| M. D. Smith and G. Holloway, "An introduction to machine SUIF and its portable libraries for analysis and optimization," in Division of Engineering and Applied Sciences. Cambridge, MA: Harvard Univ., 2002. |
....described in Section 3.1, is a directed acyclic graph, which describes the precedence relationship between the computing tasks. A task node in the task graph refers to a function, which could be written in high level languages, such as C C . It is analyzed using the SUIF [3] and Machine SUIF [23] tools; the result is imported in our environment as a control data flow graph (CDFG) CDFG reflects the control flow in a function, and may contain loops, branches, and jumps. Each node in CDFGs is a basic block, or a set of instructions that contains only one control transfer instruction and ....
M. D. Smith and G. Holloway. An Introduction to Machine SUIF and Its Portable Libraries for Analysis and Optimization. Division of Engineering and Applied Sciences, Harvard University, July 2002.
....task graph, as described in Section 3, is a directed acyclic graph, which shows the precedence relationship between tasks. Each task node is mapped to a function, which are written in a high level programming language, such as C C . The functions are analyzed using the SUIF [2] and Machine SUIF [21] tools. The resulting compiled function is imported in our environment as a control data flow graph (CDFG) A CDFG reflects the control flow within a function, and may contain loops, branches, and jumps. Each node in the CDFG is a basic block, i.e. a set of instructions that contains only one ....
M. D. Smith and G. Holloway. An Introduction to Machine SUIF and Its Portable Libraries for Analysis and Optimization. Division of Engineering and Applied Sciences, Harvard University, July 2002.
....speedup. Although quite rough, this model is also very fast to evaluate and hence adapted for use in the inner loop of our identification algorithm, where by no means one could use a computationally heavier model. 7. RESULTS The described algorithms were implemented within the MachSUIF framework [15] and tested on a subset of the MediaBench [13] suite benchmarks. Application C code is compiled to MachSUIF intermediate representation and preprocessed with a classic if conversion pass. In order to show the potentials of our algorithms with respect to the state of the art, we have implemented ....
M. D. Smith and G. Holloway. An Introduction to Machine SUIF and its Portable Libraries for Analysis and Optimization. Harvard University, Cambridge, Mass., 2000.
....DSP functions. DSP functions typically exhibit a large amount of parallelism making them ideal candidates for hardware compilation. The DSP functions were taken from the MediaBench test suite [10] The files were compiled into CDFGs using the SUIF compiler infrastructure [11] and the Machine SUIF [12] backend. Then, each of the benchmarks was given to our framework [6] which transforms the IR into synthesizable VHDL. The VHDL was then synthesized using the Synopsys Behavioral Compiler for architectural synthesis followed by the Synopsys Design Compiler for logic synthesis. We performed SSA ....
M.D. Smith and G. Holloway, An Introduction to Machine SUIF and its Portable Libraries for Analysis and Optimization, Division of Engineering and Applied Sciences, Harvard University,
....assembly level as well. In SUIF, it is possible to annotate the instructions with necessary information like data dependence, and use them in separate passes afterwards. Also, the SUIF package contains many optimization modules, which improve the quality of the code produced. We used the MachSUIF [15] framework to generate Alpha assembly code from the SUIF IR. We have implemented the pro ling phase also in the SUIF framework. 4. EXPERIMENTAL EVALUATION To study the e ectiveness of our thread partitioning schemes, we conducted a simulation based evaluation. This section details the ....
M. D. Smith and G. Holloway. An Introduction to Machine SUIF and Its Portable Libraries for Analysis and Optimization.
....consist of an operator such as addition, two source operands, and a destination, much like a typical datapath element. This step is essentially a standard compiler front end, which we implemented using the SUIF 2 system developed by Lam and her students at Stanford with the MachSUIF extensions [13]. The opcodes are mnemonics. In n 2 , sne r 0 , a, b ( set not equal ) performs the a = b comparison from line 3, setting r 0 to 1 if a and b are different, and bt n 3 , r 0 means to branch to n 3 if r 0 is non zero. Similarly, sgt in n 3 (set greater than) computes a b from line 4. The ....
Michael D. Smith and Glenn Holloway. An Introduction to Machine SUIF and its Portable Libraries for Analysis and Optimization. Harvard University, 2000. Machine SUIF documentation set.
No context found.
M. D. Smith and G. Holloway, "An introduction to machine SUIF and its portable libraries for analysis and optimization," in Division of Engineering and Applied Sciences. Cambridge, MA: Harvard Univ., 2002.
No context found.
M.D. Smith, and G. Holloway, ""An introduction to machine SUIF and its portable libraries for analysis and optimization"," Division of Engineering and Applied Sciences, Harvard University, Tech. Rep., 2002.
No context found.
M.D. Smith, and G. Holloway. "An introduction to machine SUIF and its portable libraries for analysis and optimization". Technical report, Division of Engineering and Applied Sciences, Harvard University, 2002.
No context found.
M. Smith and G. Holloway, "An introduction to machine suif and its portable libraries for analysis and optimization," Division of Engineering and Applied Sciences, Harvard University, Tech. Rep., 2002.
No context found.
Michael D. Smith and Glenn Holloway. An Introduction to Machine SUIF and Its portable Libraries for Analysis and Optimization. Division of Engineering and Applied Sciences, Harvard University, 2001.
No context found.
M. D. Smith and G. Holloway, "An introduction to machine SUIF and its portable libraries for analysis and optimization," in Division of Engineering and Applied Sciences. Cambridge, MA: Harvard Univ., 2002.
No context found.
M. D. Smith and G. Holloway. An Introduction to Machine SUIF and Its Portable Libraries for Analysis and Optimization. Division of Engineering and Applied Sciences, Harvard University, July 2002.
No context found.
Michael D. Smith and Glenn Holloway. An Introduction to Machine SUIF and Its Portable Libraries for Analysis and Optimization. Harvard University, http://www.eecs.harvard.edu/hube/software/nci/overview.pdf, July 2002.
No context found.
M. D. Smith, G. Holloway, \An Introduction to Machine SUIF and Its Portable Libraries for Analysis and Optimization".
No context found.
Michael D. Smith and Glenn Holloway. An Introduction to Machine SUIF and its Portable Libraries for Analysis and Optimization.Harvard University, Cambridge, Mass., 2000.
No context found.
Michael D. Smith and Glenn Holloway. An Introduction to Machine SUIF and Its portable Libraries for Analysis and Optimization. Division of Engineering and Applied Sciences, Harvard University, 2001.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC