Since 1987 - Covering the Fastest Computers in the World and the People Who Run Them

July 24, 2012

Oracle Ditches Supercomputing Language Project

Michael Feldman

According to a blog posted by Guy Steele on Oracle’s website, the company will begin winding down the work on Fortress, an experimental open-source programming language designed to make HPC application development more productive. Fortress was originally developed by Sun Microsystems, where Steele led the language R&D effort. The work was partially funded under DARPA’s High Productivity Computing Systems (HPCS) project, which was designed to bring productive multi-petaflop systems to the supercomputing community.

Sun never made the final cut in the HPCS program, but the Fortress work was retained as an area of research at a time when the company still had serious supercomputing aspirations. When Oracle acquired Sun, it inherited the language technology, but not the enthusiasm to pursue the HPC market.

Fortress runs on the Java Virtual Machine (JVM) and could, at least theoretically, be applied to less compute-intensive domains. But since the language syntax and design are focused primarily on highly parallel, math-heavy code, it was likely deemed expendable by the higher-ups at Oracle, who couldn’t rationalize the continued research for its database-focused business.

In his blog post, Steele implies that they had essentially reached the end of the line, in a technical sense, with the technology:

[O]ver the last few years, as we have focused on implementing a compiler targeted to the Java Virtual Machine, we encountered some severe technical challenges having to do with the mismatch between the (rather ambitious) Fortress type system and a virtual machine not designed to support it (that would be every currently available VM, not just JVM). In addressing these challenges, we learned a lot about the implications of the Fortress type system for the implementation of symmetric multimethod dispatch, and have concluded that we are now unlikely to learn more (in a research sense) from completing the implementation of Fortress for JVM.

That leaves just Cray’s Chapel and IBM’s X10 as the surviving members of the HPCS language program. Like Fortress, both languages are being developed in the open-source model. While neither Chapel nor X10 has reached anywhere near mainstream acceptance, both efforts are still active.

Since Fortress was developed as an open-source technology, according to Steele, it will “remain available for the foreseeable future to those who wish to work on it.” He also says the that they’ll spend the next few months polishing up the code and language spec and penning some academic papers before shutting down the research effort at Oracle.

You can read the entire Fortress obituary here.

Share This