SlideShare a Scribd company logo
1 of 5
Download to read offline
ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2, No 5, May 2013
1904
All Rights Reserved © 2013 IJARCET
Reconfiguration of Memory for High Speed Matrix
Multiplication
Ms. Shamshad Shirgeri1
,Ms. Pallavi Umesh Naik2
,Mrs. Rohini3
, Prof. Krishnananda Shet4
1,2,3 – Student, M.Tech in Industrial Electronics, 4 - Research Scholar,
1,2,3-KLS’s VDRIT,Haliyal affiliated to VTU,Belgaum, 4- Kuvempu University, Shivamogga, Karnataka, India.
Abstract
The paper entitled
“RECONFIGURATION OF MEMORY
FOR HIGH SPEED MATRIX
MULTIPLICATION“ is basically an
enhancement in speed of matrix
multiplication. The project aims at reducing
the time involved in the operation.
The system presented here makes
use of three RAMS designated as MAT-
RAM-A, MAT-RAM-B and MAT-RAM-C
and system also consists of a control
circuitry.
This system works in two modes i.e.,
mode-1 and mode-2. In mode-1 the RAM’s
act as extensions of existing memory of the
processor and in mode-2 the two RAMs
MAT-RAM-A and MAT-RAM-B are
configured as source for hardware with
MAT-RAM-C working as destination for
storage or resultant matrix after
multiplication. The RAMs are configured
back in mode-1, so that data can be read
from MAT-RAM-C.
Keywords: RAM, MAT-RAM-A, MAT-
RAM-B and MAT-RAM-C, SIMD.
1. INTRODUCTION
In the present electronic era, speed
plays a very crucial role. Speed size ad
economics are the present buzzwords in the
electronic industry. Any designer will have to
take all the three outlined factors into account
before starting any design. The trend now-a-
days is miniaturization coupled with fast
access time. The technique presented here uses
additional RAMs which will act as extension
of the existing RAMs of computers. These
RAM devices are switched in a special
configuration to provide high speed matrix
multiplication. In vector computers, special
vector registers within the CPU are used for
handling vector operations. The hardware and
associated software will have to take care of
transferring data approximately. This is quite
an involved task and it involves substantial
hardware and software overheads. In SIMD
architecture the multiplication of two matrices
involves a number of load and store
instructions. Moreover this requires N
processors for N x N matrix operation. Also in
this case the data will have to be stored
column wise in each machine and then load
and store instructions have to be executed for
N times. The technique presented here
overcomes these problems and provides a very
convenient and economical solution for high
speed matrix multiplication.
2. PREAMBLE
2.1 Statement of problem
Traditional method for matrix
multiplication consumes lot of time, as it
involves reading from memory, multiplying
them and again storing the result back in
memory and this process has to be repeated for
each element of the matrix. This process
ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2, No 5, May 2013
1905
All Rights Reserved © 2013 IJARCET
obviously wastes processors time. Processors
time can be judiciously utilized by the
technique presented here.
2.2 Scope of Study
This system provides a very
convenient and economical solution for high
speed matrix multiplication. The speed is quite
high and is limited only by the access time of
the RAM devices used for this purpose.
2.3 Methodology
 Splitting the design into functional
modules. The complete design to be
implemented to be broken into smaller
modules catering to the different
functions.
 Adding the designed modules to the
library. The design modules are added
to the library so that they can be
accessed whenever necessary. This
method reduces length of coding.
 Interconnecting all the modules to
complete the circuit. The various
modules required for are called from
the library and then are interconnected
as required.
2.4 Limitation of the system
 The speed of the system is limited
only by the access time of the RAMS
used in the circuit.
3. BLOCK DIAGRAM OF
RECONFIGURATION OF
MEMORY FOR HIGH SPEED
MATRIX MULTIPLICATION
Figure1. Block Diagram of
reconfiguration of memory for
high speed matrix multiplication
The system presented here consists of
 Processor
 Switching unit
 External control unit
 RAM’s
This system operates in one of the two
modes in mode-1 these can be used as RAM’s
of the main processor, connected just as an
extension RAM’s of the main processor. In
this mode the necessary control signals of the
RAM’s are connected to the processors data
bus.
In mode-2, these are used for matrix
multiplication process. In this mode the
RAM’s are isolated from the main processor
MATRAM-A and MATRAM-B store the
input data matrix and output of the operation is
stored in MATRAM-C. The necessary control
signals are obtained from the external circuit.
3.1 Switching unit
