1 April 2025: DUGL 1.01

^^^^^^^^^^^^^^^^^^^^^^^

Global:

- Changes CPU target support to SSE4.1 instead of SSE4.2

- Increase Max Target DgSurf Height from 4096 to 8192 (unlimited width)

- Changes Min Target DgSurf size to (w,h) (4,1)

- DgQuit increased compatiblity

- Add experimental DgWaitVSync (could reduce flickering, but not effectively reduce fps to screen refresh)

Graphics:

- Poly16: reject now single hline polygones; faster detection of (in, out or clipped); Optimization of some FILL funcs;

Completely rewrite UV mapping, (left,write borders) hlines to be faster and more accurate (clipped/unclipped), this allowed multi-cores rendering without

any visible change against single core; Reintroduce special case corners poly out detection, with vertical size reduction of clipped poly;

- RePoly16: Bug fixes, optimization/size reduction; stop handling 'N' Status (completely out) poly, instead REPOLY16 macro added, to avoid useless call;

- Line16,line16: size/speed optimization.

- Bar16, bar16: rewrite to use faster InBar16 instead of Poly16.

- PutXXXSurfXXX16: Performance/size optimization.

- DgSurfCPutPixel16: fix crash.

Render Cores:

- Add LastPolyStatus to cores context

- Implement the macro REPOLY16 in separate function for each core, and redirect RePoly16 to it.

GUI:

- Bug fix unaligned 16bytes local DgSurf causing occasional crashes

Timer:

- Small rewrite in attempt to smoother/more precise synching.

Samples:

- Shadow: Add Multi-core resizing; Improve Multi-Core rendering up to 4 cores; Add parametric High quality rendering from 1.1 to 3.0 ratio;

Add background panoramic sky background; Improve keyboard shortcuts; Enable double-buffer; optimize polygones sorting;

Reduce unecessera big memory geometry allocation; Fix long standing camera concurrency between rendering/control threads, bug fixes, speed improvement;

Add capability to show/hide info&help.

- Sprites: Add capability of switching/profiling any of the 6 possible Put functions (PUT, MASK PUT, COL BLND PUT, MASK COL BLND PUT, TRANSP PUT and MASK TRANSP PUT

- HelloWorld, Forest: Rework event handling and rendering loop and DWorker

19 April 2023: DUGL 1.0

^^^^^^^^^^^^^^^^^^^^^^^

All assembly functions rewritten to Target SSE4.1+ x86 CPU with much higher performance than the old Pentium MMX

Render DgSurf target could be up to 4096 pixel Height (unlimited width)

RePoly function to rerender last rendered poly, being approx 20% faster allow to reapply effects to same poly, as clothes, terrain grass, shadow casting ...

Render Cores with Up to 4 render cores, allowing to split Rendering Up to 4 DWorker(s) improving performance up to 300% against single threaded rendering

Blur function (that allow to make high quality rendering) rewritten to allow multi-DWorker(s) blurring

Sprite blitting through PutXXSurfXX function take care of both source and destination DgSurf Views

Sound Support with mixer (based mostly on SDL)

Thread Support throw DWorker Concept

Math3D Library with big set of functions optimized up to SSE4.1, including suffix NT functions that pollute less CPU cache (using MOVNTDQ ...)

Thanks to usage of LibSDL, DUGL is now X86 multi-plateforme library, binaries are available for both Windows/Linux

and more ...

What's new against last DUGL 1.0 beta 5

Graphics:

Increase MAX vertical size of render DgSurf to 4096 from 2048

Add RePoly16 function with capability to rerender last Poly16 polygone, but with a different Polygone Type/color, Source DgSurf, texture U,V mapping ... benchmark show that it's about 20% faster than regular Poly16 call.

Add BlndCol16 function to mix two RGB16 colors with 32 levels of mixing.

Bug fixes on Poly16:POLY16_MASK_TEXT_TRANS and PutMaskSurfBlnd16

Window:

Add capability to detect/handle window close request event

Add DgSetWindowIcone

Fix bug of mouse position/cursor visiblity at application start-up

Audio:

Expose DSoundState struct

Add DSoundState events callback

Add functions ReplaceDSound, IsPlayingDSound, SetEventsDSound..

Several improvements and bug fixes ...

Image:

Fix bug loading paletted/masked png file

Rendering Cores:

Adds 4 rendering cores

Add GetDGCORE function which return a filled new struct DGCORE with pointer to core index functions and global vars.

Container:

Adds DFileBuffer functions: GetBytesDFileBuffer, FseekDFileBuffer, RewindDFileBuffer

Math3d:

Add functions GetPlaneDVEC4, IntersectRayPlane, IntersectRayPlaneRes, RayProjectDVEC4Res, SubNormalizeDVEC4, SubNormalizeDVEC4Res, CrossNormalizeDVEC4, DotNormalizeDVEC4, NormalizeDVEC4Res

EqualDVEC4 now compare all the 4 components of the DVEC4.

several function optimized using SSE41 DPPS instruction.

several performance tweaks.

several functions reimplemented to return pointer to result when possible.

Tools:

Add keybmap tool, the proprietary DUGL keyboard map(or layout) editor.

Samples:

Add Sprites sample: benchmark of sprites blitting using 1, 2 or 4 cores

Add Shadow sample: a simple unpotimized "3d engine" displaying a ground with a tree that project its Shadow on the ground.