hugova
0adbbaf12c
Add memcpy and code skeleton for text.s
2025-03-28 12:05:11 +01:00
hugova
b3887a8e0d
Rename memory operation and create files for later text drawing stuff
2025-03-27 21:12:40 +01:00
hugova
801cd55541
Improved performance of line_down.s before loop.
2025-03-26 22:00:09 +01:00
hugova
89c9bc4129
Add self modifying code to draw_line for 2% performance improvments
2025-03-26 12:14:55 +01:00
hugova
ff64a9a9b0
Add missing performance improvments to line_* from line_down.s
2025-03-25 23:28:53 +01:00
hugova
20b24a299a
Clean up memory_rec.s
2025-03-25 21:59:44 +01:00
hugova
41b8fec7a8
Performance improved mult_16 without changing behaviour
2025-03-25 21:40:40 +01:00
hugova
7a5585d2c4
Add better code coments
2025-03-25 21:21:39 +01:00
hugova
842033a014
Add performance optimisiation for line_up_inv.s
2025-03-25 19:46:03 +01:00
hugova
d5c15bd0c2
Code cleanup
2025-03-25 19:38:51 +01:00
hugova
0df0e666ad
Implement earlier perfromance improvments to line_up.s
2025-03-25 19:32:51 +01:00
hugova
4497026d95
Rewrite start condition of line_down so that it start on exaktly the right pixel
2025-03-25 19:20:23 +01:00
hugova
f185c80a16
Add code skeleton for the binary-tree-api:s unittest
2025-03-24 20:16:52 +01:00
hugova
05b0ff9762
Add performance improvments to line_down_inv.s
2025-03-24 19:55:00 +01:00
hugova
9b8e3fe33f
Removed 1 instruction from draw_line_down.s 6% performance boost
2025-03-24 18:25:42 +01:00
hugova
3ce76c02a8
Rewrote 4 branch instructions to 2. 2% performance increase
2025-03-24 18:00:17 +01:00
hugova
f583423bb6
Switched CLC ROR to LSR. 2% performance improvment
2025-03-24 17:45:22 +01:00
hugova
bc59ac63b1
Fix performance by using pointer tricks for ADD. 10% faster line drawing
2025-03-24 17:08:09 +01:00
hugova
de2d597612
Change of starting condition for potential testing purposes of another person
2025-03-09 23:53:33 +01:00
hugova
a617d91183
Change 4 spaces indenting to 8 and add string for emacs to understand indenting rules
2025-03-09 23:43:59 +01:00
hugova
0797e69106
Add all optimisations done on line_down to line_down_inv. DIx a problem that the pixel was drawin a bit to late in the loop on line_down.
2025-03-06 01:50:37 +01:00
hugova
ecc138f2bc
Redused number of instructions but to small a diffrence to messure with current test
2025-03-06 00:59:29 +01:00
hugova
9694d8d55b
Redused paint_pixel to 1 call in draw line
2025-03-05 22:42:00 +01:00
hugova
76f29f6b81
Performance improvment in test by 0.1s. Mady by better arithmatic for *40 in draw_pixel and fast_unsafe parameter for most macros
2025-03-05 21:02:54 +01:00
hugova
982b6c7ea6
A fix for the timer to work by making it start with value 1. It do not increment from 0!
2025-03-05 01:27:37 +01:00
hugova
89c2b42bfe
Reorginisation of line-drawing code. Also added a macro to benchmark line drawing logic. Addition of line_test_time to benchmark line drawing time (Do not work yet)
2025-03-04 23:01:57 +01:00
hugova
494e9dc5f9
A fix of line_up_inv such that X_pos < X_end like all other line_* implementations.
2025-03-04 21:00:21 +01:00
hugova
4d12f20a6e
A Fix for D that gets the wrong initial value and the branch logic in line_* checked for the wrong value.
...
A new proc, line.s that can chose whitch line_* to use
A small test program in source.s
Wrote a smal test program in source.s (all looks good exept line_up_inv)
2025-03-03 22:08:00 +01:00
hugova
1524c837ca
A optimisation of line_up_inv.s by moving some logic out of loop
2025-03-02 23:14:51 +01:00
hugova
f3712eca9f
Move logic outside loop for performance reasons.
...
Same as earlier commit but for line_down_inv.s, line_up.s
Comment: Should do the same for line_up_inv.s
2025-03-02 18:29:58 +01:00
hugova
ee69828ec7
Optimisation by moving logic out of loop for faster execution.
...
This should be implemented on the other line_* files
2025-03-02 18:05:20 +01:00
hugova
0940d9f9f9
1. A fix for a bug where D gets the wrong initial value (2 bytes where flipped)
...
2. Implemented line drawing for the other 3 cases
2025-03-02 16:51:55 +01:00
hugova
4f396fd740
A fix to a bug, to make line_down draw lines with small slopes.
...
I needed to use 16bit logic to ashive this.
Also moved out dx*2, dy*2 calculation outside of loop for better performance.
2025-03-01 23:57:15 +01:00
hugova
9c86b570ed
Will make the indenting better. Also it will properly use dubble ; for comment blocks
2025-02-28 18:22:07 +01:00
hugova
864b0e659b
Add hugo code
2025-02-27 22:04:43 +01:00
