VMOVDQA  CC3, tmpStoreAVX2
	VPSLLD   $7, BB0, CC3; VPSRLD $25, BB0, BB0; VPXOR CC3, BB0, BB0
	VPSLLD   $7, BB1, CC3; VPSRLD $25, BB1, BB1; VPXOR CC3, BB1, BB1
	VPSLLD   $7, BB2, CC3; VPSRLD $25, BB2, BB2; VPXOR CC3, BB2, BB2
	VPSLLD   $7, BB3, CC3; VPSRLD $25, BB3, BB3; VPXOR CC3, BB3, BB3
	VMOVDQA  tmpStoreAVX2, CC3

// CHECK:   [[PRED:%.+]] = tensor.extract [[TOBOOL]][] : tensor<i1>
// CHECK:   cf.cond_br [[PRED]], ^bb1, ^bb2
// CHECK: ^bb1:
// CHECK:   [[THEN:%.+]] = call @testIf1Then(%arg1, %arg2)
// CHECK:   cf.br ^bb3([[THEN]] : tensor<*xf32>)
// CHECK: ^bb2:
// CHECK:   [[ELSE:%.+]] = call @testIf1Else(%arg1, %arg2)
// CHECK:   cf.br ^bb3([[ELSE]] : tensor<*xf32>)
// CHECK: ^bb3([[BBARG0:%.+]]: tensor<*xf32>):

// generating CondSelect instructions.
//
// Search for basic blocks that look like
//
//	bb0            bb0
//	 | \          /   \
//	 | bb1  or  bb1   bb2    <- trivial if/else blocks
//	 | /          \   /
//	bb2            bb3
//
// where the intermediate blocks are mostly empty (with no side-effects);
// rewrite Phis in the postdominator as CondSelects.
func branchelim(f *Func) {