Home
last modified time | relevance | path

Searched refs:denormal (Results 1 – 13 of 13) sorted by relevance

/freebsd/contrib/llvm-project/compiler-rt/lib/builtins/
H A Dfp_mul_impl.inc25 // Detect if a or b is zero, denormal, infinity, or NaN.
64 // One or both of a or b is denormal. The other (if applicable) is a
74 // denormal path it was already set by normalize( ), but setting it twice
98 // The result is denormal before rounding.
H A Dfp_add_impl.inc86 // implicit significand bit. If we fell through from the denormal path it
133 // The result is denormal before rounding. The exponent is zero and we
H A Dfp_div_impl.inc45 // Detect if a or b is zero, denormal, infinity, or NaN.
84 // One or both of a or b is denormal. The other (if applicable) is a
94 // denormal path it was already set by normalize( ), but setting it twice
359 // It is re-computed on denormal path that is expected to be not so
/freebsd/contrib/arm-optimized-routines/math/test/testcases/directed/
H A Dexpm1f.tst34 ; Check values for which exp goes denormal. expm1f should not report
/freebsd/contrib/gdtoa/
H A Dchanges119 that caused one-bit rounding errors in certain denormal numbers, such
128 2.2250738585072012e-308, which was rounded to the largest denormal
192 it returned 0 rather than the smallest denormal.
334 the smallest denormal).
373 denormal result should be zero, correct logic for rounding up when the
374 result is denormal, and when returning zero or Infinity, set *bp = 0;
497 NO_ERRNO being defined) set ERANGE for denormal values as well as real
599 handling of numbers that will scale to denormal values. The increments
/freebsd/contrib/netbsd-tests/lib/libm/
H A Dt_ldexp.c444 TEST(ldexp_denormal, denormal) in TEST()
/freebsd/cddl/contrib/opensolaris/lib/libdtrace/common/
H A Ddt_cc.c369 int denormal = (strcmp(dnp->dn_ident->di_name, "denormalize") == 0); in dt_action_normalize() local
377 if ((denormal && argc != 1) || (!denormal && argc != 2)) { in dt_action_normalize()
380 dnp->dn_ident->di_name, argc, denormal ? 1 : 2); in dt_action_normalize()
410 if (denormal) { in dt_action_normalize()
/freebsd/contrib/llvm-project/llvm/include/llvm/IR/
H A DAttributes.td357 def DenormalFPMath : ComplexStrAttr<"denormal-fp-math", [FnAttr]>;
358 def DenormalFPMathF32 : ComplexStrAttr<"denormal-fp-math-f32", [FnAttr]>;
H A DIntrinsicsAMDGPU.td3073 // Emit 2.5 ulp, no denormal division. Should only be inserted by
/freebsd/contrib/llvm-project/llvm/lib/Target/AMDGPU/
H A DAMDGPU.td32 def FeatureFastDenormalF32 : SubtargetFeature<"fast-denormal-f32",
799 : SubtargetFeature<"memory-atomic-fadd-f32-denormal-support",
802 "global/flat/buffer atomic fadd for float supports denormal handling"
H A DSIInstructions.td3064 // On pre-gfx9 targets, v_max_*/v_min_* did not respect the denormal
/freebsd/contrib/llvm-project/clang/include/clang/Driver/
H A DOptions.td1261 HelpText<"Flush denormal floating point values to zero in CUDA/HIP device mode.">;
5073 "respect rounding and denormal modes, and may give incorrect results "
/freebsd/contrib/llvm-project/llvm/lib/Target/PowerPC/
H A DPPCInstrInfo.td409 // to a single precision floating point non-denormal immediate without loss of