Optimizing Compilers for Modern Architectures

By Allen. 2004

484 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Automatic performance tuning of sparse matrix kernels (2003, background): “1 for i = 0 to m− 1 do 2 for j = 0 to n− 1 do 3 y[i]← y[i] + val[i+ stride · j] · x[j] All array accesses are affine functions of the iteration variables i and j, and standard compiler register- and cache-level tiling transformations may be applied statically [7].”
Sourcebook of parallel computing (2003, methodology): “In the literature on compiler construction, these kinds of races are identified as dependences [27].”
The performance impact analysis of loop unrolling (2014, background): “Compilers are the Queen of Computing Science and Technology and they have long been the bridge from applications to systems [4].”
The tao of parallelism in algorithms (2011, background): “successful for regular, dense array programs in which array subscripts are affine functions of loop indices [35]; LU with pivoting requires fractal symbol analysis [48].”
Polyhedral parallel code generation for CUDA (2013, methodology): “Classical techniques of loop distribution and loop interchange [Wolfe 1996; Allen and Kennedy 2001; Trifunović et al. 2009] can dramatically impact the profitability of vectorization.”