| ... | ... | @@ -27,482 +27,545 @@ then statically linked and therefore is a transparent dependency for the |
| 27 | 27 | programmer. |
| 28 | 28 | For details see `../compiler_rt.zig`. |
| 29 | 29 | |
| 30 | | The routines in this folder are listed below. |
| 31 | | Routines are annotated as `type source routine // description`, with `routine` |
| 32 | | being the name used in aforementioned `compiler_rt.zig`. |
| 33 | | `dev` means deviating from compiler_rt, `port` ported, `source` is the |
| 34 | | information source for the implementation, `none` means unimplemented. |
| 35 | | Some examples for the naming convention are: |
| 36 | | - dev source name_routine, name_routine2 various implementations for performance, simplicity etc |
| 37 | | - port llvm compiler-rt library routines from [LLVM](http://compiler-rt.llvm.org/) |
| 38 | | * LLVM emits library calls to compiler-rt, if the hardware lacks functionality |
| 39 | | - port musl libc routines from [musl](https://musl.libc.org/) |
| 40 | | If the library or information source is uncommon, use the entry `other` for `source`. |
| 41 | | Please do not break the search by inserting entries in another format than `impl space source`. |
| 42 | | |
| 43 | 30 | Bugs should be solved by trying to duplicate the bug upstream, if possible. |
| 44 | 31 | * If the bug exists upstream, get it fixed upstream and port the fix downstream to Zig. |
| 45 | 32 | * If the bug only exists in Zig, use the corresponding C code and debug |
| 46 | 33 | both implementations side by side to figure out what is wrong. |
| 47 | 34 | |
| 48 | | ## Integer library routines |
| 49 | | |
| 50 | | #### Integer Bit operations |
| 51 | | |
| 52 | | - dev HackersDelight __clzsi2 // count leading zeros |
| 53 | | - dev HackersDelight __clzdi2 // count leading zeros |
| 54 | | - dev HackersDelight __clzti2 // count leading zeros |
| 55 | | - dev HackersDelight __ctzsi2 // count trailing zeros |
| 56 | | - dev HackersDelight __ctzdi2 // count trailing zeros |
| 57 | | - dev HackersDelight __ctzti2 // count trailing zeros |
| 58 | | - dev __ctzsi2 __ffssi2 // find least significant 1 bit |
| 59 | | - dev __ctzsi2 __ffsdi2 // find least significant 1 bit |
| 60 | | - dev __ctzsi2 __ffsti2 // find least significant 1 bit |
| 61 | | - dev BitTwiddlingHacks __paritysi2 // bit parity |
| 62 | | - dev BitTwiddlingHacks __paritydi2 // bit parity |
| 63 | | - dev BitTwiddlingHacks __parityti2 // bit parity |
| 64 | | - dev TAOCP __popcountsi2 // bit population |
| 65 | | - dev TAOCP __popcountdi2 // bit population |
| 66 | | - dev TAOCP __popcountti2 // bit population |
| 67 | | - dev other __bswapsi2 // a byteswapped |
| 68 | | - dev other __bswapdi2 // a byteswapped |
| 69 | | - dev other __bswapti2 // a byteswapped |
| 70 | | |
| 71 | | #### Integer Comparison |
| 72 | | |
| 73 | | - port llvm __cmpsi2 // a,b: i32, (a<b)-> 0, (a==b) -> 1, (a>b) -> 2 |
| 74 | | - port llvm __cmpdi2 // a,b: i64 |
| 75 | | - port llvm __cmpti2 // a,b: i128 |
| 76 | | - port llvm __ucmpsi2 // a,b: u32, (a<b)-> 0, (a==b) -> 1, (a>b) -> 2 |
| 77 | | - port llvm __ucmpdi2 // a,b: u64 |
| 78 | | - port llvm __ucmpti2 // a,b: u128 |
| 79 | | |
| 80 | | #### Integer Arithmetic |
| 81 | | |
| 82 | | - none none __ashlsi3 // a,b: i32, a << b unused in llvm, TODO (e.g. used by rl78) |
| 83 | | - port llvm __ashldi3 // a,b: u64 |
| 84 | | - port llvm __ashlti3 // a,b: u128 |
| 85 | | - none none __ashrsi3 // a,b: i32, a >> b arithmetic (sign fill) TODO (e.g. used by rl78) |
| 86 | | - port llvm __ashrdi3 // .. |
| 87 | | - port llvm __ashrti3 // |
| 88 | | - none none __lshrsi3 // a,b: i32, a >> b logical (zero fill) TODO (e.g. used by rl78) |
| 89 | | - port llvm __lshrdi3 // |
| 90 | | - port llvm __lshrti3 // |
| 91 | | - port llvm __negdi2 // a: i32, -a, symbol-level compatibility with libgcc |
| 92 | | - port llvm __negti2 // unnecessary: unused in backends |
| 93 | | - port llvm __mulsi3 // a,b: i32, a * b |
| 94 | | - port llvm __muldi3 // |
| 95 | | - port llvm __multi3 // |
| 96 | | - port llvm __divsi3 // a,b: i32, a / b |
| 97 | | - port llvm __divdi3 // |
| 98 | | - port llvm __divti3 // |
| 99 | | - port llvm __udivsi3 // a,b: u32, a / b |
| 100 | | - port llvm __udivdi3 // |
| 101 | | - port llvm __udivti3 // |
| 102 | | - port llvm __modsi3 // a,b: i32, a % b |
| 103 | | - port llvm __moddi3 // |
| 104 | | - port llvm __modti3 // |
| 105 | | - port llvm __umodsi3 // a,b: u32, a % b |
| 106 | | - port llvm __umoddi3 // |
| 107 | | - port llvm __umodti3 // |
| 108 | | - port llvm __udivmoddi4 // a,b: u32, a / b, rem.* = a % b unsigned |
| 109 | | - port llvm __udivmodti4 // |
| 110 | | - port llvm __udivmodsi4 // |
| 111 | | - port llvm __divmodsi4 // a,b: i32, a / b, rem.* = a % b signed, ARM |
| 112 | | - port llvm __divmoddi4 // |
| 113 | | |
| 114 | | #### Integer Arithmetic with trapping overflow |
| 115 | | |
| 116 | | - dev BitTwiddlingHacks __absvsi2 // abs(a) |
| 117 | | - dev BitTwiddlingHacks __absvdi2 // abs(a) |
| 118 | | - dev BitTwiddlingHacks __absvti2 // abs(a) |
| 119 | | - port llvm __negvsi2 // -a symbol-level compatibility: libgcc |
| 120 | | - port llvm __negvdi2 // -a unnecessary: unused in backends |
| 121 | | - port llvm __negvti2 // -a |
| 122 | | - TODO upstreaming __addvsi3..__mulvti3 after testing panics works |
| 123 | | - dev HackersDelight __addvsi3 // a + b |
| 124 | | - dev HackersDelight __addvdi3 // |
| 125 | | - dev HackersDelight __addvti3 // |
| 126 | | - dev HackersDelight __subvsi3 // a - b |
| 127 | | - dev HackersDelight __subvdi3 // |
| 128 | | - dev HackersDelight __subvti3 // |
| 129 | | - dev HackersDelight __mulvsi3 // a * b |
| 130 | | - dev HackersDelight __mulvdi3 // |
| 131 | | - dev HackersDelight __mulvti3 // |
| 132 | | |
| 133 | | #### Integer Arithmetic which returns if overflow (would be faster without pointer) |
| 134 | | |
| 135 | | - dev HackersDelight __addosi4 // a + b, overflow->ov.*=1 else 0 |
| 136 | | - dev HackersDelight __addodi4 // (completeness + performance, llvm does not use them) |
| 137 | | - dev HackersDelight __addoti4 // |
| 138 | | - dev HackersDelight __subosi4 // a - b, overflow->ov.*=1 else 0 |
| 139 | | - dev HackersDelight __subodi4 // (completeness + performance, llvm does not use them) |
| 140 | | - dev HackersDelight __suboti4 // |
| 141 | | - dev HackersDelight __mulosi4 // a * b, overflow->ov.*=1 else 0 |
| 142 | | - dev HackersDelight __mulodi4 // (required by llvm) |
| 143 | | - dev HackersDelight __muloti4 // |
| 144 | | |
| 145 | | ## Float library routines |
| 146 | | |
| 147 | | TODO: review source of implementation |
| 148 | | |
| 149 | | #### Float Conversion |
| 150 | | |
| 151 | | - dev other __extendsfdf2 // a: f32 -> f64, TODO: missing tests |
| 152 | | - dev other __extendsftf2 // a: f32 -> f128 |
| 153 | | - dev llvm __extendsfxf2 // a: f32 -> f80, TODO: missing tests |
| 154 | | - dev other __extenddftf2 // a: f64 -> f128 |
| 155 | | - dev llvm __extenddfxf2 // a: f64 -> f80 |