The switching part consists of two
multiplexers one for transferring data between
RAM and external control unit. The other
multiplexer is used for routing the control
signals to RAM either from the processor or
from an external control unit.
Depending upon the select input
switching unit connect the RAM’s either to the
processor or the external control unit. The
system operates in two modes, the mode of
operation is decided by the input applied to
select pin of switching unit.
Figure2. Switching Unit
ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2, No 5, May 2013
1906
All Rights Reserved © 2013 IJARCET
3.2 External control unit
Figure3. External Control Unit
It consists of
 A Multiplier
 An 8 bit parallel adder
 A register
 Three counters
 A clock
3.2.1 Multiplier
Used for the multiplication of the matrix
elements. It takes the inputs from the two
matrices, namely MATRAM-A and
MATRAM-B. It generates the partial product
and this becomes the input for the order block.
3.2.2 8 bit Parallel adder
Used for the summation purpose. It has
two inputs one is partial product from the
multiplier block and the other input is the
previous sum from the register block.
Figure4. 8 bit parallel adder
3.2.3 Register
Used for storing the sum from the
adder block temporarily. Register initially
contains zero. Register gets reset once the
resultant matrix element is generated.
3.2.4 Counters
The counters used here are binary
up counters that generate addresses of
RAMs during Mode-2 operation along
with the addresses, and bi-directional
multiplexers used I the circuit that controls
the mode of operation of the system. The
counter has a register in it, which should
be loaded with the number of elements of
the matrix. The counter counts until the
loaded value, during which the select
signal generated by the counter is at logic
high and hence the system operates I
mode-2, as soon as the count equals the
stored value the select signal becomes low,
now the system is switched back to mode-
1.
Figure5. Counters
3.3 RAM’s
MATRAM-A, MATRAM-B and MATRAM-
C which are present in the system are just an
extension of existing RAM’s of the main
processor during mode-1 and in this mode the
address lines are generated by the processor
and input data are given by the processor
itself. MATRAM-A and MATRAM-B are
configured as source for the hardware.
MATRAM-C works as destination matrix to
store the resultant. During mode-2 their
address lines are generated by the external
circuit. The RAM’s are again configured back
in mode-1, so that data can be read from
MATRAM-C.
ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2, No 5, May 2013
1907
All Rights Reserved © 2013 IJARCET
4. MODES OF OPERATION
4.1 Mode-1
In mode-1, the input to the select pin
of external circuit is at logic low. Under this
condition the RAM’s are connected to the
processor. Under this mode, the RAM’s are
connected to the processor and are isolated
from the external circuit. In this mode the
address and control signals for the RAM’s are
obtained from the processor, and processor
data bus is connected to the RAM’s data bus.
Hence reading from or writing into the RAM
is under the control of the processor.
4.2 Mode-2
In the mode-2 the input to the select
pin is high. Under this condition the RAM’s
are connected to the external circuit. The
RAM’s are completely isolated from main
processor. In this mode, the address of the
RAM’s are obtained from the external control
unit, the data buses of RAM-A and RAM-B
are connected to the multiplier input and the
data bus of RAM-C is connected to the output
of the adder.
5. VHDL IMPLEMENTATION
Figure6. Implementation of
Reconfiguration of memory for high speed
matrix multiplication using VHDL.
6. GRAPHICAL COMPARISION
Figure7. Graphical comparison of
Reconfiguration of memory for high speed
matrix multiplication
ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2, No 5, May 2013
1908
All Rights Reserved © 2013 IJARCET
7. CONCLUSION
 It provides a much faster and an
economic technique for matrix
multiplication operation.
 The speed in this technique will be
more than ten times faster
compared to Pentium processor.
 Cost involved will be less
compared to that of a vector
machine or a SMID machine.
 The speed of the operation is only
limited by the access time of the
RAM chip used in the circuit.
8. REFERENCES
 P.H.W. Leong, M.P. Leong, O.Y.H.
Cheung, T. Tung, C.M. Kwok, and
K.H. Lee, “Pilchard – A
Reconfigurable Computing Platform
with Memory Interface”.
 Xilinx, “Virtex-E 1.8 V Field
Programmable Gate Arrays”.
 Randall Hyde, "The Art of Assembly
Language Programming".
http://webster.cs.ucr.edu/Page_AoA
Linux/HTML/AoATOC.html
 IEEE international conference on “An
FPGA based floating-point processor
array supporting a high-precision dot
product” by Fritz Mayer.

More Related Content

What's hot (20)

Ec8791 unit 5 processes and operating systems
Ec8791 unit 5 processes and operating systemsEc8791 unit 5 processes and operating systems
Ec8791 unit 5 processes and operating systems
 
Lecture 47
Lecture 47Lecture 47
Lecture 47
 
00364438
0036443800364438
00364438
 
Computer architecture
Computer architectureComputer architecture
Computer architecture
 
Co notes3 sem
Co notes3 semCo notes3 sem
Co notes3 sem
 
computer architecture
computer architecture computer architecture
computer architecture
 
Low Power Area Efficient Parallel Counter Architecture
Low Power Area Efficient Parallel Counter ArchitectureLow Power Area Efficient Parallel Counter Architecture
Low Power Area Efficient Parallel Counter Architecture
 
eet_NPU_file.PDF
eet_NPU_file.PDFeet_NPU_file.PDF
eet_NPU_file.PDF
 
Architectures for parallel
Architectures for parallelArchitectures for parallel
Architectures for parallel
 
Lecture 3
Lecture 3Lecture 3
Lecture 3
 
Direct memory access
Direct memory accessDirect memory access
Direct memory access
 
Direct Memory Access (DMA)-Working and Implementation
Direct Memory Access (DMA)-Working and ImplementationDirect Memory Access (DMA)-Working and Implementation
Direct Memory Access (DMA)-Working and Implementation
 
Dma
DmaDma
Dma
 
Lecture 48
Lecture 48Lecture 48
Lecture 48
 
Ec8791 arm 9 processor
Ec8791 arm 9 processorEc8791 arm 9 processor
Ec8791 arm 9 processor
 
INTERCONNECTION STRUCTURE
INTERCONNECTION STRUCTUREINTERCONNECTION STRUCTURE
INTERCONNECTION STRUCTURE
 
Accelerix ISSCC 1998 Paper
Accelerix ISSCC 1998 PaperAccelerix ISSCC 1998 Paper
Accelerix ISSCC 1998 Paper
 
Dma
DmaDma
Dma
 
L010236974
L010236974L010236974
L010236974
 
Unit 4-input-output organization
Unit 4-input-output organizationUnit 4-input-output organization
Unit 4-input-output organization
 

Viewers also liked

Viewers also liked (8)

PowerMagpie
PowerMagpiePowerMagpie
PowerMagpie
 
55 60
55 6055 60
55 60
 
293 299
293 299293 299
293 299
 
300 305
300 305300 305
300 305
 
Ijarcet vol-2-issue-7-2241-2245
Ijarcet vol-2-issue-7-2241-2245Ijarcet vol-2-issue-7-2241-2245
Ijarcet vol-2-issue-7-2241-2245
 
Ijarcet vol-2-issue-7-2351-2356
Ijarcet vol-2-issue-7-2351-2356Ijarcet vol-2-issue-7-2351-2356
Ijarcet vol-2-issue-7-2351-2356
 
Cli2 Bibalex
Cli2 BibalexCli2 Bibalex
Cli2 Bibalex
 
Ijarcet vol-2-issue-3-1280-1284
Ijarcet vol-2-issue-3-1280-1284Ijarcet vol-2-issue-3-1280-1284
Ijarcet vol-2-issue-3-1280-1284
 

Similar to 1904 1908

Performance comparison of row per slave and rows set per slave method in pvm ...
Performance comparison of row per slave and rows set per slave method in pvm ...Performance comparison of row per slave and rows set per slave method in pvm ...
Performance comparison of row per slave and rows set per slave method in pvm ...eSAT Journals
 
Performance comparison of row per slave and rows set
Performance comparison of row per slave and rows setPerformance comparison of row per slave and rows set
Performance comparison of row per slave and rows seteSAT Publishing House
 
Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...
Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...
Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...IRJET Journal
 
Paper id 42201619
Paper id 42201619Paper id 42201619
Paper id 42201619IJRAT
 
The Architecture Design for Find Outing Vehicle Position System using Communi...
The Architecture Design for Find Outing Vehicle Position System using Communi...The Architecture Design for Find Outing Vehicle Position System using Communi...
The Architecture Design for Find Outing Vehicle Position System using Communi...Kumar Goud
 
IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...
IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...
IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...IRJET Journal
 
Secure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfigurationSecure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfigurationeSAT Journals
 
Secure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfigurationSecure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfigurationeSAT Publishing House
 
FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...
FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...
FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...IRJET Journal
 
IRJET- Ad-hoc Based Outdoor Positioning System
IRJET- Ad-hoc Based Outdoor Positioning SystemIRJET- Ad-hoc Based Outdoor Positioning System
IRJET- Ad-hoc Based Outdoor Positioning SystemIRJET Journal
 
IRJET- Design of Low Power 32- Bit RISC Processor using Verilog HDL
IRJET-  	  Design of Low Power 32- Bit RISC Processor using Verilog HDLIRJET-  	  Design of Low Power 32- Bit RISC Processor using Verilog HDL
IRJET- Design of Low Power 32- Bit RISC Processor using Verilog HDLIRJET Journal
 
Wireless Data Acquisition System Using ARM Cortex M-3
Wireless Data Acquisition System Using ARM Cortex M-3Wireless Data Acquisition System Using ARM Cortex M-3
Wireless Data Acquisition System Using ARM Cortex M-3IOSR Journals
 
Power minimization of systems using Performance Enhancement Guaranteed Caches
Power minimization of systems using Performance Enhancement Guaranteed CachesPower minimization of systems using Performance Enhancement Guaranteed Caches
Power minimization of systems using Performance Enhancement Guaranteed CachesIJTET Journal
 
A new method for self-organized dynamic delay loop associated pipeline with ...
A new method for self-organized dynamic delay loop associated  pipeline with ...A new method for self-organized dynamic delay loop associated  pipeline with ...
A new method for self-organized dynamic delay loop associated pipeline with ...IJECEIAES
 

Similar to 1904 1908 (20)

Performance comparison of row per slave and rows set per slave method in pvm ...
Performance comparison of row per slave and rows set per slave method in pvm ...Performance comparison of row per slave and rows set per slave method in pvm ...
Performance comparison of row per slave and rows set per slave method in pvm ...
 
Performance comparison of row per slave and rows set
Performance comparison of row per slave and rows setPerformance comparison of row per slave and rows set
Performance comparison of row per slave and rows set
 
Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...
Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...
Design of a Novel Multiplier and Accumulator using Modified Booth Algorithm w...
 
Matter new
Matter newMatter new
Matter new
 
Aw25293296
Aw25293296Aw25293296
Aw25293296
 
cuTau Leaping
cuTau LeapingcuTau Leaping
cuTau Leaping
 
Paper id 42201619
Paper id 42201619Paper id 42201619
Paper id 42201619
 
The Architecture Design for Find Outing Vehicle Position System using Communi...
The Architecture Design for Find Outing Vehicle Position System using Communi...The Architecture Design for Find Outing Vehicle Position System using Communi...
The Architecture Design for Find Outing Vehicle Position System using Communi...
 
IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...
IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...
IRJET-Design of ARM Based Data Acquisition and Control System for Engine Asse...
 
Secure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfigurationSecure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfiguration
 
Secure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfigurationSecure remote protocol for fpga reconfiguration
Secure remote protocol for fpga reconfiguration
 
FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...
FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...
FPGA Implementation of High Speed AMBA Bus Architecture for Image Transmissio...
 
IRJET- Ad-hoc Based Outdoor Positioning System
IRJET- Ad-hoc Based Outdoor Positioning SystemIRJET- Ad-hoc Based Outdoor Positioning System
IRJET- Ad-hoc Based Outdoor Positioning System
 
FrackingPaper
FrackingPaperFrackingPaper
FrackingPaper
 
Implementation of MAC using Modified Booth Algorithm
Implementation of MAC using Modified Booth AlgorithmImplementation of MAC using Modified Booth Algorithm
Implementation of MAC using Modified Booth Algorithm
 
IRJET- Design of Low Power 32- Bit RISC Processor using Verilog HDL
IRJET-  	  Design of Low Power 32- Bit RISC Processor using Verilog HDLIRJET-  	  Design of Low Power 32- Bit RISC Processor using Verilog HDL
IRJET- Design of Low Power 32- Bit RISC Processor using Verilog HDL
 
Wireless Data Acquisition System Using ARM Cortex M-3
Wireless Data Acquisition System Using ARM Cortex M-3Wireless Data Acquisition System Using ARM Cortex M-3
Wireless Data Acquisition System Using ARM Cortex M-3
 
E010113235
E010113235E010113235
E010113235
 
Power minimization of systems using Performance Enhancement Guaranteed Caches
Power minimization of systems using Performance Enhancement Guaranteed CachesPower minimization of systems using Performance Enhancement Guaranteed Caches
Power minimization of systems using Performance Enhancement Guaranteed Caches
 
A new method for self-organized dynamic delay loop associated pipeline with ...
A new method for self-organized dynamic delay loop associated  pipeline with ...A new method for self-organized dynamic delay loop associated  pipeline with ...
A new method for self-organized dynamic delay loop associated pipeline with ...
 

More from Editor IJARCET

Electrically small antennas: The art of miniaturization
Electrically small antennas: The art of miniaturizationElectrically small antennas: The art of miniaturization
Electrically small antennas: The art of miniaturizationEditor IJARCET
 
Volume 2-issue-6-2205-2207
Volume 2-issue-6-2205-2207Volume 2-issue-6-2205-2207
Volume 2-issue-6-2205-2207Editor IJARCET
 
Volume 2-issue-6-2195-2199
Volume 2-issue-6-2195-2199Volume 2-issue-6-2195-2199
Volume 2-issue-6-2195-2199Editor IJARCET
 
Volume 2-issue-6-2200-2204
Volume 2-issue-6-2200-2204Volume 2-issue-6-2200-2204
Volume 2-issue-6-2200-2204Editor IJARCET
 
Volume 2-issue-6-2190-2194
Volume 2-issue-6-2190-2194Volume 2-issue-6-2190-2194
Volume 2-issue-6-2190-2194Editor IJARCET
 
Volume 2-issue-6-2186-2189
Volume 2-issue-6-2186-2189Volume 2-issue-6-2186-2189
Volume 2-issue-6-2186-2189Editor IJARCET
 
Volume 2-issue-6-2177-2185
Volume 2-issue-6-2177-2185Volume 2-issue-6-2177-2185
Volume 2-issue-6-2177-2185Editor IJARCET
 
Volume 2-issue-6-2173-2176
Volume 2-issue-6-2173-2176Volume 2-issue-6-2173-2176
Volume 2-issue-6-2173-2176Editor IJARCET
 
Volume 2-issue-6-2165-2172
Volume 2-issue-6-2165-2172Volume 2-issue-6-2165-2172
Volume 2-issue-6-2165-2172Editor IJARCET
 
Volume 2-issue-6-2159-2164
Volume 2-issue-6-2159-2164Volume 2-issue-6-2159-2164
Volume 2-issue-6-2159-2164Editor IJARCET
 
Volume 2-issue-6-2155-2158
Volume 2-issue-6-2155-2158Volume 2-issue-6-2155-2158
Volume 2-issue-6-2155-2158Editor IJARCET
 
Volume 2-issue-6-2148-2154
Volume 2-issue-6-2148-2154Volume 2-issue-6-2148-2154
Volume 2-issue-6-2148-2154Editor IJARCET
 
Volume 2-issue-6-2143-2147
Volume 2-issue-6-2143-2147Volume 2-issue-6-2143-2147
Volume 2-issue-6-2143-2147Editor IJARCET
 
Volume 2-issue-6-2119-2124
Volume 2-issue-6-2119-2124Volume 2-issue-6-2119-2124
Volume 2-issue-6-2119-2124Editor IJARCET
 
Volume 2-issue-6-2139-2142
Volume 2-issue-6-2139-2142Volume 2-issue-6-2139-2142
Volume 2-issue-6-2139-2142Editor IJARCET
 
Volume 2-issue-6-2130-2138
Volume 2-issue-6-2130-2138Volume 2-issue-6-2130-2138
Volume 2-issue-6-2130-2138Editor IJARCET
 
Volume 2-issue-6-2125-2129
Volume 2-issue-6-2125-2129Volume 2-issue-6-2125-2129
Volume 2-issue-6-2125-2129Editor IJARCET
 
Volume 2-issue-6-2114-2118
Volume 2-issue-6-2114-2118Volume 2-issue-6-2114-2118
Volume 2-issue-6-2114-2118Editor IJARCET
 
Volume 2-issue-6-2108-2113
Volume 2-issue-6-2108-2113Volume 2-issue-6-2108-2113
Volume 2-issue-6-2108-2113Editor IJARCET
 
Volume 2-issue-6-2102-2107
Volume 2-issue-6-2102-2107Volume 2-issue-6-2102-2107
Volume 2-issue-6-2102-2107Editor IJARCET
 

More from Editor IJARCET (20)

Electrically small antennas: The art of miniaturization
Electrically small antennas: The art of miniaturizationElectrically small antennas: The art of miniaturization
Electrically small antennas: The art of miniaturization
 
Volume 2-issue-6-2205-2207
Volume 2-issue-6-2205-2207Volume 2-issue-6-2205-2207
Volume 2-issue-6-2205-2207
 
Volume 2-issue-6-2195-2199
Volume 2-issue-6-2195-2199Volume 2-issue-6-2195-2199
Volume 2-issue-6-2195-2199
 
Volume 2-issue-6-2200-2204
Volume 2-issue-6-2200-2204Volume 2-issue-6-2200-2204
Volume 2-issue-6-2200-2204
 
Volume 2-issue-6-2190-2194
Volume 2-issue-6-2190-2194Volume 2-issue-6-2190-2194
Volume 2-issue-6-2190-2194
 
Volume 2-issue-6-2186-2189
Volume 2-issue-6-2186-2189Volume 2-issue-6-2186-2189
Volume 2-issue-6-2186-2189
 
Volume 2-issue-6-2177-2185
Volume 2-issue-6-2177-2185Volume 2-issue-6-2177-2185
Volume 2-issue-6-2177-2185
 
Volume 2-issue-6-2173-2176
Volume 2-issue-6-2173-2176Volume 2-issue-6-2173-2176
Volume 2-issue-6-2173-2176
 
Volume 2-issue-6-2165-2172
Volume 2-issue-6-2165-2172Volume 2-issue-6-2165-2172
Volume 2-issue-6-2165-2172
 
Volume 2-issue-6-2159-2164
Volume 2-issue-6-2159-2164Volume 2-issue-6-2159-2164
Volume 2-issue-6-2159-2164
 
Volume 2-issue-6-2155-2158
Volume 2-issue-6-2155-2158Volume 2-issue-6-2155-2158
Volume 2-issue-6-2155-2158
 
Volume 2-issue-6-2148-2154
Volume 2-issue-6-2148-2154Volume 2-issue-6-2148-2154
Volume 2-issue-6-2148-2154
 
Volume 2-issue-6-2143-2147
Volume 2-issue-6-2143-2147Volume 2-issue-6-2143-2147
Volume 2-issue-6-2143-2147
 
Volume 2-issue-6-2119-2124
Volume 2-issue-6-2119-2124Volume 2-issue-6-2119-2124
Volume 2-issue-6-2119-2124
 
Volume 2-issue-6-2139-2142
Volume 2-issue-6-2139-2142Volume 2-issue-6-2139-2142
Volume 2-issue-6-2139-2142
 
Volume 2-issue-6-2130-2138
Volume 2-issue-6-2130-2138Volume 2-issue-6-2130-2138
Volume 2-issue-6-2130-2138
 
Volume 2-issue-6-2125-2129
Volume 2-issue-6-2125-2129Volume 2-issue-6-2125-2129
Volume 2-issue-6-2125-2129
 
Volume 2-issue-6-2114-2118
Volume 2-issue-6-2114-2118Volume 2-issue-6-2114-2118
Volume 2-issue-6-2114-2118
 
Volume 2-issue-6-2108-2113
Volume 2-issue-6-2108-2113Volume 2-issue-6-2108-2113
Volume 2-issue-6-2108-2113
 
Volume 2-issue-6-2102-2107
Volume 2-issue-6-2102-2107Volume 2-issue-6-2102-2107
Volume 2-issue-6-2102-2107
 

Recently uploaded

COMPUTER 10: Lesson 7 - File Storage and Online Collaboration
COMPUTER 10: Lesson 7 - File Storage and Online CollaborationCOMPUTER 10: Lesson 7 - File Storage and Online Collaboration
COMPUTER 10: Lesson 7 - File Storage and Online Collaborationbruanjhuli
 
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019IES VE
 
NIST Cybersecurity Framework (CSF) 2.0 Workshop
NIST Cybersecurity Framework (CSF) 2.0 WorkshopNIST Cybersecurity Framework (CSF) 2.0 Workshop
NIST Cybersecurity Framework (CSF) 2.0 WorkshopBachir Benyammi
 
activity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdf
activity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdf
activity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdfJamie (Taka) Wang
 
Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.YounusS2
 
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCostKubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCostMatt Ray
 
Comparing Sidecar-less Service Mesh from Cilium and Istio
Comparing Sidecar-less Service Mesh from Cilium and IstioComparing Sidecar-less Service Mesh from Cilium and Istio
Comparing Sidecar-less Service Mesh from Cilium and IstioChristian Posta
 
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...Will Schroeder
 
IaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdf
IaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdfIaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdf
IaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdfDaniel Santiago Silva Capera
 
COMPUTER 10 Lesson 8 - Building a Website
COMPUTER 10 Lesson 8 - Building a WebsiteCOMPUTER 10 Lesson 8 - Building a Website
COMPUTER 10 Lesson 8 - Building a Websitedgelyza
 
UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...
UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...
UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...UbiTrack UK
 
Introduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptxIntroduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptxMatsuo Lab
 
Empowering Africa's Next Generation: The AI Leadership Blueprint
Empowering Africa's Next Generation: The AI Leadership BlueprintEmpowering Africa's Next Generation: The AI Leadership Blueprint
Empowering Africa's Next Generation: The AI Leadership BlueprintMahmoud Rabie
 
Building AI-Driven Apps Using Semantic Kernel.pptx
Building AI-Driven Apps Using Semantic Kernel.pptxBuilding AI-Driven Apps Using Semantic Kernel.pptx
Building AI-Driven Apps Using Semantic Kernel.pptxUdaiappa Ramachandran
 
Building Your Own AI Instance (TBLC AI )
Building Your Own AI Instance (TBLC AI )Building Your Own AI Instance (TBLC AI )
Building Your Own AI Instance (TBLC AI )Brian Pichman
 
VoIP Service and Marketing using Odoo and Asterisk PBX
VoIP Service and Marketing using Odoo and Asterisk PBXVoIP Service and Marketing using Odoo and Asterisk PBX
VoIP Service and Marketing using Odoo and Asterisk PBXTarek Kalaji
 
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve DecarbonizationUsing IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve DecarbonizationIES VE
 
OpenShift Commons Paris - Choose Your Own Observability Adventure
OpenShift Commons Paris - Choose Your Own Observability AdventureOpenShift Commons Paris - Choose Your Own Observability Adventure
OpenShift Commons Paris - Choose Your Own Observability AdventureEric D. Schabell
 
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...DianaGray10
 

Recently uploaded (20)

20230104 - machine vision
20230104 - machine vision20230104 - machine vision
20230104 - machine vision
 
COMPUTER 10: Lesson 7 - File Storage and Online Collaboration
COMPUTER 10: Lesson 7 - File Storage and Online CollaborationCOMPUTER 10: Lesson 7 - File Storage and Online Collaboration
COMPUTER 10: Lesson 7 - File Storage and Online Collaboration
 
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
IESVE Software for Florida Code Compliance Using ASHRAE 90.1-2019
 
NIST Cybersecurity Framework (CSF) 2.0 Workshop
NIST Cybersecurity Framework (CSF) 2.0 WorkshopNIST Cybersecurity Framework (CSF) 2.0 Workshop
NIST Cybersecurity Framework (CSF) 2.0 Workshop
 
activity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdf
activity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdf
activity_diagram_combine_v4_20190827.pdfactivity_diagram_combine_v4_20190827.pdf
 
Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.Basic Building Blocks of Internet of Things.
Basic Building Blocks of Internet of Things.
 
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCostKubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
KubeConEU24-Monitoring Kubernetes and Cloud Spend with OpenCost
 
Comparing Sidecar-less Service Mesh from Cilium and Istio
Comparing Sidecar-less Service Mesh from Cilium and IstioComparing Sidecar-less Service Mesh from Cilium and Istio
Comparing Sidecar-less Service Mesh from Cilium and Istio
 
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
Apres-Cyber - The Data Dilemma: Bridging Offensive Operations and Machine Lea...
 
IaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdf
IaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdfIaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdf
IaC & GitOps in a Nutshell - a FridayInANuthshell Episode.pdf
 
COMPUTER 10 Lesson 8 - Building a Website
COMPUTER 10 Lesson 8 - Building a WebsiteCOMPUTER 10 Lesson 8 - Building a Website
COMPUTER 10 Lesson 8 - Building a Website
 
UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...
UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...
UWB Technology for Enhanced Indoor and Outdoor Positioning in Physiological M...
 
Introduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptxIntroduction to Matsuo Laboratory (ENG).pptx
Introduction to Matsuo Laboratory (ENG).pptx
 
Empowering Africa's Next Generation: The AI Leadership Blueprint
Empowering Africa's Next Generation: The AI Leadership BlueprintEmpowering Africa's Next Generation: The AI Leadership Blueprint
Empowering Africa's Next Generation: The AI Leadership Blueprint
 
Building AI-Driven Apps Using Semantic Kernel.pptx
Building AI-Driven Apps Using Semantic Kernel.pptxBuilding AI-Driven Apps Using Semantic Kernel.pptx
Building AI-Driven Apps Using Semantic Kernel.pptx
 
Building Your Own AI Instance (TBLC AI )
Building Your Own AI Instance (TBLC AI )Building Your Own AI Instance (TBLC AI )
Building Your Own AI Instance (TBLC AI )
 
VoIP Service and Marketing using Odoo and Asterisk PBX
VoIP Service and Marketing using Odoo and Asterisk PBXVoIP Service and Marketing using Odoo and Asterisk PBX
VoIP Service and Marketing using Odoo and Asterisk PBX
 
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve DecarbonizationUsing IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
Using IESVE for Loads, Sizing and Heat Pump Modeling to Achieve Decarbonization
 
OpenShift Commons Paris - Choose Your Own Observability Adventure
OpenShift Commons Paris - Choose Your Own Observability AdventureOpenShift Commons Paris - Choose Your Own Observability Adventure
OpenShift Commons Paris - Choose Your Own Observability Adventure
 
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
Connector Corner: Extending LLM automation use cases with UiPath GenAI connec...
 

1904 1908

  • 1. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, No 5, May 2013 1904 All Rights Reserved © 2013 IJARCET Reconfiguration of Memory for High Speed Matrix Multiplication Ms. Shamshad Shirgeri1 ,Ms. Pallavi Umesh Naik2 ,Mrs. Rohini3 , Prof. Krishnananda Shet4 1,2,3 – Student, M.Tech in Industrial Electronics, 4 - Research Scholar, 1,2,3-KLS’s VDRIT,Haliyal affiliated to VTU,Belgaum, 4- Kuvempu University, Shivamogga, Karnataka, India. Abstract The paper entitled “RECONFIGURATION OF MEMORY FOR HIGH SPEED MATRIX MULTIPLICATION“ is basically an enhancement in speed of matrix multiplication. The project aims at reducing the time involved in the operation. The system presented here makes use of three RAMS designated as MAT- RAM-A, MAT-RAM-B and MAT-RAM-C and system also consists of a control circuitry. This system works in two modes i.e., mode-1 and mode-2. In mode-1 the RAM’s act as extensions of existing memory of the processor and in mode-2 the two RAMs MAT-RAM-A and MAT-RAM-B are configured as source for hardware with MAT-RAM-C working as destination for storage or resultant matrix after multiplication. The RAMs are configured back in mode-1, so that data can be read from MAT-RAM-C. Keywords: RAM, MAT-RAM-A, MAT- RAM-B and MAT-RAM-C, SIMD. 1. INTRODUCTION In the present electronic era, speed plays a very crucial role. Speed size ad economics are the present buzzwords in the electronic industry. Any designer will have to take all the three outlined factors into account before starting any design. The trend now-a- days is miniaturization coupled with fast access time. The technique presented here uses additional RAMs which will act as extension of the existing RAMs of computers. These RAM devices are switched in a special configuration to provide high speed matrix multiplication. In vector computers, special vector registers within the CPU are used for handling vector operations. The hardware and associated software will have to take care of transferring data approximately. This is quite an involved task and it involves substantial hardware and software overheads. In SIMD architecture the multiplication of two matrices involves a number of load and store instructions. Moreover this requires N processors for N x N matrix operation. Also in this case the data will have to be stored column wise in each machine and then load and store instructions have to be executed for N times. The technique presented here overcomes these problems and provides a very convenient and economical solution for high speed matrix multiplication. 2. PREAMBLE 2.1 Statement of problem Traditional method for matrix multiplication consumes lot of time, as it involves reading from memory, multiplying them and again storing the result back in memory and this process has to be repeated for each element of the matrix. This process
  • 2. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, No 5, May 2013 1905 All Rights Reserved © 2013 IJARCET obviously wastes processors time. Processors time can be judiciously utilized by the technique presented here. 2.2 Scope of Study This system provides a very convenient and economical solution for high speed matrix multiplication. The speed is quite high and is limited only by the access time of the RAM devices used for this purpose. 2.3 Methodology  Splitting the design into functional modules. The complete design to be implemented to be broken into smaller modules catering to the different functions.  Adding the designed modules to the library. The design modules are added to the library so that they can be accessed whenever necessary. This method reduces length of coding.  Interconnecting all the modules to complete the circuit. The various modules required for are called from the library and then are interconnected as required. 2.4 Limitation of the system  The speed of the system is limited only by the access time of the RAMS used in the circuit. 3. BLOCK DIAGRAM OF RECONFIGURATION OF MEMORY FOR HIGH SPEED MATRIX MULTIPLICATION Figure1. Block Diagram of reconfiguration of memory for high speed matrix multiplication The system presented here consists of  Processor  Switching unit  External control unit  RAM’s This system operates in one of the two modes in mode-1 these can be used as RAM’s of the main processor, connected just as an extension RAM’s of the main processor. In this mode the necessary control signals of the RAM’s are connected to the processors data bus. In mode-2, these are used for matrix multiplication process. In this mode the RAM’s are isolated from the main processor MATRAM-A and MATRAM-B store the input data matrix and output of the operation is stored in MATRAM-C. The necessary control signals are obtained from the external circuit. 3.1 Switching unit The switching part consists of two multiplexers one for transferring data between RAM and external control unit. The other multiplexer is used for routing the control signals to RAM either from the processor or from an external control unit. Depending upon the select input switching unit connect the RAM’s either to the processor or the external control unit. The system operates in two modes, the mode of operation is decided by the input applied to select pin of switching unit. Figure2. Switching Unit
  • 3. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, No 5, May 2013 1906 All Rights Reserved © 2013 IJARCET 3.2 External control unit Figure3. External Control Unit It consists of  A Multiplier  An 8 bit parallel adder  A register  Three counters  A clock 3.2.1 Multiplier Used for the multiplication of the matrix elements. It takes the inputs from the two matrices, namely MATRAM-A and MATRAM-B. It generates the partial product and this becomes the input for the order block. 3.2.2 8 bit Parallel adder Used for the summation purpose. It has two inputs one is partial product from the multiplier block and the other input is the previous sum from the register block. Figure4. 8 bit parallel adder 3.2.3 Register Used for storing the sum from the adder block temporarily. Register initially contains zero. Register gets reset once the resultant matrix element is generated. 3.2.4 Counters The counters used here are binary up counters that generate addresses of RAMs during Mode-2 operation along with the addresses, and bi-directional multiplexers used I the circuit that controls the mode of operation of the system. The counter has a register in it, which should be loaded with the number of elements of the matrix. The counter counts until the loaded value, during which the select signal generated by the counter is at logic high and hence the system operates I mode-2, as soon as the count equals the stored value the select signal becomes low, now the system is switched back to mode- 1. Figure5. Counters 3.3 RAM’s MATRAM-A, MATRAM-B and MATRAM- C which are present in the system are just an extension of existing RAM’s of the main processor during mode-1 and in this mode the address lines are generated by the processor and input data are given by the processor itself. MATRAM-A and MATRAM-B are configured as source for the hardware. MATRAM-C works as destination matrix to store the resultant. During mode-2 their address lines are generated by the external circuit. The RAM’s are again configured back in mode-1, so that data can be read from MATRAM-C.
  • 4. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, No 5, May 2013 1907 All Rights Reserved © 2013 IJARCET 4. MODES OF OPERATION 4.1 Mode-1 In mode-1, the input to the select pin of external circuit is at logic low. Under this condition the RAM’s are connected to the processor. Under this mode, the RAM’s are connected to the processor and are isolated from the external circuit. In this mode the address and control signals for the RAM’s are obtained from the processor, and processor data bus is connected to the RAM’s data bus. Hence reading from or writing into the RAM is under the control of the processor. 4.2 Mode-2 In the mode-2 the input to the select pin is high. Under this condition the RAM’s are connected to the external circuit. The RAM’s are completely isolated from main processor. In this mode, the address of the RAM’s are obtained from the external control unit, the data buses of RAM-A and RAM-B are connected to the multiplier input and the data bus of RAM-C is connected to the output of the adder. 5. VHDL IMPLEMENTATION Figure6. Implementation of Reconfiguration of memory for high speed matrix multiplication using VHDL. 6. GRAPHICAL COMPARISION Figure7. Graphical comparison of Reconfiguration of memory for high speed matrix multiplication
  • 5. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 2, No 5, May 2013 1908 All Rights Reserved © 2013 IJARCET 7. CONCLUSION  It provides a much faster and an economic technique for matrix multiplication operation.  The speed in this technique will be more than ten times faster compared to Pentium processor.  Cost involved will be less compared to that of a vector machine or a SMID machine.  The speed of the operation is only limited by the access time of the RAM chip used in the circuit. 8. REFERENCES  P.H.W. Leong, M.P. Leong, O.Y.H. Cheung, T. Tung, C.M. Kwok, and K.H. Lee, “Pilchard – A Reconfigurable Computing Platform with Memory Interface”.  Xilinx, “Virtex-E 1.8 V Field Programmable Gate Arrays”.  Randall Hyde, "The Art of Assembly Language Programming". http://webster.cs.ucr.edu/Page_AoA Linux/HTML/AoATOC.html  IEEE international conference on “An FPGA based floating-point processor array supporting a high-precision dot product” by Fritz Mayer.