Sie sind auf Seite 1von 11

Cvervlew

ara||e| rocess|ng
|pe||n|ng
frfnerlslns of Mullpronessors
lnernonnenlon Srunures
lner pronessor Arblrflon
lner pronessor ommunlnflon fnd Synnronlzflon
9frfllel 9ronesslng
eve|s of ara||e| rocess|ng
Iob or rogram |eve|
1ask or rocedure |eve|
InterInstruct|on |eve|
IntraInstruct|on |eve|
Lxecut|on of concurrent vents |n the comput|ng process to ach|eve
faster computotiono/ 5peed
1he purpose |s to speed up the computer process|ng capab|||ty and
|ncrease |ts throughput (amount of process|ng that can be
accomp||shed dur|ng a g|ven |nterva| of t|me)
9frfllel ompuers
rch|tectura| C|ass|f|cat|on
|ynns c|ass|f|cat|on
Cons|ders the organ|zat|on by the number of |nstruct|ons and data
|tems that are man|pu|ated s|mu|taneous|y
W 8ased on the mu|t|p||c|ty of nstruction 5treoms and uoto 5treoms
W Instruct|on Stream
Sequence of Instruct|ons read from memory
W Data Stream
Cperat|ons performed on the data |n the processor
Number of uoto 5treoms
Number of
nstruction
5treoms
S|ng|e
Mu|t|p|e
S|ng|e Mu|t|p|e
SISD SIMD
MISD MIMD
SlSu
Contro|
Un|t
rocessor
Un|t
Memory
Instruct|on stream
Data stream
ara||e| process|ng |s ach|eved by means of mu|t|p|e funct|ona| un|ts or by
p|pe||ne process|ng
Character|st|cs
Instruct|ons and data are stored |n memory
Cne operat|on at a t|me
|m|tat|ons
Max|mum speed of the system |s ||m|ted by the ,emory 8ondwidth (b|ts]sec or
bytes]sec)
|m|tat|on on ,emory 8ondwidth
Memory |s shared by CU and I]C
MlSu
M CU
M CU
M CU
-
-
-
-
-
-
Memory
Instruct|on stream
Data stream
Character|st|cs
1here |s no computer at present that can be c|ass|f|ed as MISD
SlMu
Contro| Un|t
Memory
||gnment network
- - -
M M M - - -
Data bus
Instruct|on stream
Data stream
rocessor un|ts
Memory modu|es
Character|st|cs
Cn|y one copy of the program ex|sts
s|ng|e contro||er executes one |nstruct|on at a t|me
Shared memory conta|ns mu|t|p|e modu|es to commun|cate s|mu|taneou|sy
w|th
MlMu
Interconnect|on Network
M M M
- - -
Shared Memory
Character|st|cs
Mu|t|p|e process|ng un|ts
severa| programs are processed at same t|me
Lxecut|on of mu|t|p|e |nstruct|ons on mu|t|p|e data
1ypes of MIMD computer systems
Shared memory mu|t|processors
Messagepass|ng mu|t|computers
9lpellnlng
81 A
l
82 8
l
Lofd A
l
fnd 8
l
83 81 * 82 84
l
Mullply fnd lofd
l
83 83 + 84 Add
A ennlque of denomposlng f sequenlfl proness lno sub operflons wl
efn sub proness belng exenued ln f pfrlfl dedlnfed segmen f
operfes nonnurrenly wl fll oer segmens

|
* 8
|
+ C
|
for | 1 2 3 7

|
k1 k2
Mu|t|p||er
k3 k4
dder
kS
Memory
8
|
C
|
Segment 1
Segment 2
Segment 3
Cperflons ln efn 9lpellne Sfge
CIock
PuIse
Segment 1 Segment 2 Segment 3
Number R1 R2 R3 R4 R5
1 A1 B1
2 A2 B2 A1 * B1 C1
3 A3 B3 A2 * B2 C2 A1 * B1 + C1
4 A4 B4 A3 * B3 C3 A2 * B2 + C2
5 A5 B5 A4 * B4 C4 A3 * B3 + C3
6 A6 B6 A5 * B5 C5 A4 * B4 + C4
7 A7 B7 A6 * B6 C6 A5 * B5 + C5
8 A7 * B7 C7 A6 * B6 + C6
9 A7 * B7 + C7
enerfl 9lpellne
enera| Structure of a 4Segment |pe||ne
S k
1 1
S k
2 2
S k
3 3
S k
4 4
Input
C|ock
Space1|me D|agram
1 2 3 4 S 6 7 8 9
11
11
11
11
12
12
12
12
13
13
13
13 14
14
14
14 1S
1S
1S
1S 16
16
16
16
C|ock cyc|es
Segment 1
2
3
4
9lpellne Speedup
n Number of tasks to be performed
Convent|ona| Mach|ne (Non|pe||ned)
t
n
C|ock cyc|e
:

1|me requ|red to comp|ete the n tasks


:

n * t
n
|pe||ned Mach|ne (k stages)
t
p
C|ock cyc|e (t|me to comp|ete each suboperat|on)
:

1|me requ|red to comp|ete the n tasks


k segments and n nos of tasks
:

(k + n 1) * t
p
Speedup
S
k
Speedup
S
k
n*t
n
] (k + n 1)*t
p

Das könnte Ihnen auch gefallen