| 156 | | - dev other __truncdfsf2 // a: f64 -> f32, rounding towards zero |
| 157 | | - dev other __trunctfdf2 // a: f128-> f64 |
| 158 | | - dev other __trunctfsf2 // a: f128-> f32 |
| 159 | | - dev llvm __truncxfsf2 // a: f80 -> f32, TODO: missing tests |
| 160 | | - dev llvm __truncxfdf2 // a: f80 -> f64, TODO: missing tests |
| 161 | | |
| 162 | | - dev unclear __fixsfsi // a: f32 -> i32, rounding towards zero |
| 163 | | - dev unclear __fixdfsi // a: f64 -> i32 |
| 164 | | - dev unclear __fixtfsi // a: f128-> i32 |
| 165 | | - dev unclear __fixxfsi // a: f80 -> i32, TODO: missing tests |
| 166 | | - dev unclear __fixsfdi // a: f32 -> i64, rounding towards zero |
| 167 | | - dev unclear __fixdfdi // .. |
| 168 | | - dev unclear __fixtfdi // |
| 169 | | - dev unclear __fixxfdi // TODO: missing tests |
| 170 | | - dev unclear __fixsfti // a: f32 -> i128, rounding towards zero |
| 171 | | - dev unclear __fixdfti // .. |
| 172 | | - dev unclear __fixtfdi // |
| 173 | | - dev unclear __fixxfti // TODO: missing tests |
| 174 | | |
| 175 | | - dev unclear __fixunssfsi // a: f32 -> u32, rounding towards zero. negative values become 0. |
| 176 | | - dev unclear __fixunsdfsi // .. |
| 177 | | - dev unclear __fixunstfsi // |
| 178 | | - dev unclear __fixunsxfsi // TODO: missing tests |
| 179 | | - dev unclear __fixunssfdi // a: f32 -> u64, rounding towards zero. negative values become 0. |
| 180 | | - dev unclear __fixunsdfdi // |
| 181 | | - dev unclear __fixunstfdi // |
| 182 | | - dev unclear __fixunsxfdi // TODO: missing tests |
| 183 | | - dev unclear __fixunssfti // a: f32 -> u128, rounding towards zero. negative values become 0. |
| 184 | | - dev unclear __fixunsdfti // |
| 185 | | - dev unclear __fixunstfdi // |
| 186 | | - dev unclear __fixunsxfti // TODO: some more tests needed for base coverage |
| 187 | | |
| 188 | | - dev unclear __floatsisf // a: i32 -> f32 |
| 189 | | - dev unclear __floatsidf // a: i32 -> f64, TODO: missing tests |
| 190 | | - dev unclear __floatsitf // .. |
| 191 | | - dev unclear __floatsixf // TODO: missing tests |
| 192 | | - dev unclear __floatdisf // a: i64 -> f32 |
| 193 | | - dev unclear __floatdidf // |
| 194 | | - dev unclear __floatditf // |
| 195 | | - dev unclear __floatdixf // TODO: missing tests |
| 196 | | - dev unclear __floattisf // a: i128-> f32 |
| 197 | | - dev unclear __floattidf // |
| 198 | | - dev unclear __floattitf // |
| 199 | | - dev unclear __floattixf // TODO: missing tests |
| 200 | | |
| 201 | | - dev unclear __floatunsisf // a: u32 -> f32 |
| 202 | | - dev unclear __floatunsidf // TODO: missing tests |
| 203 | | - dev unclear __floatunsitf // |
| 204 | | - dev unclear __floatunsixf // TODO: missing tests |
| 205 | | - dev unclear __floatundisf // a: u64 -> f32 |
| 206 | | - dev unclear __floatundidf // |
| 207 | | - dev unclear __floatunditf // |
| 208 | | - dev unclear __floatundixf // TODO: missing tests |
| 209 | | - dev unclear __floatuntisf // a: u128-> f32 |
| 210 | | - dev unclear __floatuntidf // |
| 211 | | - dev unclear __floatuntitf // |
| 212 | | - dev unclear __floatuntixf // TODO: missing tests |
| 213 | | |
| 214 | | #### Float Comparison |
| 215 | | |
| 216 | | - dev other __cmpsf2 // a,b:f32, (a<b)->-1,(a==b)->0,(a>b)->1,Nan->1 |
| 217 | | - dev other __cmpdf2 // exported from __lesf2, __ledf2, __letf2 (below) |
| 218 | | - dev other __cmptf2 // But: if NaN is a possibility, use another routine. |
| 219 | | - dev other __unordsf2 // a,b:f32, (a==+-NaN or b==+-NaN) -> !=0, else -> 0 |
| 220 | | - dev other __unorddf2 // __only reliable for (input!=NaN)__ |
| 221 | | - dev other __unordtf2 // TODO: missing tests |
| 222 | | - dev other __eqsf2 // (a!=NaN) and (b!=Nan) and (a==b) -> output=0 |
| 223 | | - dev other __eqdf2 // |
| 224 | | - dev other __eqtf2 // |
| 225 | | - dev other __nesf2 // (a==NaN) or (b==Nan) or (a!=b) -> output!=0 |
| 226 | | - dev other __nedf2 // |
| 227 | | - dev other __netf2 // __eqtf2 and __netf2 have same return value -> tested with __eqsf2 |
| 228 | | - dev other __gesf2 // (a!=Nan) and (b!=Nan) and (a>=b) -> output>=0 |
| 229 | | - dev other __gedf2 // |
| 230 | | - dev other __getf2 // TODO: missing tests |
| 231 | | - dev other __ltsf2 // (a!=Nan) and (b!=Nan) and (a<b) -> output<0 |
| 232 | | - dev other __ltdf2 // |
| 233 | | - dev other __lttf2 // TODO: missing tests |
| 234 | | - dev other __lesf2 // (a!=Nan) and (b!=Nan) and (a<=b) -> output<=0 |
| 235 | | - dev other __ledf2 // |
| 236 | | - dev other __letf2 // TODO: missing tests |
| 237 | | - dev other __gtsf2 // (a!=Nan) and (b!=Nan) and (a>b) -> output>0 |
| 238 | | - dev other __gtdf2 // |
| 239 | | - dev other __gttf2 // TODO: missing tests |
| 240 | | |
| 241 | | #### Float Arithmetic |
| 242 | | |
| 243 | | - dev unclear __addsf3 // a + b f32, TODO: missing tests |
| 244 | | - dev unclear __adddf3 // a + b f64, TODO: missing tests |
| 245 | | - dev unclear __addtf3 // a + b f128 |
| 246 | | - dev unclear __addxf3 // a + b f80 |
| 247 | | - dev unclear __aeabi_fadd // a + b f64 ARM: AAPCS |
| 248 | | - dev unclear __aeabi_dadd // a + b f64 ARM: AAPCS |
| 249 | | - dev unclear __subsf3 // a - b, TODO: missing tests |
| 250 | | - dev unclear __subdf3 // a - b, TODO: missing tests |
| 251 | | - dev unclear __subtf3 // a - b |
| 252 | | - dev unclear __subxf3 // a - b f80, TODO: missing tests |
| 253 | | - dev unclear __aeabi_fsub // a - b f64 ARM: AAPCS |
| 254 | | - dev unclear __aeabi_dsub // a - b f64 ARM: AAPCS |
| 255 | | - dev unclear __mulsf3 // a * b, TODO: missing tests |
| 256 | | - dev unclear __muldf3 // a * b, TODO: missing tests |
| 257 | | - dev unclear __multf3 // a * b |
| 258 | | - dev unclear __mulxf3 // a * b |
| 259 | | - dev unclear __divsf3 // a / b, TODO: review tests |
| 260 | | - dev unclear __divdf3 // a / b, TODO: review tests |
| 261 | | - dev unclear __divtf3 // a / b |
| 262 | | - dev unclear __divxf3 // a / b |
| 263 | | - dev unclear __negsf2 // -a symbol-level compatibility: libgcc uses this for the rl78 |
| 264 | | - dev unclear __negdf2 // -a unnecessary: can be lowered directly to a xor |
| 265 | | - dev unclear __negtf2 // -a, TODO: missing tests |
| 266 | | - dev unclear __negxf2 // -a, TODO: missing tests |
| 267 | | |
| 268 | | #### Floating point raised to integer power |
| 269 | | - dev unclear __powisf2 // a ^ b, TODO |
| 270 | | - dev unclear __powidf2 // |
| 271 | | - dev unclear __powitf2 // |
| 272 | | - dev unclear __powixf2 // |
| 273 | | - dev unclear __mulsc3 // (a+ib) * (c+id) |
| 274 | | - dev unclear __muldc3 // |
| 275 | | - dev unclear __multc3 // |
| 276 | | - dev unclear __mulxc3 // |
| 277 | | - dev unclear __divsc3 // (a+ib) * / (c+id) |
| 278 | | - dev unclear __divdc3 // |
| 279 | | - dev unclear __divtc3 // |
| 280 | | - dev unclear __divxc3 // |
| 281 | | |
| 282 | | ## Decimal float library routines |
| 35 | Routines with status are given below. Sources were besides |
