Rev 834 | Rev 1657 | Go to most recent revision | Only display areas with differences | Regard whitespace | Details | Blame | Last modification | View Log | RSS feed
Rev 834 | Rev 1031 | ||
---|---|---|---|
1 | /* |
1 | /* |
2 | * Copyright (C) 2005 Josef Cejka |
2 | * Copyright (C) 2005 Josef Cejka |
3 | * All rights reserved. |
3 | * All rights reserved. |
4 | * |
4 | * |
5 | * Redistribution and use in source and binary forms, with or without |
5 | * Redistribution and use in source and binary forms, with or without |
6 | * modification, are permitted provided that the following conditions |
6 | * modification, are permitted provided that the following conditions |
7 | * are met: |
7 | * are met: |
8 | * |
8 | * |
9 | * - Redistributions of source code must retain the above copyright |
9 | * - Redistributions of source code must retain the above copyright |
10 | * notice, this list of conditions and the following disclaimer. |
10 | * notice, this list of conditions and the following disclaimer. |
11 | * - Redistributions in binary form must reproduce the above copyright |
11 | * - Redistributions in binary form must reproduce the above copyright |
12 | * notice, this list of conditions and the following disclaimer in the |
12 | * notice, this list of conditions and the following disclaimer in the |
13 | * documentation and/or other materials provided with the distribution. |
13 | * documentation and/or other materials provided with the distribution. |
14 | * - The name of the author may not be used to endorse or promote products |
14 | * - The name of the author may not be used to endorse or promote products |
15 | * derived from this software without specific prior written permission. |
15 | * derived from this software without specific prior written permission. |
16 | * |
16 | * |
17 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR |
17 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR |
18 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES |
18 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES |
19 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. |
19 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. |
20 | * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, |
20 | * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, |
21 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT |
21 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT |
22 | * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
22 | * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
23 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
23 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
24 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
24 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
25 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF |
25 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF |
26 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
26 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
27 | */ |
27 | */ |
28 | 28 | ||
29 | #include<sftypes.h> |
29 | #include<sftypes.h> |
30 | #include<add.h> |
30 | #include<add.h> |
31 | #include<comparison.h> |
31 | #include<comparison.h> |
32 | 32 | ||
33 | /** Add two Float32 numbers with same signs |
33 | /** Add two Float32 numbers with same signs |
34 | */ |
34 | */ |
35 | float32 addFloat32(float32 a, float32 b) |
35 | float32 addFloat32(float32 a, float32 b) |
36 | { |
36 | { |
37 | int expdiff; |
37 | int expdiff; |
38 | __u32 exp1, exp2,frac1, frac2; |
38 | uint32_t exp1, exp2,frac1, frac2; |
39 | 39 | ||
40 | expdiff = a.parts.exp - b.parts.exp; |
