> N.B., with the same ifc & mkl I noticed last week that -r8 appears to > break tetra; I cannot see why it should. Can you elaborate? Does it not compile, or give different results?