| 36 | "The Art of Computer Programming" by Donald E. Knuth, "HackersDelight" by Henry S. Warren, |
| 37 | "Bit Twiddling Hacks" collected by Sean Eron Anderson, "Berkeley SoftFloat" by John R. Hauser, |
| 38 | LLVM "compiler-rt" as it was MIT-licensed, "musl libc" and thoughts + work of contributors. |
| 39 | |
| 40 | The compiler-rt routines have not yet been audited. |
| 41 | See https://github.com/ziglang/zig/issues/1504. |
| 42 | |
| 43 | From left to right the columns mean 1. if the routine is implemented (✗ or ✓), |
| 44 | 2. the name, 3. input (`a`), 4. input (`b`), 5. return value, |
| 45 | 6. an explanation of the functionality, .. to repeat the comment from the |
| 46 | column a row above and/or additional return values. |
| 47 | Some routines have more extensive comments supplemented with a reference text. |
| 48 | |
| 49 | Integer and Float Operations |
| 50 | |
| 51 | | Done | Name | a | b | Out | Comment | |
| 52 | | ------ | ------------- | ---- | ---- | ---- | ------------------------------ | |
| 53 | | | | | | | **Integer Bit Operations** | |
| 54 | | ✓ | __clzsi2 | u32 | ∅ | i32 | count leading zeroes | |
| 55 | | ✓ | __clzdi2 | u64 | ∅ | i32 | count leading zeroes | |
| 56 | | ✓ | __clzti2 | u128 | ∅ | i32 | count trailing zeros | |
| 57 | | ✓ | __ctzsi2 | u32 | ∅ | i32 | count trailing zeros | |
| 58 | | ✓ | __ctzdi2 | u64 | ∅ | i32 | count trailing zeros | |
| 59 | | ✓ | __ctzti2 | u128 | ∅ | i32 | count leading zeroes | |
| 60 | | ✓ | __ffssi2 | u32 | ∅ | i32 | count leading zeroes | |
| 61 | | ✓ | __ffsdi2 | u64 | ∅ | i32 | count leading zeroes | |
| 62 | | ✓ | __ffsti2 | u128 | ∅ | i32 | count leading zeroes | |
| 63 | | ✓ | __paritysi2 | u32 | ∅ | i32 | find least significant 1 bit | |
| 64 | | ✓ | __paritydi2 | u64 | ∅ | i32 | find least significant 1 bit | |
| 65 | | ✓ | __parityti2 | u128 | ∅ | i32 | find least significant 1 bit | |
| 66 | | ✓ | __popcountsi2 | u32 | ∅ | i32 | bit population | |
| 67 | | ✓ | __popcountdi2 | u64 | ∅ | i32 | bit population | |
| 68 | | ✓ | __popcountti2 | u128 | ∅ | i32 | bit population | |
| 69 | | ✓ | __bswapsi2 | u32 | ∅ | i32 | bit parity | |
| 70 | | ✓ | __bswapdi2 | u64 | ∅ | i32 | bit parity | |
| 71 | | ✓ | __bswapti2 | u128 | ∅ | i32 | bit parity | |
| 72 | | | | | | | **Integer Comparison** | |
| 73 | | ✓ | __cmpsi2 | i32 | i32 | i32 | `(a<b) -> 0, (a==b) -> 1, (a>b) -> 2` | |
| 74 | | ✓ | __cmpdi2 | i64 | i64 | i32 | .. | |
| 75 | | ✓ | __cmpti2 | i128 | i128 | i32 | .. | |
| 76 | | ✓ | __ucmpsi2 | i32 | i32 | i32 | `(a<b) -> 0, (a==b) -> 1, (a>b) -> 2` | |
| 77 | | ✓ | __ucmpdi2 | i64 | i64 | i32 | .. | |
| 78 | | ✓ | __ucmpti2 | i128 | i128 | i32 | .. | |
| 79 | | | | | | | **Integer Arithmetic** | |
| 80 | | ✗ | __ashlsi3 | i32 | i32 | i32 | `a << b` [^unused_rl78] | |
| 81 | | ✓ | __ashldi3 | i64 | i32 | i64 | .. | |
| 82 | | ✓ | __ashlti3 | i128 | i32 | i128 | .. | |
| 83 | | ✓ | __aeabi_llsl | i32 | i32 | i32 | .. ARM | |
| 84 | | ✗ | __ashrsi3 | i32 | i32 | i32 | `a >> b` arithmetic (sign fill) [^unused_rl78] | |
| 85 | | ✓ | __ashrdi3 | i64 | i32 | i64 | .. | |
| 86 | | ✓ | __ashrti3 | i128 | i32 | i128 | .. | |
| 87 | | ✓ | __aeabi_lasr | i64 | i32 | i64 | .. ARM | |
| 88 | | ✗ | __lshrsi3 | i32 | i32 | i32 | `a >> b` logical (zero fill) [^unused_rl78] | |
| 89 | | ✓ | __lshrdi3 | i64 | i32 | i64 | .. | |
| 90 | | ✓ | __lshrti3 | i128 | i32 | i128 | .. | |
| 91 | | ✓ | __aeabi_llsr | i64 | i32 | i64 | .. ARM | |
| 92 | | ✓ | __negsi2 | i32 | i32 | i32 | `-a` [^libgcc_compat] | |
| 93 | | ✓ | __negdi2 | i64 | i64 | i64 | .. | |
| 94 | | ✓ | __negti2 | i128 | i128 | i128 | .. | |
| 95 | | ✓ | __mulsi3 | i32 | i32 | i32 | `a * b` | |
| 96 | | ✓ | __muldi3 | i64 | i64 | i64 | .. | |
| 97 | | ✓ | __multi3 | i128 | i128 | i128 | .. | |
| 98 | | ✓ | __divsi3 | i32 | i32 | i32 | `a / b` | |
| 99 | | ✓ | __divdi3 | i64 | i64 | i64 | .. | |
| 100 | | ✓ | __divti3 | i128 | i128 | i128 | .. | |
| 101 | | ✓ | __aeabi_idiv | i32 | i32 | i32 | .. ARM | |
| 102 | | ✓ | __udivsi3 | u32 | u32 | u32 | `a / b` | |
| 103 | | ✓ | __udivdi3 | u64 | u64 | u64 | .. | |
| 104 | | ✓ | __udivti3 | u128 | u128 | u128 | .. | |
| 105 | | ✓ | __aeabi_uidiv | i32 | i32 | i32 | .. ARM | |
| 106 | | ✓ | __modsi3 | i32 | i32 | i32 | `a % b` | |
| 107 | | ✓ | __moddi3 | i64 | i64 | i64 | .. | |
| 108 | | ✓ | __modti3 | i128 | i128 | i128 | .. | |
| 109 | | ✓ | __umodsi3 | u32 | u32 | u32 | `a % b` | |
| 110 | | ✓ | __umoddi3 | u64 | u64 | u64 | .. | |
| 111 | | ✓ | __umodti3 | u128 | u128 | u128 | .. | |
| 112 | | ✓ | __udivmodsi4 | u32 | u32 | u32 | `a / b, rem.* = a % b` | |
| 113 | | ✓ | __udivmoddi4 | u64 | u64 | u64 | .. | |
| 114 | | ✓ | __udivmodti4 | u128 | u128 | u128 | .. | |
| 115 | | ✓ | __divmodsi4 | i32 | i32 | i32 | `a / b, rem.* = a % b` | |
| 116 | | ✓ | __divmoddi4 | i64 | i64 | i64 | .. | |
| 117 | | ✗ | __divmodti4 | i128 | i128 | i128 | .. [^libgcc_compat] | |
| 118 | | | | | | | **Integer Arithmetic with Trapping Overflow**| |
| 119 | | ✓ | __absvsi2 | i32 | i32 | i32 | abs(a) | |
| 120 | | ✓ | __absvdi2 | i64 | i64 | i64 | .. | |
| 121 | | ✓ | __absvti2 | i128 | i128 | i128 | .. | |
| 122 | | ✓ | __negvsi2 | i32 | i32 | i32 | `-a` [^libgcc_compat] | |
| 123 | | ✓ | __negvdi2 | i64 | i64 | i64 | .. | |
| 124 | | ✓ | __negvti2 | i128 | i128 | i128 | .. | |
| 125 | | ✗ | __addvsi3 | i32 | i32 | i32 | `a + b` | |
| 126 | | ✗ | __addvdi3 | i64 | i64 | i64 | .. | |
| 127 | | ✗ | __addvti3 | i128 | i128 | i128 | .. | |
| 128 | | ✗ | __subvsi3 | i32 | i32 | i32 | `a - b` | |
| 129 | | ✗ | __subvdi3 | i64 | i64 | i64 | .. | |
| 130 | | ✗ | __subvti3 | i128 | i128 | i128 | .. | |
| 131 | | ✗ | __mulvsi3 | i32 | i32 | i32 | `a * b` | |
| 132 | | ✗ | __mulvdi3 | i64 | i64 | i64 | .. | |
| 133 | | ✗ | __mulvti3 | i128 | i128 | i128 | .. | |
| 134 | | | | | | | **Integer Arithmetic which Return on Overflow** [^noptr_faster] | |
| 135 | | ✓ | __addosi4 | i32 | i32 | i32 | `a + b`, overflow->ov.*=1 else 0 [^perf_addition] | |
| 136 | | ✓ | __addodi4 | i64 | i64 | i64 | .. | |
| 137 | | ✓ | __addoti4 | i128 | i128 | i128 | .. | |
| 138 | | ✓ | __subosi4 | i32 | i32 | i32 | `a - b`, overflow->ov.*=1 else 0 [^perf_addition] | |
| 139 | | ✓ | __subodi4 | i64 | i64 | i64 | .. | |
| 140 | | ✓ | __suboti4 | i128 | i128 | i128 | .. | |
| 141 | | ✓ | __mulosi4 | i32 | i32 | i32 | `a * b`, overflow->ov.*=1 else 0 | |
| 142 | | ✓ | __mulodi4 | i64 | i64 | i64 | .. | |
| 143 | | ✓ | __muloti4 | i128 | i128 | i128 | .. | |
| 144 | | | | | | | **Float Conversion** | |
| 145 | | ✓ | __extendsfdf2 | f32 | ∅ | f64 | .. | |
| 146 | | ✓ | __extendsftf2 | f32 | ∅ | f128 | .. | |
| 147 | | ✓ | __extendsfxf2 | f32 | ∅ | f80 | .. | |
| 148 | | ✓ | __extenddftf2 | f64 | ∅ | f128 | .. | |
| 149 | | ✓ | __extenddfxf2 | f64 | ∅ | f80 | .. | |
| 150 | | ✓ | __truncsfhf2 | f32 | ∅ | f16 | rounding towards zero | |
| 151 | | ✓ | __truncdfhf2 | f64 | ∅ | f16 | .. | |
| 152 | | ✓ | __truncdfsf2 | f64 | ∅ | f32 | .. | |
| 153 | | ✓ | __trunctfhf2 | f128 | ∅ | f16 | .. | |
