-
Notifications
You must be signed in to change notification settings - Fork 35
/
ASMMatrixAbsOperations.pas
373 lines (272 loc) · 9.18 KB
/
ASMMatrixAbsOperations.pas
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
// ###################################################################
// #### This file is part of the mathematics library project, and is
// #### offered under the licence agreement described on
// #### http://www.mrsoft.org/
// ####
// #### Copyright:(c) 2017, Michael R. . All rights reserved.
// ####
// #### Unless required by applicable law or agreed to in writing, software
// #### distributed under the License is distributed on an "AS IS" BASIS,
// #### WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// #### See the License for the specific language governing permissions and
// #### limitations under the License.
// ###################################################################
unit ASMMatrixAbsOperations;
// #####################################################
// #### Abs opertaion applied to every element in a matrix
// #####################################################
interface
{$I 'mrMath_CPU.inc'}
{$IFNDEF x64}
procedure ASMMatrixAbsAlignedEvenW(Dest : PDouble; const LineWidth, Width, Height : NativeInt); {$IFDEF FPC} assembler; {$ELSE} register; {$ENDIF}
procedure ASMMatrixAbsUnAlignedEvenW(Dest : PDouble; const LineWidth, Width, Height : NativeInt); {$IFDEF FPC} assembler; {$ELSE} register; {$ENDIF}
procedure ASMMatrixAbsAlignedOddW(Dest : PDouble; const LineWidth, Width, Height : NativeInt); {$IFDEF FPC} assembler; {$ELSE} register; {$ENDIF}
procedure ASMMatrixAbsUnAlignedOddW(Dest : PDouble; const LineWidth, Width, Height : NativeInt); {$IFDEF FPC} assembler; {$ELSE} register; {$ENDIF}
{$ENDIF}
implementation
{$IFNDEF x64}
const cLocSignBits : Array[0..1] of int64 = ($7FFFFFFFFFFFFFFF, $7FFFFFFFFFFFFFFF);
procedure ASMMatrixAbsAlignedEvenW(Dest : PDouble; const LineWidth, Width, Height : NativeInt);
// eax=dest, edx=LineWidth, width=ecx
asm
push ebx;
// width* -8
imul ecx, -8;
// helper registers for the mt1, mt2 and dest pointers
sub eax, ecx;
movupd xmm0, cLocSignBits;
// for y := 0 to height - 1:
@@addforyloop:
// for x := 0 to w - 1;
// prepare for reverse loop
mov ebx, ecx;
@addforxloop:
add ebx, 128;
jg @loopEnd;
// Abs:
movapd xmm4, [eax + ebx - 128];
Andpd xmm4, xmm0;
movapd [eax + ebx - 128], xmm4;
movapd xmm1, [eax + ebx - 112];
andpd xmm1, xmm0;
movapd [eax + ebx - 112], xmm1;
movapd xmm2, [eax + ebx - 96];
andpd xmm2, xmm0;
movapd [eax + ebx - 96], xmm2;
movapd xmm3, [eax + ebx - 80];
andpd xmm3, xmm0;
movapd [eax + ebx - 80], xmm3;
movapd xmm4, [eax + ebx - 64];
andpd xmm4, xmm0;
movapd [eax + ebx - 64], xmm4;
movapd xmm5, [eax + ebx - 48];
andpd xmm5, xmm0;
movapd [eax + ebx - 48], xmm5;
movapd xmm6, [eax + ebx - 32];
andpd xmm6, xmm0;
movapd [eax + ebx - 32], xmm6;
movapd xmm7, [eax + ebx - 16];
andpd xmm7, xmm0;
movapd [eax + ebx - 16], xmm7;
jmp @addforxloop
@loopEnd:
sub ebx, 128;
jz @nextLine;
@addforxloop2:
movapd xmm1, [eax + ebx];
andpd xmm1, xmm0;
movapd [eax + ebx], xmm1;
add ebx, 16;
jnz @addforxloop2;
@nextLine:
// next line:
add eax, edx;
// loop y end
dec height;
jnz @@addforyloop;
pop ebx;
end;
procedure ASMMatrixAbsUnAlignedEvenW(Dest : PDouble; const LineWidth, Width, Height : NativeInt);
asm
push ebx;
imul ecx, -8;
// helper registers for the mt1, mt2 and dest pointers
sub eax, ecx;
movupd xmm0, cLocSignBits;
// for y := 0 to height - 1:
@@addforyloop:
// for x := 0 to w - 1;
// prepare for reverse loop
mov ebx, ecx;
@addforxloop:
add ebx, 128;
jg @loopEnd;
// Abs:
movupd xmm4, [eax + ebx - 128];
Andpd xmm4, xmm0;
movupd [eax + ebx - 128], xmm4;
movupd xmm1, [eax + ebx - 112];
andpd xmm1, xmm0;
movupd [eax + ebx - 112], xmm1;
movupd xmm2, [eax + ebx - 96];
andpd xmm2, xmm0;
movupd [eax + ebx - 96], xmm2;
movupd xmm3, [eax + ebx - 80];
andpd xmm3, xmm0;
movupd [eax + ebx - 80], xmm3;
movupd xmm4, [eax + ebx - 64];
andpd xmm4, xmm0;
movupd [eax + ebx - 64], xmm4;
movupd xmm5, [eax + ebx - 48];
andpd xmm5, xmm0;
movupd [eax + ebx - 48], xmm5;
movupd xmm6, [eax + ebx - 32];
andpd xmm6, xmm0;
movupd [eax + ebx - 32], xmm6;
movupd xmm7, [eax + ebx - 16];
andpd xmm7, xmm0;
movupd [eax + ebx - 16], xmm7;
jmp @addforxloop
@loopEnd:
sub ebx, 128;
jz @nextLine;
@addforxloop2:
movupd xmm1, [eax + ebx];
andpd xmm1, xmm0;
movupd [eax + ebx], xmm1;
add ebx, 16;
jnz @addforxloop2;
@nextLine:
// next line:
add eax, edx;
// loop y end
dec Height;
jnz @@addforyloop;
pop ebx;
end;
procedure ASMMatrixAbsAlignedOddW(Dest : PDouble; const LineWidth, Width, Height : NativeInt);
asm
push ebx;
dec ecx;
imul ecx, -8;
sub eax, ecx;
movupd xmm0, cLocSignBits;
// for y := 0 to height - 1:
@@addforyloop:
// for x := 0 to w - 1;
// prepare for reverse loop
mov ebx, ecx;
@addforxloop:
add ebx, 128;
jg @loopEnd;
// Abs:
movapd xmm4, [eax + ebx - 128];
Andpd xmm4, xmm0;
movapd [eax + ebx - 128], xmm4;
movapd xmm1, [eax + ebx - 112];
andpd xmm1, xmm0;
movapd [eax + ebx - 112], xmm1;
movapd xmm2, [eax + ebx - 96];
andpd xmm2, xmm0;
movapd [eax + ebx - 96], xmm2;
movapd xmm3, [eax + ebx - 80];
andpd xmm3, xmm0;
movapd [eax + ebx - 80], xmm3;
movapd xmm4, [eax + ebx - 64];
andpd xmm4, xmm0;
movapd [eax + ebx - 64], xmm4;
movapd xmm5, [eax + ebx - 48];
andpd xmm5, xmm0;
movapd [eax + ebx - 48], xmm5;
movapd xmm6, [eax + ebx - 32];
andpd xmm6, xmm0;
movapd [eax + ebx - 32], xmm6;
movapd xmm7, [eax + ebx - 16];
andpd xmm7, xmm0;
movapd [eax + ebx - 16], xmm7;
jmp @addforxloop
@loopEnd:
sub ebx, 128;
jz @nextLine;
@addforxloop2:
movapd xmm1, [eax + ebx - 16];
andpd xmm1, xmm0;
movapd [eax + ebx], xmm1;
add ebx, 16;
jnz @addforxloop2;
@nextLine:
// special care of the last column:
movsd xmm1, [eax];
andpd xmm1, xmm0;
movsd [eax], xmm1;
// next line:
add eax, edx;
// loop y end
dec Height;
jnz @@addforyloop;
pop ebx;
end;
procedure ASMMatrixAbsUnAlignedOddW(Dest : PDouble; const LineWidth, Width, Height : NativeInt);
asm
push ebx;
dec ecx;
imul ecx, -8;
sub eax, ecx;
movupd xmm0, cLocSignBits;
// for y := 0 to height - 1:
@@addforyloop:
// for x := 0 to w - 1;
// prepare for reverse loop
mov ebx, ecx;
@addforxloop:
add ebx, 128;
jg @loopEnd;
// Abs:
movupd xmm4, [eax + ebx - 128];
Andpd xmm4, xmm0;
movupd [eax + ebx - 128], xmm4;
movupd xmm1, [eax + ebx - 112];
andpd xmm1, xmm0;
movupd [eax + ebx - 112], xmm1;
movupd xmm2, [eax + ebx - 96];
andpd xmm2, xmm0;
movupd [eax + ebx - 96], xmm2;
movupd xmm3, [eax + ebx - 80];
andpd xmm3, xmm0;
movupd [eax + ebx - 80], xmm3;
movupd xmm4, [eax + ebx - 64];
andpd xmm4, xmm0;
movupd [eax + ebx - 64], xmm4;
movupd xmm5, [eax + ebx - 48];
andpd xmm5, xmm0;
movupd [eax + ebx - 48], xmm5;
movupd xmm6, [eax + ebx - 32];
andpd xmm6, xmm0;
movupd [eax + ebx - 32], xmm6;
movupd xmm7, [eax + ebx - 16];
andpd xmm7, xmm0;
movupd [eax + ebx - 16], xmm7;
jmp @addforxloop
@loopEnd:
sub ebx, 128;
jz @nextLine;
@addforxloop2:
movupd xmm1, [eax + ebx];
andpd xmm1, xmm0;
movupd [eax + ebx], xmm1;
add ebx, 16;
jnz @addforxloop2;
@nextLine:
// special care of the last column:
movsd xmm1, [eax];
andpd xmm1, xmm0;
movsd [eax], xmm1;
// next line:
add eax, edx;
// loop y end
dec Height;
jnz @@addforyloop;
pop ebx;
end;
{$ENDIF}
end.