lli: Switch to using ParseIRFile, for consistency with other LLVM tools.

 - Also, switch tests to not using llvm-as. They run 20% faster now, not that it matters.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@118952 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/ExecutionEngine/fpbitcast.ll b/test/ExecutionEngine/fpbitcast.ll
index 34ca129..47cbb02 100644
--- a/test/ExecutionEngine/fpbitcast.ll
+++ b/test/ExecutionEngine/fpbitcast.ll
@@ -1,5 +1,4 @@
-; RUN: llvm-as < %s > %t.bc
-; RUN: lli -force-interpreter=true %t.bc | grep 40091eb8
+; RUN: lli -force-interpreter=true %s | grep 40091eb8
 ;
 define i32 @test(double %x) {
 entry: