Lines Matching refs:VFP

38 // FP loads/stores/moves, shared between VFP and MVE (even in the integer-only
43 // 16-bit FP loads/stores/moves, shared between VFP (with the v8.2A FP16
63 /// Versions of the VFP flags restricted to single precision, or to
236 // Cyclone can zero VFP registers in 0 cycles.
238 // particularly effective at zeroing a VFP register.
292 // True if use a wider stride when allocating VFP registers.
295                                                "Use a wide stride when allocating VFP registers">;
304 // VFP register widths.
305 // True if splat a register between VFP and NEON instructions.
308                                              "Splat register from VFP to NEON",
311 // Whether or not it is profitable to expand VFP/NEON MLA/MLS instructions.
315                                              "Expand VFP/NEON MLA/MLS instructions">;
317 // Some targets have special RAW hazards for VFP/NEON VMLA/VMLS.
318 // True if VFP/NEON VMLA/VMLS have special RAW hazards.
323 // VFP to NEON, as an execution domain optimization.
324 // True if VMOVRS, VMOVSR and VMOVS will be converted from VFP to NEON.
346 // Some processors have a nonpipelined VFP coprocessor.
347 // True if VFP instructions are not pipelined.
350                                               "VFP instructions are not pipelined">;
353 // play nicely with other VFP / NEON instructions, and it's generally better
358                                              "Disable VFP / NEON MAC instructions">;
364                                              "Disable VFP / NEON FMA instructions">;