summaryrefslogtreecommitdiffstatshomepage
path: root/src/osd/eminline.h
blob: bb0aec63468075062e5ec8da7b7abc41172a6060 (plain) (blame)
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
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
// license:BSD-3-Clause
// copyright-holders:Aaron Giles
/***************************************************************************

    eminline.h

    Definitions for inline functions that can be overridden by OSD-
    specific code.

***************************************************************************/

#ifndef MAME_OSD_EMINLINE_H
#define MAME_OSD_EMINLINE_H

#pragma once

#include "osdcomm.h"
#include "osdcore.h"

#if !defined(MAME_NOASM)

#if defined(__GNUC__)

#if defined(__i386__) || defined(__x86_64__)
#include "eigccx86.h"
#elif defined(__ppc__) || defined (__PPC__) || defined(__ppc64__) || defined(__PPC64__)
#include "eigccppc.h"
#else
#error "no matching assembler implementations found - please compile with NOASM=1"
#endif

#elif defined(_MSC_VER)

#if (defined(_M_IX86) || defined(_M_X64))
#include "eivcx86.h"
#endif

#include "eivc.h"

#else

#error "no matching assembler implementations found - please compile with NOASM=1"

#endif

#endif // !defined(MAME_NOASM)


/***************************************************************************
    INLINE MATH FUNCTIONS
***************************************************************************/

/*-------------------------------------------------
    mul_32x32 - perform a signed 32 bit x 32 bit
    multiply and return the full 64 bit result
-------------------------------------------------*/

#ifndef mul_32x32
inline int64_t mul_32x32(int32_t a, int32_t b)
{
	return int64_t(a) * int64_t(b);
}
#endif


/*-------------------------------------------------
    mulu_32x32 - perform an unsigned 32 bit x
    32 bit multiply and return the full 64 bit
    result
-------------------------------------------------*/

#ifndef mulu_32x32
inline uint64_t mulu_32x32(uint32_t a, uint32_t b)
{
	return uint64_t(a) * uint64_t(b);
}
#endif


/*-------------------------------------------------
    mul_32x32_hi - perform a signed 32 bit x 32 bit
    multiply and return the upper 32 bits of the
    result
-------------------------------------------------*/

#ifndef mul_32x32_hi
inline int32_t mul_32x32_hi(int32_t a, int32_t b)
{
	return uint32_t((int64_t(a) * int64_t(b)) >> 32);
}
#endif


/*-------------------------------------------------
    mulu_32x32_hi - perform an unsigned 32 bit x
    32 bit multiply and return the upper 32 bits
    of the result
-------------------------------------------------*/

#ifndef mulu_32x32_hi
inline uint32_t mulu_32x32_hi(uint32_t a, uint32_t b)
{
	return uint32_t((uint64_t(a) * uint64_t(b)) >> 32);
}
#endif


/*-------------------------------------------------
    mul_32x32_shift - perform a signed 32 bit x
    32 bit multiply and shift the result by the
    given number of bits before truncating the
    result to 32 bits
-------------------------------------------------*/

#ifndef mul_32x32_shift
inline int32_t mul_32x32_shift(int32_t a, int32_t b, uint8_t shift)
{
	return int32_t((int64_t(a) * int64_t(b)) >> shift);
}
#endif


/*-------------------------------------------------
    mulu_32x32_shift - perform an unsigned 32 bit x
    32 bit multiply and shift the result by the
    given number of bits before truncating the
    result to 32 bits
-------------------------------------------------*/

#ifndef mulu_32x32_shift
inline uint32_t mulu_32x32_shift(uint32_t a, uint32_t b, uint8_t shift)
{
	return uint32_t((uint64_t(a) * uint64_t(b)) >> shift);
}
#endif


/*-------------------------------------------------
    div_64x32 - perform a signed 64 bit x 32 bit
    divide and return the 32 bit quotient
-------------------------------------------------*/

#ifndef div_64x32
inline int32_t div_64x32(int64_t a, int32_t b)
{
	return a / int64_t(b);
}
#endif


/*-------------------------------------------------
    divu_64x32 - perform an unsigned 64 bit x 32 bit
    divide and return the 32 bit quotient
-------------------------------------------------*/

#ifndef divu_64x32
inline uint32_t divu_64x32(uint64_t a, uint32_t b)
{
	return a / uint64_t(b);
}
#endif


/*-------------------------------------------------
    div_64x32_rem - perform a signed 64 bit x 32
    bit divide and return the 32 bit quotient and
    32 bit remainder
-------------------------------------------------*/

#ifndef div_64x32_rem
inline int32_t div_64x32_rem(int64_t a, int32_t b, int32_t *remainder)
{
	int32_t const res = div_64x32(a, b);
	*remainder = a - (int64_t(b) * res);
	return res;
}
#endif


/*-------------------------------------------------
    divu_64x32_rem - perform an unsigned 64 bit x
    32 bit divide and return the 32 bit quotient
    and 32 bit remainder
-------------------------------------------------*/

#ifndef divu_64x32_rem
inline uint32_t divu_64x32_rem(uint64_t a, uint32_t b, uint32_t *remainder)
{
	uint32_t const res = divu_64x32(a, b);
	*remainder = a - (uint64_t(b) * res);
	return res;
}
#endif


/*-------------------------------------------------
    div_32x32_shift - perform a signed divide of
    two 32 bit values, shifting the first before
    division, and returning the 32 bit quotient
-------------------------------------------------*/

