emulation

in PC Emulation

Thanks for your detailed explanation. Two questions: 1. How did the original g-sync work, did it use a buffer in the monitor or something, as monitors were expensive... Native G-SYNC historically worked pretty mostly the same way -- variable blanking interval. The difference is in the low frame …

---

Re: HOWTO: Possible Lagless VSYNC for PC/DOS Emulator Devs (implemented in WinUAE/etc), via beam-raced tearingless VSYNC

Posted on 2023-07-05, 19:17

by

by mdrejhon

in PC Emulation

in PC Emulation

Thanks for the detailed explanation, I've learned a lot from it. It all makes sense now, and it never occurred to me that on a say 240Hz display the latency would be effectively reduced to 1/4th of that of a 60Hz display... But yeah, it makes perfect sense; it couldn't work any other way. Thank you …

---