40 | expdiff = a.parts.exp - b.parts.exp; |
41 | if (expdiff < 0) { |
41 | if (expdiff < 0) { |
42 | if (isFloat32NaN(b)) { |
42 | if (isFloat32NaN(b)) { |
43 | /* TODO: fix SigNaN */ |
43 | /* TODO: fix SigNaN */ |
44 | if (isFloat32SigNaN(b)) { |
44 | if (isFloat32SigNaN(b)) { |
45 | }; |
45 | }; |
46 | 46 | ||
47 | return b; |
47 | return b; |
48 | }; |
48 | }; |
49 | 49 | ||
50 | if (b.parts.exp == FLOAT32_MAX_EXPONENT) { |
50 | if (b.parts.exp == FLOAT32_MAX_EXPONENT) { |
51 | return b; |
51 | return b; |
52 | } |
52 | } |
53 | 53 | ||
54 | frac1 = b.parts.fraction; |
54 | frac1 = b.parts.fraction; |
55 | exp1 = b.parts.exp; |
55 | exp1 = b.parts.exp; |
56 | frac2 = a.parts.fraction; |
56 | frac2 = a.parts.fraction; |
57 | exp2 = a.parts.exp; |
57 | exp2 = a.parts.exp; |
58 | expdiff *= -1; |
58 | expdiff *= -1; |
59 | } else { |
59 | } else { |
60 | if ((isFloat32NaN(a)) || (isFloat32NaN(b))) { |
60 | if ((isFloat32NaN(a)) || (isFloat32NaN(b))) { |
61 | /* TODO: fix SigNaN */ |
61 | /* TODO: fix SigNaN */ |
62 | if (isFloat32SigNaN(a) || isFloat32SigNaN(b)) { |
62 | if (isFloat32SigNaN(a) || isFloat32SigNaN(b)) { |
63 | }; |
63 | }; |
64 | return (isFloat32NaN(a)?a:b); |
64 | return (isFloat32NaN(a)?a:b); |
65 | }; |
65 | }; |
66 | 66 | ||
67 | if (a.parts.exp == FLOAT32_MAX_EXPONENT) { |
67 | if (a.parts.exp == FLOAT32_MAX_EXPONENT) { |
68 | return a; |
68 | return a; |
69 | } |
69 | } |
70 | 70 | ||
71 | frac1 = a.parts.fraction; |
71 | frac1 = a.parts.fraction; |
72 | exp1 = a.parts.exp; |
72 | exp1 = a.parts.exp; |
73 | frac2 = b.parts.fraction; |
73 | frac2 = b.parts.fraction; |
74 | exp2 = b.parts.exp; |
74 | exp2 = b.parts.exp; |
75 | }; |
75 | }; |
76 | 76 | ||
77 | if (exp1 == 0) { |
77 | if (exp1 == 0) { |
78 | /* both are denormalized */ |
78 | /* both are denormalized */ |
79 | frac1 += frac2; |
79 | frac1 += frac2; |
80 | if (frac1 & FLOAT32_HIDDEN_BIT_MASK ) { |
80 | if (frac1 & FLOAT32_HIDDEN_BIT_MASK ) { |
81 | /* result is not denormalized */ |
81 | /* result is not denormalized */ |
82 | a.parts.exp = 1; |
82 | a.parts.exp = 1; |
83 | }; |
83 | }; |
84 | a.parts.fraction = frac1; |
84 | a.parts.fraction = frac1; |
85 | return a; |
85 | return a; |
86 | }; |
86 | }; |
87 | 87 | ||
88 | frac1 |= FLOAT32_HIDDEN_BIT_MASK; /* add hidden bit */ |
88 | frac1 |= FLOAT32_HIDDEN_BIT_MASK; /* add hidden bit */ |
89 | 89 | ||
90 | if (exp2 == 0) { |
90 | if (exp2 == 0) { |
91 | /* second operand is denormalized */ |
91 | /* second operand is denormalized */ |
92 | --expdiff; |
92 | --expdiff; |
93 | } else { |
93 | } else { |
94 | /* add hidden bit to second operand */ |
94 | /* add hidden bit to second operand */ |
95 | frac2 |= FLOAT32_HIDDEN_BIT_MASK; |
95 | frac2 |= FLOAT32_HIDDEN_BIT_MASK; |
96 | }; |
96 | }; |
97 | 97 | ||
98 | /* create some space for rounding */ |
98 | /* create some space for rounding */ |
99 | frac1 <<= 6; |
99 | frac1 <<= 6; |
100 | frac2 <<= 6; |
100 | frac2 <<= 6; |
101 | 101 | ||
102 | if (expdiff < (FLOAT32_FRACTION_SIZE + 2) ) { |
102 | if (expdiff < (FLOAT32_FRACTION_SIZE + 2) ) { |
103 | frac2 >>= expdiff; |
103 | frac2 >>= expdiff; |
104 | frac1 += frac2; |
104 | frac1 += frac2; |
105 | } else { |
105 | } else { |
106 | a.parts.exp = exp1; |
106 | a.parts.exp = exp1; |
107 | a.parts.fraction = (frac1 >> 6) & (~(FLOAT32_HIDDEN_BIT_MASK)); |
107 | a.parts.fraction = (frac1 >> 6) & (~(FLOAT32_HIDDEN_BIT_MASK)); |
108 | return a; |
108 | return a; |
109 | } |
109 | } |
110 | 110 | ||
111 | if (frac1 & (FLOAT32_HIDDEN_BIT_MASK << 7) ) { |
111 | if (frac1 & (FLOAT32_HIDDEN_BIT_MASK << 7) ) { |
112 | ++exp1; |
