2 * Copyright © 2014 Timothy Arceri
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
24 * Timothy Arceri <t_arceri@yahoo.com.au>
29 #include "main/sse_minmax.h"
30 #include <smmintrin.h>
34 #if !defined(__x86_64__)
35 __attribute__((force_align_arg_pointer
))
37 _mesa_uint_array_min_max(const unsigned *ui_indices
, unsigned *min_index
,
38 unsigned *max_index
, const unsigned count
)
41 unsigned min_ui
= ~0U;
43 unsigned aligned_count
= count
;
45 /* handle the first few values without SSE until the pointer is aligned */
46 while (((uintptr_t)ui_indices
& 15) && aligned_count
) {
47 if (*ui_indices
> max_ui
)
49 if (*ui_indices
< min_ui
)
56 /* TODO: The actual threshold for SSE begin useful may be higher than 8.
57 * Some careful microbenchmarks and measurement are required to
58 * find the actual tipping point.
60 if (aligned_count
>= 8) {
61 unsigned max_arr
[4] __attribute__ ((aligned (16)));
62 unsigned min_arr
[4] __attribute__ ((aligned (16)));
64 __m128i max_ui4
= _mm_setzero_si128();
65 __m128i min_ui4
= _mm_set1_epi32(~0U);
67 __m128i
*ui_indices_ptr
;
69 vec_count
= aligned_count
& ~0x3;
70 ui_indices_ptr
= (__m128i
*)ui_indices
;
71 for (i
= 0; i
< vec_count
/ 4; i
++) {
72 ui_indices4
= _mm_load_si128(&ui_indices_ptr
[i
]);
73 max_ui4
= _mm_max_epu32(ui_indices4
, max_ui4
);
74 min_ui4
= _mm_min_epu32(ui_indices4
, min_ui4
);
77 _mm_store_si128((__m128i
*)max_arr
, max_ui4
);
78 _mm_store_si128((__m128i
*)min_arr
, min_ui4
);
80 for (i
= 0; i
< 4; i
++) {
81 if (max_arr
[i
] > max_ui
)
83 if (min_arr
[i
] < min_ui
)
89 for (; i
< aligned_count
; i
++) {
90 if (ui_indices
[i
] > max_ui
)
91 max_ui
= ui_indices
[i
];
92 if (ui_indices
[i
] < min_ui
)
93 min_ui
= ui_indices
[i
];