InstrItinData<IIC_SSE_SQRTSD_RR, [InstrStage<62, [Port0, Port1]>] >,
InstrItinData<IIC_SSE_SQRTSD_RM, [InstrStage<62, [Port0, Port1]>] >,
+ InstrItinData<IIC_SSE_RSQRTPS_RR, [InstrStage<9, [Port0, Port1]>] >,
+ InstrItinData<IIC_SSE_RSQRTPS_RM, [InstrStage<10, [Port0, Port1]>] >,
+ InstrItinData<IIC_SSE_RSQRTSS_RR, [InstrStage<4, [Port0]>] >,
+ InstrItinData<IIC_SSE_RSQRTSS_RM, [InstrStage<4, [Port0]>] >,
+
InstrItinData<IIC_SSE_RCPP_RR, [InstrStage<9, [Port0, Port1]>] >,
InstrItinData<IIC_SSE_RCPP_RM, [InstrStage<10, [Port0, Port1]>] >,
InstrItinData<IIC_SSE_RCPS_RR, [InstrStage<4, [Port0]>] >,
// On the Atom, the throughput for taken branches is 2 cycles. For small
// simple loops, expand by a small factor to hide the backedge cost.
let LoopMicroOpBufferSize = 10;
+ let PostRAScheduler = 1;
let Itineraries = AtomItineraries;
}