Message-ID: <814651670.4510.1490660823502.JavaMail.firstname.lastname@example.org> Subject: Exported From Confluence MIME-Version: 1.0 Content-Type: multipart/related; boundary="----=_Part_4509_217390631.1490660823499" ------=_Part_4509_217390631.1490660823499 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Content-Location: file:///C:/exported.html
The computer industry is at a major crossroads. Historically, single-core = architectures have delivered sustained performance improvements at an expon= ential rate over multiple generations, and sequential programming languages= and compilers have been able to pass these performance increases on to a w= ide range of applications. However, this is no longer the case due to= current hardware trends and power efficiency limits. Instead of buil= ding processors with faster clock speeds, all computers--- embedded, mainst= ream, and high-end --- are being built using chips with an increasing numbe= r of processor cores, with little or no increase in clock speed per core.&n= bsp; This trend poses a tremendous challenge for software enablement on fut= ure extreme scale systems as the number of cores per socket continues to gr= ow, and the cores become more heterogeneous.
The Habanero project at Rice University was initiated in Fall 2007 to addr= ess the multicore software challenge by developing new programming technolo= gies --- languages, compilers, runtime systems, and tools --- that support = portable parallel abstractions for future hardware with high productivity a= nd high performance. Our goal is to ensure that future software rewri= tes are done on software platforms that enable application developers to re= use their investment across multiple generations of homogeneous and heterog= eneous extreme scale hardware. We also envision broader impact of thi= s research including: updating the foundations of parallel software in intr= oductory Computer Science courses, building an open source testbed to grow = the ecosystem of researchers in the parallel software area, and using our r= esearch infrastructure as the basis for building reference implementations = of future industry standards.