| 154 | | ✓ | __trunctfsf2 | f128 | ∅ | f32 | .. | |
| 155 | | ✓ | __trunctfdf2 | f128 | ∅ | f64 | .. | |
| 156 | | ✓ | __trunctfxf2 | f128 | ∅ | f80 | .. | |
| 157 | | ✓ | __truncxfhf2 | f80 | ∅ | f16 | .. | |
| 158 | | ✓ | __truncxfsf2 | f80 | ∅ | f32 | .. | |
| 159 | | ✓ | __truncxfdf2 | f80 | ∅ | f64 | .. | |
| 160 | | ✓ | __aeabi_f2h | f32 | ∅ | f16 | .. ARM | |
| 161 | | ✓ | __gnu_f2h_ieee | f32 | ∅ | f16 | ..GNU naming convention | |
| 162 | | ✓ | __aeabi_d2h | f64 | ∅ | f16 | .. ARM | |
| 163 | | ✓ | __aeabi_d2f | f64 | ∅ | f32 | .. ARM | |
| 164 | | ✓ | __trunckfsf2 | f128 | ∅ | f32 | .. PPC | |
| 165 | | ✓ | _Qp_qtos |*f128 | ∅ | f32 | .. SPARC | |
| 166 | | ✓ | __trunckfdf2 | f128 | ∅ | f64 | .. PPC | |
| 167 | | ✓ | _Qp_qtod |*f128 | ∅ | f64 | .. SPARC | |
| 168 | | ✓ | __fixhfsi | f16 | ∅ | i32 | rounding towards zero | |
| 169 | | ✓ | __fixsfsi | f32 | ∅ | i32 | .. | |
| 170 | | ✓ | __fixdfsi | f64 | ∅ | i32 | .. | |
| 171 | | ✓ | __fixtfsi | f128 | ∅ | i32 | .. | |
| 172 | | ✓ | __fixxfsi | f80 | ∅ | i32 | .. | |
| 173 | | ✓ | __fixhfdi | f16 | ∅ | i64 | .. | |
| 174 | | ✓ | __fixsfdi | f32 | ∅ | i64 | .. | |
| 175 | | ✓ | __fixdfdi | f64 | ∅ | i64 | .. | |
| 176 | | ✓ | __fixtfdi | f128 | ∅ | i64 | .. | |
| 177 | | ✓ | __fixxfdi | f80 | ∅ | i64 | .. | |
| 178 | | ✓ | __fixhfti | f16 | ∅ | i128 | .. | |
| 179 | | ✓ | __fixsfti | f32 | ∅ | i128 | .. | |
| 180 | | ✓ | __fixdfti | f64 | ∅ | i128 | .. | |
| 181 | | ✓ | __fixtfti | f128 | ∅ | i128 | .. | |
| 182 | | ✓ | __fixxfti | f80 | ∅ | i128 | .. | |
| 183 | | ✓ | __fixunshfsi | f16 | ∅ | u32 | rounding towards zero. negative values become 0. | |
| 184 | | ✓ | __fixunssfsi | f32 | ∅ | u32 | .. | |
| 185 | | ✓ | __fixunsdfsi | f64 | ∅ | u32 | .. | |
| 186 | | ✓ | __fixunstfsi | f128 | ∅ | u32 | .. | |
| 187 | | ✓ | __fixunsxfsi | f80 | ∅ | u32 | .. | |
| 188 | | ✓ | __fixunshfdi | f16 | ∅ | u64 | .. | |
| 189 | | ✓ | __fixunssfdi | f32 | ∅ | u64 | .. | |
| 190 | | ✓ | __fixunsdfdi | f64 | ∅ | u64 | .. | |
| 191 | | ✓ | __fixunstfdi | f128 | ∅ | u64 | .. | |
| 192 | | ✓ | __fixunsxfdi | f80 | ∅ | u64 | .. | |
| 193 | | ✓ | __fixunshfti | f16 | ∅ | u128 | .. | |
| 194 | | ✓ | __fixunssfti | f32 | ∅ | u128 | .. | |
| 195 | | ✓ | __fixunsdfti | f64 | ∅ | u128 | .. | |
| 196 | | ✓ | __fixunstfti | f128 | ∅ | u128 | .. | |
| 197 | | ✓ | __fixunsxfti | f80 | ∅ | u128 | .. | |
| 198 | | ✓ | __floatsihf | i32 | ∅ | f16 | int_to_float conversions | |
| 199 | | ✓ | __floatsisf | i32 | ∅ | f32 | .. | |
| 200 | | ✓ | __floatsidf | i32 | ∅ | f64 | .. | |
| 201 | | ✓ | __floatsitf | i32 | ∅ | f128 | .. | |
| 202 | | ✓ | __floatsixf | i32 | ∅ | f80 | .. | |
| 203 | | ✓ | __floatdisf | i64 | ∅ | f32 | .. | |
| 204 | | ✓ | __floatdidf | i64 | ∅ | f64 | .. | |
| 205 | | ✓ | __floatditf | i64 | ∅ | f128 | .. | |
| 206 | | ✓ | __floatdixf | i64 | ∅ | f80 | .. | |
| 207 | | ✓ | __floattihf | i128 | ∅ | f16 | .. | |
| 208 | | ✓ | __floattisf | i128 | ∅ | f32 | .. | |
| 209 | | ✓ | __floattidf | i128 | ∅ | f64 | .. | |
| 210 | | ✓ | __floattitf | i128 | ∅ | f128 | .. | |
| 211 | | ✓ | __floattixf | i128 | ∅ | f80 | .. | |
| 212 | | ✓ | __floatunsihf | u32 | ∅ | f16 | uint_to_float conversions | |
| 213 | | ✓ | __floatunsisf | u32 | ∅ | f32 | .. | |
| 214 | | ✓ | __floatunsidf | u32 | ∅ | f64 | .. | |
| 215 | | ✓ | __floatunsitf | u32 | ∅ | f128 | .. | |
| 216 | | ✓ | __floatunsixf | u32 | ∅ | f80 | .. | |
| 217 | | ✓ | __floatundihf | u64 | ∅ | f16 | .. | |
| 218 | | ✓ | __floatundisf | u64 | ∅ | f32 | .. | |
| 219 | | ✓ | __floatundidf | u64 | ∅ | f64 | .. | |
| 220 | | ✓ | __floatunditf | u64 | ∅ | f128 | .. | |
| 221 | | ✓ | __floatundixf | u64 | ∅ | f80 | .. | |
| 222 | | ✓ | __floatuntihf | u128 | ∅ | f16 | .. | |
| 223 | | ✓ | __floatuntisf | u128 | ∅ | f32 | .. | |
| 224 | | ✓ | __floatuntidf | u128 | ∅ | f64 | .. | |
| 225 | | ✓ | __floatuntitf | u128 | ∅ | f128 | .. | |
| 226 | | ✓ | __floatuntixf | u128 | ∅ | f80 | .. | |
| 227 | | | | | | | **Float Comparison** | |
| 228 | | ✓ | __cmphf2 | f16 | f16 | i32 | `(a<b)->-1, (a==b)->0, (a>b)->1, Nan->1` | |
| 229 | | ✓ | __cmpsf2 | f32 | f32 | i32 | exported from __lesf2, __ledf2, __letf2 (below) | |
| 230 | | ✓ | __cmpdf2 | f64 | f64 | i32 | But: if NaN is a possibility, use another routine. | |
| 231 | | ✓ | __cmptf2 | f128 | f128 | i32 | .. | |
| 232 | | ✓ | __cmpxf2 | f80 | f80 | i32 | .. | |
| 233 | | ✓ | _Qp_cmp |*f128 |*f128 | i32 | .. SPARC | |
| 234 | | ✓ | __unordhf2 | f16 | f16 | i32 | `(a==+-NaN or b==+-NaN) -> !=0, else -> 0` | |
| 235 | | ✓ | __unordsf2 | f32 | f32 | i32 | .. | |
| 236 | | ✓ | __unorddf2 | f64 | f64 | i32 | Note: only reliable for (input!=NaN) | |
| 237 | | ✓ | __unordtf2 | f128 | f128 | i32 | .. | |
| 238 | | ✓ | __unordxf2 | f80 | f80 | i32 | .. | |
| 239 | | ✓ | __aeabi_fcmpun | f32 | f32 | i32 | .. ARM | |
| 240 | | ✓ | __aeabi_dcmpun | f32 | f32 | i32 | .. ARM | |
| 241 | | ✓ | __unordkf2 | f128 | f128 | i32 | .. PPC | |
| 242 | | ✓ | __eqhf2 | f16 | f16 | i32 | `(a!=NaN) and (b!=Nan) and (a==b) -> output=0` | |
| 243 | | ✓ | __eqsf2 | f32 | f32 | i32 | .. | |
| 244 | | ✓ | __eqdf2 | f64 | f64 | i32 | .. | |
| 245 | | ✓ | __eqtf2 | f128 | f128 | i32 | .. | |
| 246 | | ✓ | __eqxf2 | f80 | f80 | i32 | .. | |
| 247 | | ✓ | __aeabi_fcmpeq | f32 | f32 | i32 | .. ARM | |
| 248 | | ✓ | __aeabi_dcmpeq | f32 | f32 | i32 | .. ARM | |
| 249 | | ✓ | __eqkf2 | f128 | f128 | i32 | .. PPC | |
| 250 | | ✓ | _Qp_feq |*f128 |*f128 | bool | .. SPARC | |
| 251 | | ✓ | __nehf2 | f16 | f16 | i32 | `(a==NaN) or (b==Nan) or (a!=b) -> output!=0` | |
| 252 | | ✓ | __nesf2 | f32 | f32 | i32 | Note: __eqXf2 and __neXf2 have same return value | |
| 253 | | ✓ | __nedf2 | f64 | f64 | i32 | .. | |
| 254 | | ✓ | __netf2 | f128 | f128 | i32 | .. | |
| 255 | | ✓ | __nexf2 | f80 | f80 | i32 | .. | |
| 256 | | ✓ | __nekf2 | f128 | f128 | i32 | .. PPC | |
| 257 | | ✓ | _Qp_fne |*f128 |*f128 | bool | .. SPARC | |
| 258 | | ✓ | __gehf2 | f16 | f16 | i32 | `(a!=Nan) and (b!=Nan) and (a>=b) -> output>=0` | |
| 259 | | ✓ | __gesf2 | f32 | f32 | i32 | .. | |
| 260 | | ✓ | __gedf2 | f64 | f64 | i32 | .. | |
| 261 | | ✓ | __getf2 | f128 | f128 | i32 | .. | |
| 262 | | ✓ | __gexf2 | f80 | f80 | i32 | .. | |
| 263 | | ✓ | __gekf2 | f128 | f128 | i32 | .. PPC | |
| 264 | | ✓ | _Qp_fge |*f128 |*f128 | bool | .. SPARC | |
| 265 | | ✓ | __lthf2 | f16 | f16 | i32 | `(a!=Nan) and (b!=Nan) and (a<b) -> output<0` | |
| 266 | | ✓ | __ltsf2 | f32 | f32 | i32 | .. | |
| 267 | | ✓ | __ltdf2 | f64 | f64 | i32 | .. | |
| 268 | | ✓ | __lttf2 | f128 | f128 | i32 | .. | |
| 269 | | ✓ | __ltxf2 | f80 | f80 | i32 | .. | |
| 270 | | ✓ | __ltkf2 | f128 | f128 | i32 | .. PPC | |
| 271 | | ✓ | __aeabi_fcmplt | f32 | f32 | i32 | .. ARM | |
| 272 | | ✓ | __aeabi_dcmplt | f32 | f32 | i32 | .. ARM | |
| 273 | | ✓ | _Qp_flt |*f128 |*f128 | bool | .. SPARC | |
| 274 | | ✓ | __lehf2 | f16 | f16 | i32 | `(a!=Nan) and (b!=Nan) and (a<=b) -> output<=0` | |
