Subtract with Borrow-In Arithmetic
Subtract two operands minus a borrow-in from a previous subtract-with-borrow-out, for extended-precision subtraction chains.
Vendor-Neutral Definition
d = a - b - borrow_in (mod 2^width); optionally also produces its own borrow-out for further chaining.
⚠️ Cross-vendor mappings describe semantic relationships and are not guaranteed one-to-one compiler
translations.
equivalent with restrictions
Both subtract two operands minus an incoming borrow bit, for chaining extended-precision subtraction across multiple words.
| Property | NVIDIA PTX | AMDGPU |
|---|---|---|
| Instruction | subc | v_subb_co_u32, s_subb_u32 |
| ISA Layer | Virtual | Native |
| Data Types | - | u32 |
| Version / Target Introduced | PTX ISA 1.2 | gfx1100 |
Important Differences
- Same CC-register-vs-VCC/SCC convention difference as add-with-carry-out applies here.
Source Evidence
- Parallel Thread Execution ISA ↗ - NVIDIA Corporation
- User Guide for AMDGPU Backend ↗ - LLVM Project
Verification method: documentation