Without pain, a lot of gain! Two simple changes make things much faster (12% faster overall gain in my startup code, as measured by CPU cycles needed to process the 'retroforth.f' (reva.f in my case) startup code processing). In 'dict.pl', make the addition of 'ALIGN 4' before each dictionary entry. Likewise, in the FORTH and MCODE macros (code and mcode in rf), put an ALIGN 4 before the code label). In combination, this results in a greater than 10% speedup in compilation on both my AMD and Intel machines. The only cost is that the binary is a little bigger. Best regards, Ron the Optimizer -- My GPG public key is at http://ronware.org/ fingerprint: 8130 734C 69A3 6542 0853 CB42 3ECF 9259 AD29 415D