| 275 | | ✓ | __lesf2 | f32 | f32 | i32 | .. | |
| 276 | | ✓ | __ledf2 | f64 | f64 | i32 | .. | |
| 277 | | ✓ | __letf2 | f128 | f128 | i32 | .. | |
| 278 | | ✓ | __lexf2 | f80 | f80 | i32 | .. | |
| 279 | | ✓ | __aeabi_fcmple | f32 | f32 | i32 | .. ARM | |
| 280 | | ✓ | __aeabi_dcmple | f32 | f32 | i32 | .. ARM | |
| 281 | | ✓ | __lekf2 | f128 | f128 | i32 | .. PPC | |
| 282 | | ✓ | _Qp_fle |*f128 |*f128 | bool | .. SPARC | |
| 283 | | ✓ | __gthf2 | f16 | f16 | i32 | `(a!=Nan) and (b!=Nan) and (a>b) -> output>0` | |
| 284 | | ✓ | __gtsf2 | f32 | f32 | i32 | .. | |
| 285 | | ✓ | __gtdf2 | f64 | f64 | i32 | .. | |
| 286 | | ✓ | __gttf2 | f128 | f128 | i32 | .. | |
| 287 | | ✓ | __gtxf2 | f80 | f80 | i32 | .. | |
| 288 | | ✓ | __gtkf2 | f128 | f128 | i32 | .. PPC | |
| 289 | | ✓ | _Qp_fgt |*f128 |*f128 | bool | .. SPARC | |
| 290 | | | | | | | **Float Arithmetic** | |
| 291 | | ✓ | __addhf3 | f32 | f32 | f32 | `a + b` | |
| 292 | | ✓ | __addsf3 | f32 | f32 | f32 | .. | |
| 293 | | ✓ | __adddf3 | f64 | f64 | f64 | .. | |
| 294 | | ✓ | __addtf3 | f128 | f128 | f128 | .. | |
| 295 | | ✓ | __addxf3 | f80 | f80 | f80 | .. | |
| 296 | | ✓ | __aeabi_fadd | f32 | f32 | f32 | .. ARM | |
| 297 | | ✓ | __aeabi_dadd | f64 | f64 | f64 | .. ARM | |
| 298 | | ✓ | __addkf3 | f128 | f128 | f128 | .. PPC | |
| 299 | | ✓ | _Qp_add |*f128 |*f128 | void | .. SPARC args *c,*a,*b c=a+b | |
| 300 | | ✓ | __subhf3 | f32 | f32 | f32 | `a - b` | |
| 301 | | ✓ | __subsf3 | f32 | f32 | f32 | .. | |
| 302 | | ✓ | __subdf3 | f64 | f64 | f64 | .. | |
| 303 | | ✓ | __subtf3 | f128 | f128 | f128 | .. | |
| 304 | | ✓ | __subxf3 | f80 | f80 | f80 | .. | |
| 305 | | ✓ | __aeabi_fsub | f32 | f32 | f32 | .. ARM | |
| 306 | | ✓ | __aeabi_dsub | f64 | f64 | f64 | .. ARM | |
| 307 | | ✓ | __subkf3 | f128 | f128 | f128 | .. PPC | |
| 308 | | ✓ | _Qp_sub |*f128 |*f128 | void | .. SPARC args *c,*a,*b c=a-b | |
| 309 | | ✓ | __mulhf3 | f32 | f32 | f32 | `a * b` | |
| 310 | | ✓ | __mulsf3 | f32 | f32 | f32 | .. | |
| 311 | | ✓ | __muldf3 | f64 | f64 | f64 | .. | |
| 312 | | ✓ | __multf3 | f128 | f128 | f128 | .. | |
| 313 | | ✓ | __mulxf3 | f80 | f80 | f80 | .. | |
| 314 | | ✓ | __aeabi_fmul | f32 | f32 | f32 | .. ARM | |
| 315 | | ✓ | __aeabi_dmul | f64 | f64 | f64 | .. ARM | |
| 316 | | ✓ | __mulkf3 | f128 | f128 | f128 | .. PPC | |
| 317 | | ✓ | _Qp_mul |*f128 |*f128 | void | .. SPARC args *c,*a,*b c=a*b | |
| 318 | | ✓ | __divsf3 | f32 | f32 | f32 | `a / b` | |
| 319 | | ✓ | __divdf3 | f64 | f64 | f64 | .. | |
| 320 | | ✓ | __divtf3 | f128 | f128 | f128 | .. | |
| 321 | | ✓ | __divxf3 | f80 | f80 | f80 | .. | |
| 322 | | ✓ | __aeabi_fdiv | f32 | f32 | f32 | .. ARM | |
| 323 | | ✓ | __aeabi_ddiv | f64 | f64 | f64 | .. ARM | |
| 324 | | ✓ | __divkf3 | f128 | f128 | f128 | .. PPC | |
| 325 | | ✓ | _Qp_div |*f128 |*f128 | void | .. SPARC args *c,*a,*b c=a*b | |
| 326 | | ✓ | __negsf2 | f32 | ∅ | f32[^unused_rl78] | -a (can be lowered directly to a xor) | |
| 327 | | ✓ | __negdf2 | f64 | ∅ | f64 | .. | |
| 328 | | ✓ | __negtf2 | f128 | ∅ | f128 | .. | |
| 329 | | ✓ | __negxf2 | f80 | ∅ | f80 | .. | |
| 330 | | | | | | | **Floating point raised to integer power** | |
| 331 | | ✗ | __powihf2 | f16 | f16 | f16 | `a ^ b` | |
| 332 | | ✗ | __powisf2 | f32 | f32 | f32 | .. | |
| 333 | | ✗ | __powidf2 | f64 | f64 | f64 | .. | |
| 334 | | ✗ | __powitf2 | f128 | f128 | f128 | .. | |
| 335 | | ✗ | __powixf2 | f80 | f80 | f80 | .. | |
| 336 | | ✓ | __mulhc3 | all4 | f16 | f16 | `(a+ib) * (c+id)` | |
| 337 | | ✓ | __mulsc3 | all4 | f32 | f32 | .. | |
| 338 | | ✓ | __muldc3 | all4 | f64 | f64 | .. | |
| 339 | | ✓ | __multc3 | all4 | f128 | f128 | .. | |
| 340 | | ✓ | __mulxc3 | all4 | f80 | f80 | .. | |
| 341 | | ✓ | __divhc3 | all4 | f16 | f16 | `(a+ib) / (c+id)` | |
| 342 | | ✓ | __divsc3 | all4 | f32 | f32 | .. | |
| 343 | | ✓ | __divdc3 | all4 | f64 | f64 | .. | |
| 344 | | ✓ | __divtc3 | all4 | f128 | f128 | .. | |
| 345 | | ✓ | __divxc3 | all4 | f80 | f80 | .. | |
| 346 | |
| 347 | [^unused_rl78]: Unused in LLVM, but used for example by rl78. |
| 348 | [^libgcc_compat]: Unused in backends and for symbol-level compatibility with libgcc. |
| 349 | [^noptr_faster]: Operations without pointer and without C struct semantics lead to better optimizations. |
| 350 | [^perf_addition]: Has better performance than standard method due to 2s complement semantics. |
| 351 | Not provided by LLVM and libgcc. |
| 352 | |
| 353 | Decimal float library routines |
| 283 | 354 | |
| 284 | 355 | BID means Binary Integer Decimal encoding, DPD means Densely Packed Decimal encoding. |
| 285 | 356 | BID should be only chosen for binary data, DPD for decimal data (ASCII, Unicode etc). |
| 286 | | If possible, use BCD instead of DPD to represent numbers not accurately representable |
| 287 | | in binary like the number 0.2. |
| 288 | | |
| 289 | | All routines are TODO. |
| 290 | | |
| 291 | | #### Decimal float Conversion |
| 292 | | |
| 293 | | - __dpd_extendsddd2 // dec32->dec64 |
| 294 | | - __bid_extendsddd2 // dec32->dec64 |
| 295 | | - __dpd_extendsdtd2 // dec32->dec128 |
| 296 | | - __bid_extendsdtd2 // dec32->dec128 |
| 297 | | - __dpd_extendddtd2 // dec64->dec128 |
| 298 | | - __bid_extendddtd2 // dec64->dec128 |
| 299 | | - __dpd_truncddsd2 // dec64->dec32 |
| 300 | | - __bid_truncddsd2 // dec64->dec32 |
| 301 | | - __dpd_trunctdsd2 // dec128->dec32 |
| 302 | | - __bid_trunctdsd2 // dec128->dec32 |
| 303 | | - __dpd_trunctddd2 // dec128->dec64 |
| 304 | | - __bid_trunctddd2 // dec128->dec64 |
| 305 | | |
| 306 | | - __dpd_extendsfdd // float->dec64 |
| 307 | | - __bid_extendsfdd // float->dec64 |
| 308 | | - __dpd_extendsftd // float->dec128 |
| 309 | | - __bid_extendsftd // float->dec128 |
| 310 | | - __dpd_extenddftd // double->dec128 |
| 311 | | - __bid_extenddftd // double->dec128 |
| 312 | | - __dpd_extendxftd // long double->dec128 |
| 313 | | - __bid_extendxftd // long double->dec128 |
| 314 | | - __dpd_truncdfsd // double->dec32 |
| 315 | | - __bid_truncdfsd // double->dec32 |
| 316 | | - __dpd_truncxfsd // long double->dec32 |
| 317 | | - __bid_truncxfsd // long double->dec32 |
| 318 | | - __dpd_trunctfsd // long double->dec32 |
| 319 | | - __bid_trunctfsd // long double->dec32 |
| 320 | | - __dpd_truncxfdd // long double->dec64 |
| 321 | | - __bid_truncxfdd // long double->dec64 |
| 322 | | - __dpd_trunctfdd // long double->dec64 |
| 323 | | - __bid_trunctfdd // long double->dec64 |
| 324 | | |
| 325 | | - __dpd_truncddsf // dec64->float |
| 326 | | - __bid_truncddsf // dec64->float |
| 327 | | - __dpd_trunctdsf // dec128->float |
| 328 | | - __bid_trunctdsf // dec128->float |
| 329 | | - __dpd_extendsddf // dec32->double |
| 330 | | - __bid_extendsddf // dec32->double |
| 331 | | - __dpd_trunctddf // dec128->double |
| 332 | | - __bid_trunctddf // dec128->double |
| 333 | | - __dpd_extendsdxf // dec32->long double |
| 334 | | - __bid_extendsdxf // dec32->long double |
| 335 | | - __dpd_extendddxf // dec64->long double |
