Hi Guys, <br><br>I was working on some floating point intensive benchmarks and realize that the floating point register allocation in llvm assumes that there are only 7 floating point registers in X86, whereas the hardware has 8. <br>
<br>Line number <br><pre class="fragment">00266 assert(Reg >= X86::FP0 && Reg <= X86::FP6 && <span class="stringliteral">"Expected FP register!"</span>); </pre>
<br>of X86FloatingPoint.cpp. <br><br>Is there any reason for only counting from 0 to 6, when there are actually 8 in hardware ? <br><br>Is there an assumption somewhere else, that I am missing. <br><br><br><br>Thanks and Regards<br>
<br>Aparna Kotha <br><br>Graduate Student <br>University of Maryland, College Park <br><br><pre class="fragment"><br></pre>