Advanced Fortran Programming : 008 : Forall statement and comparison with do loop Video Lecture | Introduction to Fortran Programming (AdvancedLevel) - Database Management

hello guys my name is Aaron and welcome
to my channel this series is a series of
tutorials on advanced Fortran
programming now in this tutorial we be
looking at the for all for all statement
in Fortran now up to four times 77 and
and before and after that there wasn't
any far loop statement available in
Fortran and that was actually taken care
by this do loop but after the
improvement of the development and
initiation of far loop statements in
other programming languages like C or
C++ and several others okay
the importance of for loop statement
became even more widely acknowledged so
what happened is that in Fortran in the
Fortran synastry photon development
history from 495 onwards for to 90 and
95 onwards this for all statement came
into picture after four and 77 by
definitions you don't even have this
four loop statement in Fortran 90 and 95
and from the subsequent video tutorial
feature series of Fortran for all
statement was included and what it does
is that instead of instead of the
conventional far loop statement that
follow-up statements provide another for
programming languages this far for our
statement acts like
acts as a as a single state single
structure or a constant the looping
looping construct that makes that allows
you to have what you call nesting in
within a single line without having that
without having to use multiple lines
okay
now I'll explain what I'll explain you
guys what do you mean what I mean this
in this tutorial what I have here as I
just have a small program I have defined
an integer a parameter of integer car
with valid three and because of this
value these four matrices a B b1 and b2
these are all real matrices of kind
eight and these are all three double 3
cross 3 matrices and I mean it is and
define 2 into this I comma J here now I
define a to be
3 cross 3 matrix which in this reshaped
comment setting be to to be at equal to
a and I setting all the values of B to B
minus 1 now I'm using this print
underscore back to subroutine to print
the matrix a proper matrix format and if
you guys know why I want to know where
this matrix printed asked where my to is
done this subroutine is at the bottom
this is exactly the same subroutine we
saw in one of the previous tutorials to
print the matrix in the proper matrix
from it nothing more nothing more okay
and now what I do what I have here is
that I just have a looping statement the
for all statement it starts here and
ends here the notation the syntax is
follows you have to write the first
keyword and then within brackets you
have to write the I traitors waiting
from where to where
here here instead of this comma state
come on you have to give it in :
that's man that's a necessity in for all
if you just give it and comma and if you
have to compile this you get an error
you guys notice over here you'll get an
error but if I were to but here since
because of some important considerations
okay they give they give the kollel
notation over here which you have to
keep in mind okay and you had to write
the I iterator variable from from the
hydration variable how it varies from
one value to the other and just give a
comma and you have a second hydrating
variable use it over here
use it over here suppose you have a
third hydrating variable give a comma
and say K something equals like one is
two and something of that sort and so on
if you have a for that dating will give
a comma proceed proceed on and just like
that go on with it yes they will go on
yeah and simple as that so instead of a
nester developing kind of a thing you
can just do a simple for all statement
like this okay
now what it does is that it is this loop
the for all the values of I and J listed
with between between this iteration
range okay this statement expression
inside this for all an end for
is done accordingly and then those
missiles are stored in j those those
little store done done accordingly now
to prove you guys this this one single
for our statement is same as that of
this do loop state this nested do loop
statement I written this say I returned
the same notation same stuff same stuff
with the same operation I'm doing here
with in a nested do loop simple as that
simple as that
but instead of B I'm storing it in b1
now if I print a B & B 1 and see how
what how they look like we'll get an
idea to get an idea whether b and b b1
are different or not okay that's about
it now what I have done here is they're
using the same command subroutine cover
on the top I am printing the matrix B
and printing the matrix B 1 and we'll
get to this part in a minute we'll get
to this part in a minute
it's slightly different but now just
constant it on this part for now we were
to compile build and execute this okay
now what it does is that it prints the
matrix a as per this format trying
because this is column contagious column
contagious in Fortran so it prints the
value 1 2 3 4 by 6 4 5 6 and 10
along the columns and then it prints the
B matrix it does all the calculations
for B matrix along this under this for
loop and prints the B matrix over here
the sublet B matrix subroutine call is
over here and and that is getting
