Icecream95
0499dbd4cf
Align Scan to be a multiple of four bytes
...
Some GPUs (e.g. r600) require the stride of vertex attributes to be a
multiple of four bytes, add two bytes of padding to the Scan struct to
meet this alignment requirement and reduce driver CPU overhead.
2022-08-29 16:32:25 +12:00
Thomas Harte
c7ab3d4075
Reduces cost of bookending video data.
2020-11-18 17:32:11 -05:00
Thomas Harte
5698aa6499
Corrects colour mapping and improves documentation for self.
2020-11-11 20:41:30 -05:00
Thomas Harte
a17d0e428f
Protects against some further uninitialised values.
2020-09-16 18:15:57 -04:00
Thomas Harte
637ec35d6a
Adds getters for standard colour-space conversion matrices.
...
These are just more details on the meaning of the colour spaces, so I think they belong here.
2020-08-11 19:58:57 -04:00
Thomas Harte
8374a5e579
Adds superficially correct compositeSampleLuminance8Phase8 function.
...
Thereby uncovering a minor error in my decoding of colour phase.
2020-08-10 21:33:59 -04:00
Thomas Harte
525233e10b
Ensures all input data types are parseable in Metal.
...
Though now I need to think a bit more about the best way to compose signal-type conversions, and whether output-type calculations (i.e. gamma, brightness) are applied.
2020-08-10 19:47:47 -04:00
Thomas Harte
c1dc42a094
Add comment on latent design aim.
...
In the hope that I don't forget why I did this.
2020-08-09 21:18:23 -04:00
Thomas Harte
bdcf266e45
Having learnt a bit more: eliminates Metal attribute tags, switches to more natural expression of structs.
...
Also thereby eliminates the need for a forced alignas(4) on various structs.
2020-08-08 17:27:32 -04:00
Thomas Harte
4c9418f59a
Guarantees alignof(4) on all GPU-bound structures.
...
Taken as given: Metal's requirement here is reasonable enough that it'll either be the same as other frameworks, or at least possibly help them down a fast path.
2020-08-07 21:18:08 -04:00
Thomas Harte
7431d56166
Ensures is_in_retrace is properly initialised.
2020-06-19 23:09:02 -04:00
Thomas Harte
530ff7471d
Adds a virtual destructor, given how these things might be held.
2020-06-14 21:14:51 -04:00
Thomas Harte
267006782f
Starts to add Qt target; resolves many build warnings.
2020-05-30 00:37:06 -04:00
Thomas Harte
512a52e88d
Increases const correctness, marks some additional constructors as constexpr, switches std::atomic construction style.
2020-05-20 23:34:26 -04:00
Thomas Harte
b41920990f
Moves submit step to end of line, rather than end of scan.
2020-02-04 22:15:20 -05:00
Thomas Harte
709c229cd7
Gets a bit more explicit with ScanTarget documentation.
2020-02-04 20:19:46 -05:00
Thomas Harte
9136917f00
Enables the Oric for 50/60Hz mode switching, inventing PAL60 for the purpose.
2020-01-23 22:14:02 -05:00
Thomas Harte
a5f285b4ce
Enhances reported data.
2020-01-22 22:01:17 -05:00
Thomas Harte
a71c5946f0
Ensures proper manipulation of scan_statuses, leading to the correct result out of a CRTMachine.
...
Possibly with the exception of the TMS, as I appear to have uncovered an unrelated issue there.
2020-01-21 22:28:25 -05:00
Thomas Harte
d97a073d1b
Adds the necessary routine for all machines to be able to respond to get_scan_status.
...
They all just as the CRT, as all are currently based on the CRT. Which doesn't currently know the total clock rate it would need to in order properly to scale the answer to the question. Further thought coming.
2020-01-20 21:45:10 -05:00
Thomas Harte
f6f9024631
Corrects Macintosh aspect ratio (and framing).
2019-06-13 18:41:38 -04:00
Thomas Harte
b75ad3def2
Updates the multimachine for the ScanTarget world.
2019-02-03 15:07:22 -05:00
Thomas Harte
a321ff3037
Adds some default values.
2019-01-25 20:21:24 -05:00
Thomas Harte
5d9521fcb9
Advances back to a semi-complete monochrome composite output.
...
i.e. composite phase and amplitude is ostensibly flowing to its new destination.
2019-01-11 22:02:15 -05:00
Thomas Harte
e9d9ff0da0
Enhances ScanTarget to provide additional timing information.
2019-01-05 23:09:17 -05:00
Thomas Harte
a25470ee41
Permits tweaking of PhaseLinkedLuminance8 sampling offset.
2018-11-29 16:29:28 -08:00
Thomas Harte
fd579a019b
Introduces a new scan source data type, motivated by the reasoning used by the Oric.
...
Specifically: it'll allow PCM sampling of the potentially arbitrary composite generation logic of various machines.
2018-11-28 20:40:22 -08:00
Thomas Harte
64465f97b6
Starts towards reintroducing the proper mechanisms for selecting a display type at runtime.
2018-11-28 17:53:33 -08:00
Thomas Harte
a66a20f7fe
Manages to get a brilliant white out of the new pipeline.
2018-11-23 22:54:52 -05:00
Thomas Harte
d4ac79b0af
Attempts to introduce a full-on processing pipeline, in theory putting me two shaders away from completion.
...
Well, subject to finding the last flashing bug and updating the multimachine, anyway.
2018-11-23 22:34:38 -05:00
Thomas Harte
b9aca39eb0
Reintroduces Vic-20 output.
...
Resolving errors in shader generation while I'm here.
2018-11-22 22:43:42 -05:00
Thomas Harte
a697a2e4f6
Attempts to complete all input processing — an RGB, S-Video or composite input buffer is now produced.
...
... for all input data types.
2018-11-22 17:20:31 -05:00
Thomas Harte
396cf72029
Renames OutputType as DisplayType and promotes it to a scan target modal.
2018-11-22 14:36:46 -05:00
Thomas Harte
8f6664f0d7
Starts towards picking an input shader based on data type and pipeline.
2018-11-15 21:02:46 -05:00
Thomas Harte
8a699b6072
Kills setup_output definitively, saving some indirection. set_scan_target takes its place.
2018-11-14 21:52:57 -05:00
Thomas Harte
02f9cada43
Communicates the colour subcarrier frequency, and uses it to pick a buffer width.
2018-11-13 18:33:44 -05:00
Thomas Harte
6d277fecd5
Makes ScanTarget a little more communicative and orthogonal.
2018-11-10 19:52:57 -05:00
Thomas Harte
55da1e9c0f
Simplifies semantics a little and starts accepting a single buffer of pixel data.
2018-11-06 22:23:38 -05:00
Thomas Harte
014da41471
Ensures scan positions are communicated with a specified range, and switches manner of pixel clock communication.
2018-11-04 21:06:25 -05:00
Thomas Harte
f6562de325
Possibly adds enough for the Electron and ZX80 to start outputting dummy lines.
...
Let's see!
2018-11-03 23:40:39 -04:00
Thomas Harte
b40211d2c0
Starts to bend 'CRTMachine' to a world farther from owning the GPU relationship.
2018-11-03 21:54:25 -04:00
