interpret floating point binop instructions

[can-lehmann]

No description provided.

[github-actions]

Coverage

File	Lines	Functions	Regions	Branches
genext.hpp	🔴 54.42%	🟡 83.33%	🟡 76.14%	🟡 76.81%
jitir.hpp	🔴 51.20%	🔴 61.44%	🔴 55.81%	🔴 54.97%
jitir_llvmapi.hpp	🔴 56.19%	🔴 66.67%	🟢 92.98%	🔴 66.67%
llvmgen.hpp	🔴 64.40%	🟡 70.00%	🟡 83.71%	🟡 83.90%
lowerllvm.hpp	🔴 63.60%	🟡 75.00%	🔴 41.32%	🔴 56.49%
x86gen.hpp	🟡 74.12%	🟡 81.82%	🟡 75.78%	🟡 83.30%
x86insts.inc.hpp	-	-	-	-
TOTAL	🔴 56.21%	🔴 65.73%	🔴 64.26%	🔴 65.46%

Diff Coverage

Diff Coverage

Diff: origin/main...HEAD, staged and unstaged changes

• jitir.hpp (87.5%): Missing lines 293,6316,6321,6323-6324
• llvmgen.hpp (100%)

Summary

• Total: 50 lines
• Missing: 5 lines
• Coverage: 90%

jitir.hpp

Lines 289-297

  289   }
  290 
  291   inline Type int_type_of_size(size_t size) {
  292     return int_type_of_width(size * 8);
! 293   }
  294 
  295   template <class T, class S>
  296   inline T bit_cast(S value) {
  297     static_assert(sizeof(T) == sizeof(S));

---

Lines 6312-6328

  6312             case Type::Float32: { \
  6313               float32_t a = bit_cast<float32_t>((uint32_t) _a); \
  6314               float32_t b = bit_cast<float32_t>((uint32_t) _b); \
  6315               return Bits::constant(expr); \
! 6316             } \
  6317             case Type::Float64: { \
  6318               float64_t a = bit_cast<float64_t>(_a); \
  6319               float64_t b = bit_cast<float64_t>(_b); \
  6320               return Bits::constant(expr); \
! 6321             } \
  6322             default: \
! 6323               assert(false && "Unsupported type"); \
! 6324               return Bits::constant(type, 0); \
  6325           } \
  6326         }
  6327       
  6328       float_binop(add_f, a + b)

---