DocumentCode :
2459615
Title :
Teaching computer architecture/organisation using simulators
Author :
Grünbacher, Herbert
Author_Institution :
Wien Univ., Austria
Volume :
3
fYear :
1998
fDate :
4-7 Nov. 1998
Firstpage :
1107
Abstract :
Experience shows that many students, especially those with little hardware background, encounter difficulties in understanding the consequences and even concepts of conventional instruction pipelining; superscalar instruction processing is even more complicated and harder to understand. It is particularly difficult to statically teach the concept of a pipeline. Therefore we developed software to simulate and dynamically visualize the processing of instructions by pipelined (superscalar) processors. Three simulators have been developed: WinDLX is based on Hennessy/Pattersons DLX architecture and is modeled at the architecture level, therefore very little processor-internal information is given. MIPSim is based on Patterson/Hennessy´s MIPS processor book and is modeled at the computer organization level, functional units like register file, pipeline registers, multiplexers are visible and MIPSim displays content and dynamic behavior of such units. M10kSim is based on the MIPS R10000 architecture and models the instruction decode and dispatch unit, the branch unit, the instruction queues and the functional units. Concepts like register renaming, branch history table, branch resume buffer, out of order execution can be explained easily using the simulator. Teaching cache organization is an easier task, nevertheless visualising cache activities helps understanding the dynamics of a cache memory, Xcache is a simulator which displays the interactions between instruction memory and instruction cache, data memory and data cache, respectively.
Keywords :
cache storage; computer architecture; computer science education; digital simulation; pipeline processing; teaching; Hennessy/Pattersons DLX architecture; M10kSim; MIPS R10000 architecture; MIPSim; MIPSim displays content; Patterson/Hennessy´s MIPS processor book; WinDLX; Xcache; address calculation; branch history table; branch resume buffer; branch unit; cache organization teaching; computer architecture teaching; computer organisation teaching; computer organization level; data cache; data memory; dispatch unit; dynamic behavior; floating-point adder; floating-point multiply/divide/square-root unit; functional units; instruction cache; instruction decode unit; instruction memory; instruction pipelining; instruction queues; instructions processing visualisation; multiplexers; out of order execution; pipeline registers; register file; register renaming; simulators; superscalar instruction processing; Books; Computational modeling; Computer architecture; Computer displays; Computer simulation; Education; Hardware; Pipeline processing; Registers; Visualization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Frontiers in Education Conference, 1998. FIE '98. 28th Annual
Conference_Location :
Tempe, AZ, USA
ISSN :
0190-5848
Print_ISBN :
0-7803-4762-5
Type :
conf
DOI :
10.1109/FIE.1998.738576
Filename :
738576
Link To Document :
بازگشت