You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: llvm/docs/SPIRVUsage.rst
+6Lines changed: 6 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -207,6 +207,8 @@ list of supported SPIR-V extensions, sorted alphabetically by their extension na
207
207
- Allows support for additional group operations within uniform control flow.
208
208
* - ``SPV_KHR_non_semantic_info``
209
209
- Adds the ability to declare extended instruction sets that have no semantic impact and can be safely removed from a module.
210
+
* - ``SPV_INTEL_fp_max_error``
211
+
- Adds the ability to specify the maximum error for floating-point operations.
210
212
211
213
To enable multiple extensions, list them separated by comma. For example, to enable support for atomic operations on floating-point numbers and arbitrary precision integers, use:
212
214
@@ -307,6 +309,10 @@ SPIR-V backend, along with their descriptions and argument details.
307
309
- None
308
310
- `[Type, 32-bit Integer, Metadata]`
309
311
- Assigns one of two memory aliasing decorations (specified by the second argument) to instructions using original aliasing metadata node. Not emitted directly but used to support SPIR-V representation in LLVM IR.
312
+
* - `int_spv_assign_fpmaxerror_decoration`
313
+
- None
314
+
- `[Type, Metadata]`
315
+
- Assigns the maximum error decoration to floating-point instructions using the original metadata node. Not emitted directly but used to support SPIR-V representation in LLVM IR.
; Confirm that we handle fpmath metadata correctly
2
+
; This is a copy of https://github.com/KhronosGroup/SPIRV-LLVM-Translator/blob/main/test/extensions/INTEL/SPV_INTEL_fp_max_error/IntelFPMaxErrorFPMath.ll
0 commit comments