112 | ++exp1; |
113 | frac1 >>= 1; |
113 | frac1 >>= 1; |
114 | }; |
114 | }; |
115 | 115 | ||
116 | /* rounding - if first bit after fraction is set then round up */ |
116 | /* rounding - if first bit after fraction is set then round up */ |
117 | frac1 += (0x1 << 5); |
117 | frac1 += (0x1 << 5); |
118 | 118 | ||
119 | if (frac1 & (FLOAT32_HIDDEN_BIT_MASK << 7)) { |
119 | if (frac1 & (FLOAT32_HIDDEN_BIT_MASK << 7)) { |
120 | /* rounding overflow */ |
120 | /* rounding overflow */ |
121 | ++exp1; |
121 | ++exp1; |
122 | frac1 >>= 1; |
122 | frac1 >>= 1; |
123 | }; |
123 | }; |
124 | 124 | ||
125 | 125 | ||
126 | if ((exp1 == FLOAT32_MAX_EXPONENT ) || (exp2 > exp1)) { |
126 | if ((exp1 == FLOAT32_MAX_EXPONENT ) || (exp2 > exp1)) { |
127 | /* overflow - set infinity as result */ |
127 | /* overflow - set infinity as result */ |
128 | a.parts.exp = FLOAT32_MAX_EXPONENT; |
128 | a.parts.exp = FLOAT32_MAX_EXPONENT; |
129 | a.parts.fraction = 0; |
129 | a.parts.fraction = 0; |
130 | return a; |
130 | return a; |
131 | } |
131 | } |
132 | 132 | ||
133 | a.parts.exp = exp1; |
133 | a.parts.exp = exp1; |
134 | 134 | ||
135 | /*Clear hidden bit and shift */ |
135 | /*Clear hidden bit and shift */ |
136 | a.parts.fraction = ((frac1 >> 6) & (~FLOAT32_HIDDEN_BIT_MASK)) ; |
136 | a.parts.fraction = ((frac1 >> 6) & (~FLOAT32_HIDDEN_BIT_MASK)) ; |
137 | return a; |
137 | return a; |
138 | } |
138 | } |
139 | 139 | ||
140 | 140 | ||
141 | /** Add two Float64 numbers with same signs |
141 | /** Add two Float64 numbers with same signs |
142 | */ |
142 | */ |
143 | float64 addFloat64(float64 a, float64 b) |
143 | float64 addFloat64(float64 a, float64 b) |
144 | { |
144 | { |
145 | int expdiff; |
145 | int expdiff; |
146 | __u32 exp1, exp2; |
146 | uint32_t exp1, exp2; |
147 | __u64 frac1, frac2; |
147 | uint64_t frac1, frac2; |
148 | 148 | ||
149 | expdiff = ((int )a.parts.exp) - b.parts.exp; |
149 | expdiff = ((int )a.parts.exp) - b.parts.exp; |
150 | if (expdiff < 0) { |
150 | if (expdiff < 0) { |
151 | if (isFloat64NaN(b)) { |
151 | if (isFloat64NaN(b)) { |
152 | /* TODO: fix SigNaN */ |
152 | /* TODO: fix SigNaN */ |
153 | if (isFloat64SigNaN(b)) { |
153 | if (isFloat64SigNaN(b)) { |
154 | }; |
154 | }; |
155 | 155 | ||
156 | return b; |
156 | return b; |
157 | }; |
157 | }; |
158 | 158 | ||
159 | /* b is infinity and a not */ |
159 | /* b is infinity and a not */ |
160 | if (b.parts.exp == FLOAT64_MAX_EXPONENT ) { |
160 | if (b.parts.exp == FLOAT64_MAX_EXPONENT ) { |
161 | return b; |
161 | return b; |
162 | } |
162 | } |
163 | 163 | ||
164 | frac1 = b.parts.fraction; |
164 | frac1 = b.parts.fraction; |
165 | exp1 = b.parts.exp; |
165 | exp1 = b.parts.exp; |
166 | frac2 = a.parts.fraction; |
166 | frac2 = a.parts.fraction; |
167 | exp2 = a.parts.exp; |
167 | exp2 = a.parts.exp; |
168 | expdiff *= -1; |
168 | expdiff *= -1; |
169 | } else { |
169 | } else { |
170 | if (isFloat64NaN(a)) { |
170 | if (isFloat64NaN(a)) { |
171 | /* TODO: fix SigNaN */ |
171 | /* TODO: fix SigNaN */ |
172 | if (isFloat64SigNaN(a) || isFloat64SigNaN(b)) { |
172 | if (isFloat64SigNaN(a) || isFloat64SigNaN(b)) { |
173 | }; |
173 | }; |
174 | return a; |
174 | return a; |
175 | }; |
175 | }; |
176 | 176 | ||
177 | /* a is infinity and b not */ |
177 | /* a is infinity and b not */ |
178 | if (a.parts.exp == FLOAT64_MAX_EXPONENT ) { |
178 | if (a.parts.exp == FLOAT64_MAX_EXPONENT ) { |
179 | return a; |
179 | return a; |
180 | } |
180 | } |
181 | 181 | ||
182 | frac1 = a.parts.fraction; |
182 | frac1 = a.parts.fraction; |
183 | exp1 = a.parts.exp; |
183 | exp1 = a.parts.exp; |
184 | frac2 = b.parts.fraction; |