| 336 | | - __bid_extendddxf // dec64->long double |
| 337 | | - __dpd_trunctdxf // dec128->long double |
| 338 | | - __bid_trunctdxf // dec128->long double |
| 339 | | - __dpd_extendsdtf // dec32->long double |
| 340 | | - __bid_extendsdtf // dec32->long double |
| 341 | | - __dpd_extendddtf // dec64->long double |
| 342 | | - __bid_extendddtf // dec64->long double |
| 343 | | |
| 344 | | Same size conversion: |
| 345 | | - __dpd_extendsfsd // float->dec32 |
| 346 | | - __bid_extendsfsd // float->dec32 |
| 347 | | - __dpd_extenddfdd // double->dec64 |
| 348 | | - __bid_extenddfdd // double->dec64 |
| 349 | | - __dpd_extendtftd //long double->dec128 |
| 350 | | - __bid_extendtftd //long double->dec128 |
| 351 | | - __dpd_truncsdsf // dec32->float |
| 352 | | - __bid_truncsdsf // dec32->float |
| 353 | | - __dpd_truncdddf // dec64->float |
| 354 | | - __bid_truncdddf // dec64->float |
| 355 | | - __dpd_trunctdtf // dec128->long double |
| 356 | | - __bid_trunctdtf // dec128->long double |
| 357 | | |
| 358 | | - __dpd_fixsdsi // dec32->int |
| 359 | | - __bid_fixsdsi // dec32->int |
| 360 | | - __dpd_fixddsi // dec64->int |
| 361 | | - __bid_fixddsi // dec64->int |
| 362 | | - __dpd_fixtdsi // dec128->int |
| 363 | | - __bid_fixtdsi // dec128->int |
| 364 | | |
| 365 | | - __dpd_fixsddi // dec32->long |
| 366 | | - __bid_fixsddi // dec32->long |
| 367 | | - __dpd_fixdddi // dec64->long |
| 368 | | - __bid_fixdddi // dec64->long |
| 369 | | - __dpd_fixtddi // dec128->long |
| 370 | | - __bid_fixtddi // dec128->long |
| 371 | | |
| 372 | | - __dpd_fixunssdsi // dec32->unsigned int, All negative values become zero. |
| 373 | | - __bid_fixunssdsi // dec32->unsigned int |
| 374 | | - __dpd_fixunsddsi // dec64->unsigned int |
| 375 | | - __bid_fixunsddsi // dec64->unsigned int |
| 376 | | - __dpd_fixunstdsi // dec128->unsigned int |
| 377 | | - __bid_fixunstdsi // dec128->unsigned int |
| 378 | | |
| 379 | | - __dpd_fixunssddi // dec32->unsigned long, All negative values become zero. |
| 380 | | - __bid_fixunssddi // dec32->unsigned long |
| 381 | | - __dpd_fixunsdddi // dec64->unsigned long |
| 382 | | - __bid_fixunsdddi // dec64->unsigned long |
| 383 | | - __dpd_fixunstddi // dec128->unsigned long |
| 384 | | - __bid_fixunstddi // dec128->unsigned long |
| 385 | | |
| 386 | | - __dpd_floatsisd // int->dec32 |
| 387 | | - __bid_floatsisd // int->dec32 |
| 388 | | - __dpd_floatsidd // int->dec64 |
| 389 | | - __bid_floatsidd // int->dec64 |
| 390 | | - __dpd_floatsitd // int->dec128 |
| 391 | | - __bid_floatsitd // int->dec128 |
| 392 | | |
| 393 | | - __dpd_floatdisd // long->dec32 |
| 394 | | - __bid_floatdisd // long->dec32 |
| 395 | | - __dpd_floatdidd // long->dec64 |
| 396 | | - __bid_floatdidd // long->dec64 |
| 397 | | - __dpd_floatditd // long->dec128 |
| 398 | | - __bid_floatditd // long->dec128 |
| 399 | | |
| 400 | | - __dpd_floatunssisd // unsigned int->dec32 |
| 401 | | - __bid_floatunssisd // unsigned int->dec32 |
| 402 | | - __dpd_floatunssidd // unsigned int->dec64 |
| 403 | | - __bid_floatunssidd // unsigned int->dec64 |
| 404 | | - __dpd_floatunssitd // unsigned int->dec128 |
| 405 | | - __bid_floatunssitd // unsigned int->dec128 |
| 406 | | |
| 407 | | - __dpd_floatunsdisd // unsigned long->dec32 |
| 408 | | - __bid_floatunsdisd // unsigned long->dec32 |
| 409 | | - __dpd_floatunsdidd // unsigned long->dec64 |
| 410 | | - __bid_floatunsdidd // unsigned long->dec64 |
| 411 | | - __dpd_floatunsditd // unsigned long->dec128 |
| 412 | | - __bid_floatunsditd // unsigned long->dec128 |
| 413 | | |
| 414 | | #### Decimal float Comparison |
| 415 | | |
| 416 | | All decimal float comparison routines return c_int. |
| 417 | | |
| 418 | | - __dpd_unordsd2 // a,b: dec32, a +-NaN or a +-NaN -> 1(nonzero), else -> 0 |
| 419 | | - __bid_unordsd2 // a,b: dec32 |
| 420 | | - __dpd_unorddd2 // a,b: dec64 |
| 421 | | - __bid_unorddd2 // a,b: dec64 |
| 422 | | - __dpd_unordtd2 // a,b: dec128 |
| 423 | | - __bid_unordtd2 // a,b: dec128 |
| 424 | | |
| 425 | | - __dpd_eqsd2 // a,b: dec32, a!=+-NaN and b!=+-Nan and a==b -> 0, else -> 1(nonzero) |
| 426 | | - __bid_eqsd2 // a,b: dec32 |
| 427 | | - __dpd_eqdd2 // a,b: dec64 |
| 428 | | - __bid_eqdd2 // a,b: dec64 |
| 429 | | - __dpd_eqtd2 // a,b: dec128 |
| 430 | | - __bid_eqtd2 // a,b: dec128 |
| 431 | | |
| 432 | | - __dpd_nesd2 // a,b: dec32, a==+-NaN or b==+-NaN or a!=b -> 1(nonzero), else -> 0 |
| 433 | | - __bid_nesd2 // a,b: dec32 |
| 434 | | - __dpd_nedd2 // a,b: dec64 |
| 435 | | - __bid_nedd2 // a,b: dec64 |
| 436 | | - __dpd_netd2 // a,b: dec128 |
| 437 | | - __bid_netd2 // a,b: dec128 |
| 438 | | |
| 439 | | - __dpd_gesd2 // a,b: dec32, a!=+-NaN and b!=+-NaN and a>=b -> >=0, else -> <0 |
| 440 | | - __bid_gesd2 // a,b: dec32 |
| 441 | | - __dpd_gedd2 // a,b: dec64 |
| 442 | | - __bid_gedd2 // a,b: dec64 |
| 443 | | - __dpd_getd2 // a,b: dec128 |
| 444 | | - __bid_getd2 // a,b: dec128 |
| 445 | | |
| 446 | | - __dpd_ltsd2 // a,b: dec32, a!=+-NaN and b!=+-NaN and a<b -> <0, else -> >=0 |
| 447 | | - __bid_ltsd2 // a,b: dec32 |
| 448 | | - __dpd_ltdd2 // a,b: dec64 |
| 449 | | - __bid_ltdd2 // a,b: dec64 |
| 450 | | - __dpd_lttd2 // a,b: dec128 |
| 451 | | - __bid_lttd2 // a,b: dec128 |
| 452 | | |
| 453 | | - __dpd_lesd2 // a,b: dec32, a!=+-NaN and b!=+-NaN and a<=b -> <=0, else -> >=0 |
| 454 | | - __bid_lesd2 // a,b: dec32 |
| 455 | | - __dpd_ledd2 // a,b: dec64 |
| 456 | | - __bid_ledd2 // a,b: dec64 |
| 457 | | - __dpd_letd2 // a,b: dec128 |
| 458 | | - __bid_letd2 // a,b: dec128 |
| 459 | | |
| 460 | | - __dpd_gtsd2 // a,b: dec32, a!=+-NaN and b!=+-NaN and a>b -> >0, else -> <=0 |
| 461 | | - __bid_gtsd2 // a,b: dec32 |
| 462 | | - __dpd_gtdd2 // a,b: dec64 |
| 463 | | - __bid_gtdd2 // a,b: dec64 |
| 464 | | - __dpd_gttd2 // a,b: dec128 |
| 465 | | - __bid_gttd2 // a,b: dec128 |
| 466 | | |
| 467 | | #### Decimal float Arithmetic |
| 468 | | |
| 469 | | These numbers include options with routines for +-0 and +-Nan. |
| 470 | | |
| 471 | | - __dpd_addsd3 // a,b: dec32 -> dec32, a + b |
| 472 | | - __bid_addsd3 // a,b: dec32 -> dec32 |
| 473 | | - __dpd_adddd3 // a,b: dec64 -> dec64 |
| 474 | | - __bid_adddd3 // a,b: dec64 -> dec64 |
| 475 | | - __dpd_addtd3 // a,b: dec128-> dec128 |
| 476 | | - __bid_addtd3 // a,b: dec128-> dec128 |
| 477 | | - __dpd_subsd3 // a,b: dec32, a - b |
| 478 | | - __bid_subsd3 // a,b: dec32 -> dec32 |
| 479 | | - __dpd_subdd3 // a,b: dec64 .. |
| 480 | | - __bid_subdd3 // a,b: dec64 |
| 481 | | - __dpd_subtd3 // a,b: dec128 |
| 482 | | - __bid_subtd3 // a,b: dec128 |
| 483 | | - __dpd_mulsd3 // a,b: dec32, a * b |
| 484 | | - __bid_mulsd3 // a,b: dec32 -> dec32 |
| 485 | | - __dpd_muldd3 // a,b: dec64 .. |
| 486 | | - __bid_muldd3 // a,b: dec64 |
| 487 | | - __dpd_multd3 // a,b: dec128 |
| 488 | | - __bid_multd3 // a,b: dec128 |
| 489 | | - __dpd_divsd3 // a,b: dec32, a / b |
| 490 | | - __bid_divsd3 // a,b: dec32 -> dec32 |
| 491 | | - __dpd_divdd3 // a,b: dec64 .. |