printed over here so this a so look at
the values now to ensure that this
matrix is same and that this is nested
do looping over here with b1 okay and if
you guys notice beavin is over here and
if you guys noticed these two matrices B
& B 1 are exactly the same because we're
doing exactly I we get exactly the same
operation on both parts so it does the
same now how is it different from do
loop
well there are a few differences okay
we'll come to a and B too little later
to come to that now what you can do is
that you can do a masking statement also
masking statement also something like
this or you can put a condition as to
that to skip and Van not to skip and all
for the follow-up statement for this for
all statement whereas in do loop if you
want to do you wanna do a condition like
that you may have to write a condition
over here with an X if statement or
something of that sort and if you have
to write sometimes you may have to write
in a statement over here depending on
the kind of program but here when for
all statement all I will do is that if
you have a condition just had to give a
condition like this okay
now are then doing is that I've setting
B value to be negative one over here for
all the values of B I'm setting it to be
negative one okay and this loop will
execute all the time except except when
yeah I comma J equals not equals eight
only when that value reaches B will not
be this B I comma J set will not be
executed but for all other statements
here this looping will execute don't
believe me watch compiled I'm comparing
and building and executing this watch
now matrix in matrix a a I comma J is a
eight at here at the when I equals two
and J equals three now to show that that
this part is to show that and this index
B I comma J does not work look at the B
matrix over here look at the B matrix
over here in this question aware where
it corresponds to invalid corresponds in
a a this this value has eight here and
wherever wherever eight comes in matrix
a okay
that value is remains unaltered and this
-1 is because we initially said that the
entire matrix should be minus one except
in that location everywhere else the
operation the operation which is
supposed to give us a matter similar to
be one has app has happened ok simple as
that
this is the kind of a masking thing can
do and if you guys if you guys want to
make this a little more
complicated some I'll let me put a new
condition and I not equal to two
no not equal to two something of that
sort
so this will do all the statement
although job in such a manner that I is
not equal to two and they come a I comma
J not equal to eight
except for them and these two conditions
are reached for everywhere else this
looping statement works okay and if you
guys don't believe me again compile
build and execute this now watch now in
this entire row I is 2 so in this
operation what happens it is the entire
row is skipped and fortunately
fortunately this value is also skipped
ok now ok let me write this as I not
equal to 3 compile build and execute
this because of this condition over here
this value is an altered because of this
condition over here this value is
unaltered kind of ok only when this
entire condition has worked it works
kind of fine now let me just you know
come on this comment this part out come
on this part out thereby they are
getting the same answer now what what I
also want to tell you is that suppose if
you are using this for all statement to
modify that matter the same matrix alone
same as a matrix alone the results will
be little different when you do it in
the do loop I now I will explain this
concept with in a comparison with the do
loop and do loop and the new matrix ok
now watch here now if you guys notice
I've set the value of a here and I'm
assigning the same value of a to the
matrix B 2 so identically matrix a and B
matrix B 2 are exactly the same they
just have the same value but they have
different names nothing that's all ok so
that being said okay let me walk you
guys over here this
in this loop in this for all statement I
modifying the matrix contents of AI a
using the matrix a value itself okay and
now I'm printing the matrix a over here
now in this nested do loop what I'm
doing is that I'm printing and modifying
the matrix B to using the values of
matrix B to itself now intuition might
have might have said that intuition
might have said that okay we as we saw
and look last we saw in the top since
for all and do loop can behave sim sim
in a similar manner this modified a and
this modified B will be same okay but to
be honest they will not be the same okay
now watch throughout the compile and
execute this okay this is whether even
these two sections come in temperature
and if you guys notice notice in this
look in this location in this location
this is where a is used this is
different whereas in this location and
in this location b2 is different so the
Mattocks be a 1 and B 2 after them after
the modification using their own values
they're entirely different reason it's
because do loop when one map when the
new loop when using this do loop when B
to us okay when b2 is modified okay
it updates the value of b2 as and when
it's modified okay but in this for all
statement then a is modified in this
command using the mod using the values