#ifndef div_32x32_shift
inline int32_t div_32x32_shift(int32_t a, int32_t b, uint8_t shift)
{
	return (int64_t(a) << shift) / int64_t(b);
}
#endif


/*-------------------------------------------------
    divu_32x32_shift - perform an unsigned divide of
    two 32 bit values, shifting the first before
    division, and returning the 32 bit quotient
-------------------------------------------------*/

#ifndef divu_32x32_shift
inline uint32_t divu_32x32_shift(uint32_t a, uint32_t b, uint8_t shift)
{
	return (uint64_t(a) << shift) / uint64_t(b);
}
#endif


/*-------------------------------------------------
    mod_64x32 - perform a signed 64 bit x 32 bit
    divide and return the 32 bit remainder
-------------------------------------------------*/

#ifndef mod_64x32
inline int32_t mod_64x32(int64_t a, int32_t b)
{
	return a - (b * div_64x32(a, b));
}
#endif


/*-------------------------------------------------
    modu_64x32 - perform an unsigned 64 bit x 32 bit
    divide and return the 32 bit remainder
-------------------------------------------------*/

#ifndef modu_64x32
inline uint32_t modu_64x32(uint64_t a, uint32_t b)
{
	return a - (b * divu_64x32(a, b));
}
#endif


/*-------------------------------------------------
    recip_approx - compute an approximate floating
    point reciprocal
-------------------------------------------------*/

#ifndef recip_approx
inline float recip_approx(float value)
{
	return 1.0f / value;
}
#endif



/***************************************************************************
    INLINE BIT MANIPULATION FUNCTIONS
***************************************************************************/

/*-------------------------------------------------
    count_leading_zeros - return the number of
    leading zero bits in a 32-bit value
-------------------------------------------------*/

#ifndef count_leading_zeros
inline uint8_t count_leading_zeros(uint32_t val)
{
	if (!val) return 32U;
	uint8_t count;
	for (count = 0; int32_t(val) >= 0; count++) val <<= 1;
	return count;
}
#endif


/*-------------------------------------------------
    count_leading_ones - return the number of
    leading one bits in a 32-bit value
-------------------------------------------------*/

#ifndef count_leading_ones
inline uint8_t count_leading_ones(uint32_t val)
{
	uint8_t count;
	for (count = 0; int32_t(val) < 0; count++) val <<= 1;
	return count;
}
#endif


/*-------------------------------------------------
    population_count_32 - return the number of
    one bits in a 32-bit value
-------------------------------------------------*/

#ifndef population_count_32
#if defined(__NetBSD__)
#define population_count_32 popcount32
#else
inline unsigned population_count_32(uint32_t val)
{
#if defined(__GNUC__)
	// uses CPU feature if available, otherwise falls back to implementation similar to what follows
	static_assert(sizeof(val) == sizeof(unsigned), "expected 32-bit unsigned int");
	return unsigned(__builtin_popcount(static_cast<unsigned>(val)));
#else
	// optimal Hamming weight assuing fast 32*32->32
	constexpr uint32_t m1(0x55555555);
	constexpr uint32_t m2(0x33333333);
	constexpr uint32_t m4(0x0f0f0f0f);
	constexpr uint32_t h01(0x01010101);
	val -= (val >> 1) & m1;
	val = (val & m2) + ((val >> 2) & m2);
	val = (val + (val >> 4)) & m4;
	return unsigned((val * h01) >> 24);
#endif
}
#endif
#endif


/*-------------------------------------------------
    population_count_64 - return the number of
    one bits in a 64-bit value
-------------------------------------------------*/

#ifndef population_count_64
#if defined(__NetBSD__)
#define population_count_64 popcount64
#else
inline unsigned population_count_64(uint64_t val)
{
#if defined(__GNUC__)
	// uses CPU feature if available, otherwise falls back to implementation similar to what follows
	static_assert(sizeof(val) == sizeof(unsigned long long), "expected 64-bit unsigned long long int");
	return unsigned(__builtin_popcountll(static_cast<unsigned long long>(val)));
#else
	// guess that architectures with 64-bit pointers have 64-bit multiplier
	if (sizeof(void *) >= sizeof(uint64_t))
	{
		// optimal Hamming weight assuming fast 64*64->64
		constexpr uint64_t m1(0x5555555555555555);
		constexpr uint64_t m2(0x3333333333333333);
		constexpr uint64_t m4(0x0f0f0f0f0f0f0f0f);
		constexpr uint64_t h01(0x0101010101010101);
		val -= (val >> 1) & m1;
		val = (val & m2) + ((val >> 2) & m2);
		val = (val + (val >> 4)) & m4;
		return unsigned((val * h01) >> 56);
	}
	else
	{
		// fall back to two 32-bit operations to avoid slow multiply
		return population_count_32(uint32_t(val)) + population_count_32(uint32_t(val >> 32));
	}
#endif
}
#endif
#endif


/***************************************************************************
    INLINE TIMING FUNCTIONS
***************************************************************************/

/*-------------------------------------------------
    get_profile_ticks - return a tick counter
    from the processor that can be used for
    profiling. It does not need to run at any
    particular rate.
-------------------------------------------------*/

#ifndef get_profile_ticks
inline int64_t get_profile_ticks()
{
	return osd_ticks();
}
#endif

#endif // MAME_OSD_EMINLINE_H