184 | frac2 = b.parts.fraction; |
185 | exp2 = b.parts.exp; |
185 | exp2 = b.parts.exp; |
186 | }; |
186 | }; |
187 | 187 | ||
188 | if (exp1 == 0) { |
188 | if (exp1 == 0) { |
189 | /* both are denormalized */ |
189 | /* both are denormalized */ |
190 | frac1 += frac2; |
190 | frac1 += frac2; |
191 | if (frac1 & FLOAT64_HIDDEN_BIT_MASK) { |
191 | if (frac1 & FLOAT64_HIDDEN_BIT_MASK) { |
192 | /* result is not denormalized */ |
192 | /* result is not denormalized */ |
193 | a.parts.exp = 1; |
193 | a.parts.exp = 1; |
194 | }; |
194 | }; |
195 | a.parts.fraction = frac1; |
195 | a.parts.fraction = frac1; |
196 | return a; |
196 | return a; |
197 | }; |
197 | }; |
198 | 198 | ||
199 | /* add hidden bit - frac1 is sure not denormalized */ |
199 | /* add hidden bit - frac1 is sure not denormalized */ |
200 | frac1 |= FLOAT64_HIDDEN_BIT_MASK; |
200 | frac1 |= FLOAT64_HIDDEN_BIT_MASK; |
201 | 201 | ||
202 | /* second operand ... */ |
202 | /* second operand ... */ |
203 | if (exp2 == 0) { |
203 | if (exp2 == 0) { |
204 | /* ... is denormalized */ |
204 | /* ... is denormalized */ |
205 | --expdiff; |
205 | --expdiff; |
206 | } else { |
206 | } else { |
207 | /* is not denormalized */ |
207 | /* is not denormalized */ |
208 | frac2 |= FLOAT64_HIDDEN_BIT_MASK; |
208 | frac2 |= FLOAT64_HIDDEN_BIT_MASK; |
209 | }; |
209 | }; |
210 | 210 | ||
211 | /* create some space for rounding */ |
211 | /* create some space for rounding */ |
212 | frac1 <<= 6; |
212 | frac1 <<= 6; |
213 | frac2 <<= 6; |
213 | frac2 <<= 6; |
214 | 214 | ||
215 | if (expdiff < (FLOAT64_FRACTION_SIZE + 2) ) { |
215 | if (expdiff < (FLOAT64_FRACTION_SIZE + 2) ) { |
216 | frac2 >>= expdiff; |
216 | frac2 >>= expdiff; |
217 | frac1 += frac2; |
217 | frac1 += frac2; |
218 | } else { |
218 | } else { |
219 | a.parts.exp = exp1; |
219 | a.parts.exp = exp1; |
220 | a.parts.fraction = (frac1 >> 6) & (~(FLOAT64_HIDDEN_BIT_MASK)); |
220 | a.parts.fraction = (frac1 >> 6) & (~(FLOAT64_HIDDEN_BIT_MASK)); |
221 | return a; |
221 | return a; |
222 | } |
222 | } |
223 | 223 | ||
224 | if (frac1 & (FLOAT64_HIDDEN_BIT_MASK << 7) ) { |
224 | if (frac1 & (FLOAT64_HIDDEN_BIT_MASK << 7) ) { |
225 | ++exp1; |
225 | ++exp1; |
226 | frac1 >>= 1; |
226 | frac1 >>= 1; |
227 | }; |
227 | }; |
228 | 228 | ||
229 | /* rounding - if first bit after fraction is set then round up */ |
229 | /* rounding - if first bit after fraction is set then round up */ |
230 | frac1 += (0x1 << 5); |
230 | frac1 += (0x1 << 5); |
231 | 231 | ||
232 | if (frac1 & (FLOAT64_HIDDEN_BIT_MASK << 7)) { |
232 | if (frac1 & (FLOAT64_HIDDEN_BIT_MASK << 7)) { |
233 | /* rounding overflow */ |
233 | /* rounding overflow */ |
234 | ++exp1; |
234 | ++exp1; |
235 | frac1 >>= 1; |
235 | frac1 >>= 1; |
236 | }; |
236 | }; |
237 | 237 | ||
238 | if ((exp1 == FLOAT64_MAX_EXPONENT ) || (exp2 > exp1)) { |
238 | if ((exp1 == FLOAT64_MAX_EXPONENT ) || (exp2 > exp1)) { |
239 | /* overflow - set infinity as result */ |
239 | /* overflow - set infinity as result */ |
240 | a.parts.exp = FLOAT64_MAX_EXPONENT; |
240 | a.parts.exp = FLOAT64_MAX_EXPONENT; |
241 | a.parts.fraction = 0; |
241 | a.parts.fraction = 0; |
242 | return a; |
242 | return a; |
243 | } |
243 | } |
244 | 244 | ||
245 | a.parts.exp = exp1; |
245 | a.parts.exp = exp1; |
246 | /*Clear hidden bit and shift */ |
246 | /*Clear hidden bit and shift */ |
247 | a.parts.fraction = ( (frac1 >> 6 ) & (~FLOAT64_HIDDEN_BIT_MASK)); |
247 | a.parts.fraction = ( (frac1 >> 6 ) & (~FLOAT64_HIDDEN_BIT_MASK)); |
248 | 248 | ||
249 | return a; |
249 | return a; |
250 | } |
250 | } |
251 | 251 | ||
252 | 252 | ||
253 | 253 |