| 492 | | - __bid_divdd3 // a,b: dec64 |
| 493 | | - __dpd_divtd3 // a,b: dec128 |
| 494 | | - __bid_divtd3 // a,b: dec128 |
| 495 | | - __dpd_negsd2 // a,b: dec32, -a |
| 496 | | - __bid_negsd2 // a,b: dec32 -> dec32 |
| 497 | | - __dpd_negdd2 // a,b: dec64 .. |
| 498 | | - __bid_negdd2 // a,b: dec64 |
| 499 | | - __dpd_negtd2 // a,b: dec128 |
| 500 | | - __bid_negtd2 // a,b: dec128 |
| 501 | | |
| 502 | | ## Fixed-point fractional library routines |
| 503 | | |
| 504 | | TODO |
| 505 | | |
| 506 | | Too unclear for work items: |
| 507 | | - Miscellaneous routines => unclear, if supported (cache control and stack functions) |
| 508 | | - Zig-specific language runtime features, for example "Arbitrary length integer library routines" |
| 357 | For example the number 0.2 is not accurately representable in binary data. |
| 358 | |
| 359 | | Done | Name | a | b | Out | Comment | |
| 360 | | ------ | ------------- | --------- | --------- | --------- | ---------------------------- | |
| 361 | | | | | | | **Decimal Float Conversion** | |
| 362 | | ✗ | __dpd_extendsddd2 | dec32 | ∅ | dec64 | conversion | |
| 363 | | ✗ | __bid_extendsddd2 | dec32 | ∅ | dec64 | .. | |
| 364 | | ✗ | __dpd_extendsdtd2 | dec32 | ∅ | dec128| .. | |
| 365 | | ✗ | __bid_extendsdtd2 | dec32 | ∅ | dec128| .. | |
| 366 | | ✗ | __dpd_extendddtd2 | dec64 | ∅ | dec128| .. | |
| 367 | | ✗ | __bid_extendddtd2 | dec64 | ∅ | dec128| .. | |
| 368 | | ✗ | __dpd_truncddsd2 | dec64 | ∅ | dec32 | .. | |
| 369 | | ✗ | __bid_truncddsd2 | dec64 | ∅ | dec32 | .. | |
| 370 | | ✗ | __dpd_trunctdsd2 | dec128 | ∅ | dec32 | .. | |
| 371 | | ✗ | __bid_trunctdsd2 | dec128 | ∅ | dec32 | .. | |
| 372 | | ✗ | __dpd_trunctddd2 | dec128 | ∅ | dec64 | .. | |
| 373 | | ✗ | __bid_trunctddd2 | dec128 | ∅ | dec64 | .. | |
| 374 | | ✗ | __dpd_extendsfdd | float | ∅ | dec64 | .. | |
| 375 | | ✗ | __bid_extendsfdd | float | ∅ | dec64 | .. | |
| 376 | | ✗ | __dpd_extendsftd | float | ∅ | dec128| .. | |
| 377 | | ✗ | __bid_extendsftd | float | ∅ | dec128| .. | |
| 378 | | ✗ | __dpd_extenddftd | double | ∅ | dec128| .. | |
| 379 | | ✗ | __bid_extenddftd | double | ∅ | dec128| .. | |
| 380 | | ✗ | __dpd_extendxftd |long double | ∅ | dec128| .. | |
| 381 | | ✗ | __bid_extendxftd |long double | ∅ | dec128| .. | |
| 382 | | ✗ | __dpd_truncdfsd | double | ∅ | dec32 | .. | |
| 383 | | ✗ | __bid_truncdfsd | double | ∅ | dec32 | .. | |
| 384 | | ✗ | __dpd_truncxfsd |long double | ∅ | dec32 | .. | |
| 385 | | ✗ | __bid_truncxfsd |long double | ∅ | dec32 | .. | |
| 386 | | ✗ | __dpd_trunctfsd |long double | ∅ | dec32 | .. | |
| 387 | | ✗ | __bid_trunctfsd |long double | ∅ | dec32 | .. | |
| 388 | | ✗ | __dpd_truncxfdd |long double | ∅ | dec64 | .. | |
| 389 | | ✗ | __bid_truncxfdd |long double | ∅ | dec64 | .. | |
| 390 | | ✗ | __dpd_trunctfdd |long double | ∅ | dec64 | .. | |
| 391 | | ✗ | __bid_trunctfdd |long double | ∅ | dec64 | .. | |
| 392 | | ✗ | __dpd_truncddsf | dec64 | ∅ | float | .. | |
| 393 | | ✗ | __bid_truncddsf | dec64 | ∅ | float | .. | |
| 394 | | ✗ | __dpd_trunctdsf | dec128 | ∅ | float | .. | |
| 395 | | ✗ | __bid_trunctdsf | dec128 | ∅ | float | .. | |
| 396 | | ✗ | __dpd_extendsddf | dec32 | ∅ | double| .. | |
| 397 | | ✗ | __bid_extendsddf | dec32 | ∅ | double| .. | |
| 398 | | ✗ | __dpd_trunctddf | dec128 | ∅ | double| .. | |
| 399 | | ✗ | __bid_trunctddf | dec128 | ∅ | double| .. | |
| 400 | | ✗ | __dpd_extendsdxf | dec32 | ∅ |long double| .. | |
| 401 | | ✗ | __bid_extendsdxf | dec32 | ∅ |long double| .. | |
| 402 | | ✗ | __dpd_extendddxf | dec64 | ∅ |long double| .. | |
| 403 | | ✗ | __bid_extendddxf | dec64 | ∅ |long double| .. | |
| 404 | | ✗ | __dpd_trunctdxf | dec128 | ∅ |long double| .. | |
| 405 | | ✗ | __bid_trunctdxf | dec128 | ∅ |long double| .. | |
| 406 | | ✗ | __dpd_extendsdtf | dec32 | ∅ |long double| .. | |
| 407 | | ✗ | __bid_extendsdtf | dec32 | ∅ |long double| .. | |
| 408 | | ✗ | __dpd_extendddtf | dec64 | ∅ |long double| .. | |
| 409 | | ✗ | __bid_extendddtf | dec64 | ∅ |long double| .. | |
| 410 | | ✗ | __dpd_extendsfsd | float | ∅ | dec32 | same size conversions | |
| 411 | | ✗ | __bid_extendsfsd | float | ∅ | dec32 | .. | |
| 412 | | ✗ | __dpd_extenddfdd | double | ∅ | dec64 | .. | |
| 413 | | ✗ | __bid_extenddfdd | double | ∅ | dec64 | .. | |
| 414 | | ✗ | __dpd_extendtftd |long double | ∅ | dec128| .. | |
| 415 | | ✗ | __bid_extendtftd |long double | ∅ | dec128| .. | |
| 416 | | ✗ | __dpd_truncsdsf | dec32 | ∅ | float | .. | |
| 417 | | ✗ | __bid_truncsdsf | dec32 | ∅ | float | .. | |
| 418 | | ✗ | __dpd_truncdddf | dec64 | ∅ | float | conversion | |
| 419 | | ✗ | __bid_truncdddf | dec64 | ∅ | float | .. | |
| 420 | | ✗ | __dpd_trunctdtf | dec128 | ∅ |long double| .. | |
| 421 | | ✗ | __bid_trunctdtf | dec128 | ∅ |long double| .. | |
| 422 | | ✗ | __dpd_fixsdsi | dec32 | ∅ | int | .. | |
| 423 | | ✗ | __bid_fixsdsi | dec32 | ∅ | int | .. | |
| 424 | | ✗ | __dpd_fixddsi | dec64 | ∅ | int | .. | |
| 425 | | ✗ | __bid_fixddsi | dec64 | ∅ | int | .. | |
| 426 | | ✗ | __dpd_fixtdsi | dec128 | ∅ | int | .. | |
| 427 | | ✗ | __bid_fixtdsi | dec128 | ∅ | int | .. | |
| 428 | | ✗ | __dpd_fixsddi | dec32 | ∅ | long | .. | |
| 429 | | ✗ | __bid_fixsddi | dec32 | ∅ | long | .. | |
| 430 | | ✗ | __dpd_fixdddi | dec64 | ∅ | long | .. | |
| 431 | | ✗ | __bid_fixdddi | dec64 | ∅ | long | .. | |
| 432 | | ✗ | __dpd_fixtddi | dec128 | ∅ | long | .. | |
| 433 | | ✗ | __bid_fixtddi | dec128 | ∅ | long | .. | |
| 434 | | ✗ | __dpd_fixunssdsi | dec32 | ∅ |unsigned int | .. All negative values become zero. | |
| 435 | | ✗ | __bid_fixunssdsi | dec32 | ∅ |unsigned int | .. | |
| 436 | | ✗ | __dpd_fixunsddsi | dec64 | ∅ |unsigned int | .. | |
| 437 | | ✗ | __bid_fixunsddsi | dec64 | ∅ |unsigned int | .. | |
| 438 | | ✗ | __dpd_fixunstdsi | dec128 | ∅ |unsigned int | .. | |
| 439 | | ✗ | __bid_fixunstdsi | dec128 | ∅ |unsigned int | .. | |
| 440 | | ✗ | __dpd_fixunssddi | dec32 | ∅ |unsigned long| .. | |
| 441 | | ✗ | __bid_fixunssddi | dec32 | ∅ |unsigned long| .. | |
| 442 | | ✗ | __dpd_fixunsdddi | dec64 | ∅ |unsigned long| .. | |
| 443 | | ✗ | __bid_fixunsdddi | dec64 | ∅ |unsigned long| .. | |
| 444 | | ✗ | __dpd_fixunstddi | dec128 | ∅ |unsigned long| .. | |
| 445 | | ✗ | __bid_fixunstddi | dec128 | ∅ |unsigned long| .. | |
| 446 | | ✗ | __dpd_floatsisd | int | ∅ | dec32 | .. | |
| 447 | | ✗ | __bid_floatsisd | int | ∅ | dec32 | .. | |
| 448 | | ✗ | __dpd_floatsidd | int | ∅ | dec64 | .. | |
| 449 | | ✗ | __bid_floatsidd | int | ∅ | dec64 | .. | |
| 450 | | ✗ | __dpd_floatsitd | int | ∅ | dec128 | .. | |
| 451 | | ✗ | __bid_floatsitd | int | ∅ | dec128 | .. | |
| 452 | | ✗ | __dpd_floatdisd | long | ∅ | dec32 | .. | |
| 453 | | ✗ | __bid_floatdisd | long | ∅ | dec32 | .. | |
| 454 | | ✗ | __dpd_floatdidd | long | ∅ | dec64 | .. | |
| 455 | | ✗ | __bid_floatdidd | long | ∅ | dec64 | .. | |
| 456 | | ✗ | __dpd_floatditd | long | ∅ | dec128 | .. | |
| 457 | | ✗ | __bid_floatditd | long | ∅ | dec128 | .. | |
