[291] | 1 | \begin{taskinfo} |
---|
| 2 | \let\LIP\leader |
---|
| 3 | \let\IRISA\enable |
---|
| 4 | \let\UBS\enable |
---|
| 5 | \let\UPMC\enable |
---|
| 6 | \let\TIMA\enable |
---|
| 7 | \end{taskinfo} |
---|
| 8 | % |
---|
| 9 | \begin{objectif} |
---|
| 10 | The objective of this task is to convert the input specification of |
---|
| 11 | an hardware accelerator, which must be written in a familiar language |
---|
| 12 | (C/C++) with as few constraints as possible, into a form suitable for |
---|
| 13 | the HLS tools (i.e. HAS back-end tools of the COACH project). If the |
---|
| 14 | target is an ASIP, the frontend has to extract |
---|
| 15 | patterns from the source code and convert them into the definition |
---|
| 16 | of an extensible processor. If the target is a process network, the |
---|
| 17 | front end has to distribute the workload and the data sets as fairly |
---|
| 18 | as possible, identify communication channels, and output an \xcoach |
---|
| 19 | description. |
---|
| 20 | \end{objectif} |
---|
| 21 | % |
---|
| 22 | \begin{workpackage} |
---|
| 23 | \subtask{ASIP compiler} |
---|
| 24 | This sub-task aims at providing compiler support for custom instructions |
---|
| 25 | within the HAS front-end. It will take as input the COACH intermediate |
---|
| 26 | representation, and will output an annotated COACH IR containing the custom |
---|
| 27 | instructions definitions along with their occurrence in the application. |
---|
| 28 | \begin{livrable} |
---|
| 29 | \itemV{0}{18}{x}{\Sirisa}{ASIP compilation flow} |
---|
| 30 | In this first version of the software, the computations patterns corresponding to |
---|
| 31 | custom instructions are specified by the user, and then automatically extracted (when |
---|
| 32 | beneficial) from the application intermediate representation. |
---|
[304] | 33 | \itemL{18}{27}{x}{\Sirisa}{ASIP compilation flow}{0:6:3} |
---|
[291] | 34 | In this second version, the software will also be able to automatically identify |
---|
| 35 | interesting pattern candidates in the application code, and use them as custom |
---|
| 36 | instructions. |
---|
| 37 | \end{livrable} |
---|
| 38 | |
---|
| 39 | \subtask{Micro-architectural template models for ASIP} |
---|
| 40 | In this sub-task, we provide micro-architectural template models for the two target |
---|
| 41 | processor architectures (NIOS-II and MIPS) supported within the COACH-ASIP design flow. |
---|
| 42 | For each processor, we provide a simulation model (System-C) and a synthesizable model (VHDL) |
---|
| 43 | of the architecture, along with its architectural extensions |
---|
| 44 | \begin{livrable} |
---|
| 45 | \itemV{0}{12}{x}{\Sirisa}{SystemC for extensible MIPS } |
---|
| 46 | { A SystemC simulation model for a simple extensible MIPS architectural template } |
---|
| 47 | \itemL{12}{20}{x}{\Sirisa}{SystemC for extensible MIPS}{2:3:0} |
---|
| 48 | {A SystemC simulation model for an extensible MIPS with a tight architectural integration of |
---|
| 49 | its instruction set extensions} |
---|
| 50 | \itemV{3}{18}{h}{\Sirisa}{VHDL for an extensible MIPS} |
---|
| 51 | {A synthesizable VHDL model for a simple extensible MIPS architectural template} |
---|
[304] | 52 | \itemL{18}{27}{h}{\Sirisa}{VHDL for extensible MIPS}{9:9:3} |
---|
[291] | 53 | {A synthesizable VHDL model for an extensible MIPS with a tight architectural integration of |
---|
| 54 | its instruction set extensions} |
---|
[304] | 55 | \itemL{27}{36}{d}{\Sirisa}{Evaluation report }{0:0:2} |
---|
[291] | 56 | {An evaluation report with quantitative analysis of the performance/area trade-off induced by |
---|
| 57 | the different approaches} |
---|
| 58 | \end{livrable} |
---|
| 59 | |
---|
| 60 | \subtask{Parallelism optimization} |
---|
| 61 | Extraction of parallelism in polyhedral loops and conversion into a process network. |
---|
| 62 | \begin{livrable} |
---|
| 63 | \itemV{0}{6}{d}{\Slip}{Method, Preliminary Definition} |
---|
| 64 | Description and specification of a process construction method for programs with |
---|
| 65 | polyhedral loops. |
---|
| 66 | \itemL{30}{36}{d}{\Slip}{Process generation method}{10:0:9} |
---|
| 67 | Final assessment of the method and improved version of the specification. |
---|
| 68 | \itemV{6}{12}{x}{\Slip}{Process construction} |
---|
| 69 | Preliminary implementation in the Syntol framework. |
---|
| 70 | At this step the software will just implement a single constructor. |
---|
| 71 | \itemV{12}{18}{x}{\Slip} {Arrays and FIFO} |
---|
| 72 | Implementation of the array contraction and FIFO construction algorithm. |
---|
| 73 | Conversion of the input and output to the \xcoach format. |
---|
| 74 | \itemV{18}{30}{d+x}{\Slip}{Non-polyhedral extension} |
---|
| 75 | Extension of automatic parallelization and array contraction |
---|
| 76 | to non-polyhedral loops. Implementation in the Bee framework. |
---|
| 77 | \itemL{30}{36}{x}{\Slip} {Process/FIFO construction}{10:20:12} |
---|
| 78 | Final release taking into account the feedbacks from the |
---|
| 79 | demonstrator \STs. |
---|
| 80 | \end{livrable} |
---|
| 81 | |
---|
| 82 | \end{workpackage} |
---|
| 83 | |
---|