I assess the vulnerability of applications for a living and I have been told over and over again by dev groups and ops that they can't upgrade their systems because it is too expensive. That's fine, but you need to understand you are making a security for money tradeoff and realize the risks you are bringing on in doing so. Make your own cost benefit analysis, but know that you are making a tradeoff and what that entails.
This is a solved problem: use 10,000+ rounds of PBKDF2, 4-5+ rounds of Bcrypt, or a sufficient work factor for Scrypt.