| 458 | | ✗ | __dpd_floatunssisd | unsigned int| ∅ | dec32 | .. | |
| 459 | | ✗ | __bid_floatunssisd | unsigned int| ∅ | dec32 | .. | |
| 460 | | ✗ | __dpd_floatunssidd | unsigned int| ∅ | dec64 | .. | |
| 461 | | ✗ | __bid_floatunssidd | unsigned int| ∅ | dec64 | .. | |
| 462 | | ✗ | __dpd_floatunssitd | unsigned int| ∅ | dec128 | .. | |
| 463 | | ✗ | __bid_floatunssitd | unsigned int| ∅ | dec128 | .. | |
| 464 | | ✗ | __dpd_floatunsdisd |unsigned long| ∅ | dec32 | .. | |
| 465 | | ✗ | __bid_floatunsdisd |unsigned long| ∅ | dec32 | .. | |
| 466 | | ✗ | __dpd_floatunsdidd |unsigned long| ∅ | dec64 | .. | |
| 467 | | ✗ | __bid_floatunsdidd |unsigned long| ∅ | dec64 | .. | |
| 468 | | ✗ | __dpd_floatunsditd |unsigned long| ∅ | dec128 | .. | |
| 469 | | ✗ | __bid_floatunsditd |unsigned long| ∅ | dec128 | .. | |
| 470 | | | | | | | **Decimal Float Comparison** | |
| 471 | | ✗ | __dpd_unordsd2 | dec32 | dec32 | c_int | `a +-NaN or a +-NaN -> 1(nonzero), else -> 0` | |
| 472 | | ✗ | __bid_unordsd2 | dec32 | dec32 | c_int | .. | |
| 473 | | ✗ | __dpd_unorddd2 | dec64 | dec64 | c_int | .. | |
| 474 | | ✗ | __bid_unorddd2 | dec64 | dec64 | c_int | .. | |
| 475 | | ✗ | __dpd_unordtd2 | dec128 | dec128 | c_int | .. | |
| 476 | | ✗ | __bid_unordtd2 | dec128 | dec128 | c_int | .. | |
| 477 | | ✗ | __dpd_eqsd2 | dec32 | dec32 | c_int |`a!=+-NaN and b!=+-Nan and a==b -> 0, else -> 1(nonzero)`| |
| 478 | | ✗ | __bid_eqsd2 | dec32 | dec32 | c_int | .. | |
| 479 | | ✗ | __dpd_eqdd2 | dec64 | dec64 | c_int | .. | |
| 480 | | ✗ | __bid_eqdd2 | dec64 | dec64 | c_int | .. | |
| 481 | | ✗ | __dpd_eqtd2 | dec128 | dec128 | c_int | .. | |
| 482 | | ✗ | __bid_eqtd2 | dec128 | dec128 | c_int | .. | |
| 483 | | ✗ | __dpd_nesd2 | dec32 | dec32 | c_int | `a==+-NaN or b==+-NaN or a!=b -> 1(nonzero), else -> 0` | |
| 484 | | ✗ | __bid_nesd2 | dec32 | dec32 | c_int | .. | |
| 485 | | ✗ | __dpd_nedd2 | dec64 | dec64 | c_int | .. | |
| 486 | | ✗ | __bid_nedd2 | dec64 | dec64 | c_int | .. | |
| 487 | | ✗ | __dpd_netd2 | dec128 | dec128 | c_int | .. | |
| 488 | | ✗ | __bid_netd2 | dec128 | dec128 | c_int | .. | |
| 489 | | ✗ | __dpd_gesd2 | dec32 | dec32 | c_int | `a!=+-NaN and b!=+-NaN and a>=b -> >=0, else -> <0` | |
| 490 | | ✗ | __bid_gesd2 | dec32 | dec32 | c_int | .. | |
| 491 | | ✗ | __dpd_gedd2 | dec64 | dec64 | c_int | .. | |
| 492 | | ✗ | __bid_gedd2 | dec64 | dec64 | c_int | .. | |
| 493 | | ✗ | __dpd_getd2 | dec128 | dec128 | c_int | .. | |
| 494 | | ✗ | __bid_getd2 | dec128 | dec128 | c_int | .. | |
| 495 | | ✗ | __dpd_ltsd2 | dec32 | dec32 | c_int | `a!=+-NaN and b!=+-NaN and a<b -> <0, else -> >=0` | |
| 496 | | ✗ | __bid_ltsd2 | dec32 | dec32 | c_int | .. | |
| 497 | | ✗ | __dpd_ltdd2 | dec64 | dec64 | c_int | .. | |
| 498 | | ✗ | __bid_ltdd2 | dec64 | dec64 | c_int | .. | |
| 499 | | ✗ | __dpd_lttd2 | dec128 | dec128 | c_int | .. | |
| 500 | | ✗ | __bid_lttd2 | dec128 | dec128 | c_int | .. | |
| 501 | | ✗ | __dpd_lesd2 | dec32 | dec32 | c_int | `a!=+-NaN and b!=+-NaN and a<=b -> <=0, else -> >=0` | |
| 502 | | ✗ | __bid_lesd2 | dec32 | dec32 | c_int | .. | |
| 503 | | ✗ | __dpd_ledd2 | dec64 | dec64 | c_int | .. | |
| 504 | | ✗ | __bid_ledd2 | dec64 | dec64 | c_int | .. | |
| 505 | | ✗ | __dpd_letd2 | dec128 | dec128 | c_int | .. | |
| 506 | | ✗ | __bid_letd2 | dec128 | dec128 | c_int | .. | |
| 507 | | ✗ | __dpd_gtsd2 | dec32 | dec32 | c_int | `a!=+-NaN and b!=+-NaN and a>b -> >0, else -> <=0` | |
| 508 | | ✗ | __bid_gtsd2 | dec32 | dec32 | c_int | .. | |
| 509 | | ✗ | __dpd_gtdd2 | dec64 | dec64 | c_int | .. | |
| 510 | | ✗ | __bid_gtdd2 | dec64 | dec64 | c_int | .. | |
| 511 | | ✗ | __dpd_gttd2 | dec128 | dec128 | c_int | .. | |
| 512 | | ✗ | __bid_gttd2 | dec128 | dec128 | c_int | .. | |
| 513 | | | | | | | **Decimal Float Arithmetic**[^options] | |
| 514 | | ✗ | __dpd_addsd3 | dec32 | dec32 | dec32 |`a + b`| |
| 515 | | ✗ | __bid_addsd3 | dec32 | dec32 | dec32 | .. | |
| 516 | | ✗ | __dpd_adddd3 | dec64 | dec64 | dec64 | .. | |
| 517 | | ✗ | __bid_adddd3 | dec64 | dec64 | dec64 | .. | |
| 518 | | ✗ | __dpd_addtd3 | dec128 | dec128 | dec128 | .. | |
| 519 | | ✗ | __bid_addtd3 | dec128 | dec128 | dec128 | .. | |
| 520 | | ✗ | __dpd_subsd3 | dec32 | dec32 | dec32 |`a - b`| |
| 521 | | ✗ | __bid_subsd3 | dec32 | dec32 | dec32 | .. | |
| 522 | | ✗ | __dpd_subdd3 | dec64 | dec64 | dec64 | .. | |
| 523 | | ✗ | __bid_subdd3 | dec64 | dec64 | dec64 | .. | |
| 524 | | ✗ | __dpd_subtd3 | dec128 | dec128 | dec128 | .. | |
| 525 | | ✗ | __bid_subtd3 | dec128 | dec128 | dec128 | .. | |
| 526 | | ✗ | __dpd_mulsd3 | dec32 | dec32 | dec32 |`a * b`| |
| 527 | | ✗ | __bid_mulsd3 | dec32 | dec32 | dec32 | .. | |
| 528 | | ✗ | __dpd_muldd3 | dec64 | dec64 | dec64 | .. | |
| 529 | | ✗ | __bid_muldd3 | dec64 | dec64 | dec64 | .. | |
| 530 | | ✗ | __dpd_multd3 | dec128 | dec128 | dec128 | .. | |
| 531 | | ✗ | __bid_multd3 | dec128 | dec128 | dec128 | .. | |
| 532 | | ✗ | __dpd_divsd3 | dec32 | dec32 | dec32 |`a / b`| |
| 533 | | ✗ | __bid_divsd3 | dec32 | dec32 | dec32 | .. | |
| 534 | | ✗ | __dpd_divdd3 | dec64 | dec64 | dec64 | .. | |
| 535 | | ✗ | __bid_divdd3 | dec64 | dec64 | dec64 | .. | |
| 536 | | ✗ | __dpd_divtd3 | dec128 | dec128 | dec128 | .. | |
| 537 | | ✗ | __bid_divtd3 | dec128 | dec128 | dec128 | .. | |
| 538 | | ✗ | __dpd_negsd2 | dec32 | dec32 | dec32 | `-a` | |
| 539 | | ✗ | __bid_negsd2 | dec32 | dec32 | dec32 | .. | |
| 540 | | ✗ | __dpd_negdd2 | dec64 | dec64 | dec64 | .. | |
| 541 | | ✗ | __bid_negdd2 | dec64 | dec64 | dec64 | .. | |
| 542 | | ✗ | __dpd_negtd2 | dec128 | dec128 | dec128 | .. | |
| 543 | | ✗ | __bid_negtd2 | dec128 | dec128 | dec128 | .. | |
| 544 | |
| 545 | [^options]: These numbers include options with routines for +-0 and +-Nan. |
| 546 | |
| 547 | Fixed-point fractional library routines |
| 548 | |
| 549 | TODO brief explanation + implementation |
| 550 | |
| 551 | | Done | Name | a | b | Out | Comment | |
| 552 | | ------ | ------------- | --------- | --------- | --------- | -------------------------- | |
| 553 | | | | | | | **Fixed-Point Fractional** | |
| 554 | |
| 555 | Further content: |
| 556 | - aarch64 outline atomics |
| 557 | - atomics |
| 558 | - msvc things like _alldiv, _aulldiv, _allrem |
| 559 | - clear cache |
| 560 | - tls emulation |
| 561 | - math routines (cos, sin, tan, ceil, floor, exp, exp2, fabs, log, log10, log2, sincos, sqrt) |
| 562 | - bcmp |
| 563 | - ieee float routines (fma, fmax, fmin, fmod, fabs, float rounding, ) |
| 564 | - arm routines (memory routines + memclr [setting to 0], divmod routines and stubs for unwind_cpp) |
| 565 | - memory routines (memcmp, memcpy, memset, memmove) |
| 566 | - objective-c __isPlatformVersionAtLeast check |
| 567 | - stack probe routines |
| 568 | |
| 569 | Future work |
| 570 | |
| 571 | Arbitrary length integer library routines |
Jan Philipp Hafer <jan.hafer@rwth-aachen.de> 2023-01-26 22:44:39+01:00
Andrew Kelley <andrew@ziglang.org> 2023-02-02 17:45:38-05:00
lib