of a itself what happens is that before
the loop entire loop begins to operate a
separate copy of a it stores elsewhere
okay and then using those values nuke
new corresponding values of a are being
calculated using this loop and this new
set of Catalan value is being stored to
a and as a consequence as a consequence
of that the for every I tration your a
IJ will not be modifying will not be
will not at all
modified okay and only at the end of
this loop you're AI J gets modified you
mean you're all there and all the
entries of your mattress gets modified
whereas a do loop it gets modified after
each and every hydration so this kind of
updates each and every edition whereas
this in this operation this for all
statement updates the updates entire
matrix only after the entire exam lupus
a entire construct is over so this is
something of importance you have to keep
in mind and because of this region this
for all is kind of a hybrid between you
do do loop and the for loop and other
languages and all so this kind of stuff
might be useful will be very useful
especially do it especially when you
were trying to calculate something what
do you call or you call time time magic
solutions and other stuff this kind of
loop will be definitely useful for
storing time time matching solutions our
finite element difference procedures on
2d finite difference methods procedure
for 2d and the similar stuff you know
and even for doing Gauss elimination of
God's Jacobi as a Gauss siedel goddess
if God siedel Gauss Jacobi methods and
all this kind of be useful and then this
will this kind of follow-up experiment
might be used to express is specifically
in Gauss Jacobi method whereas this kind
of a looping statement can be used in
Gauss siedel method at all okay I'm just
giving an example but you guys will work
around work around with it
so for these kind of statements where
and you wanna have to have this query
and we have we need a condition where
and you're Maddox has to be not should
not be changed during all the
intermediate values and should be kind
of it go for for all statement whereas
if you want a man if you want to alter
the matrix value such that each and
every value has been modified and the
same modification is being updated go
for do loop so these are kind of this is
a kind of a comparison and the stark
difference with
meanwhile up for all statement on a do
loop statement which have to keep in
mind and application wise some does in
some places this might work and some
places this kind of feature words use
them accordingly and use them
accordingly and this that that does not
all there are other features also and
you can also go for a single line
statement of for all what he can do is
that in this line sorry
control said okay I think this this also
okay it doesn't work okay I think this
has to be in the same line now let's see
yeah it works suppose if you were to
write a four-hour statement in a single
line you can write like this probably
there's only one condition suppose if
you had to write this in a multiple line
or something you have to use end for all
okay this is clear-cut distinction you
have to keep in mind another clear-cut
distinction have to keep in mind that's
it that's it now that's all that's and
if you want to write a for all statement
like a do loop statement then again it's
possible but when you write it in that
format I think the result gets modified
very much so to prove that let's do one
thing okay let's write it as B three
again new matrix B 3 that B 3 B equal to
a also ok now what I do is that I copy
this statement over here printed over
here now let me just change this okay
now I use a nested for for all nested
for all J equals 1 is to n ok and then I
said and for all
and for all now now I'm setting this to
beat Abby what was this value B 3 P 3 P
3 and this is again B 3 now I'm printing
matrix B 2 over here copy this copy this
paste this and I print this matrix B 3
ok let's see if there are any
differences executing it any user for
all statement or something of that sort
okay even your mat even your matrix B 3
even your matrix B 3 is getting updated
not only at the end so even when you
even when you new do a nested operation
okay even when you do a nested operation
with for all okay you can see that this
is entirely the same as this so even
when you do a nested operation the
matrix values do not get updated only
after you do not get updated after each
iteration whether they get updated only
at the end of the sequence of for all
statements so that way if you guys not
this matrix a and matrix B are same
indicating that this updation does not
take place immediately rather it takes
place at the end no that's all I have
for you guys in this tutorial hope you
guys learned a lot of differences and
similarities between for for all
statement and do loop and hope you guys
know now you realize to where you can
use or for all statement Val to use a do
loop statement and all thank you guys
for watching and I'll see you guys in
next tutorial and next tutor will be
looking at some important concepts
regarding procedure in current concepts
regarding procedures or sub procedures
and procedures like functions and sub so
yeah I will look at procedures some
features in it and
most possibly the save function the save
function available in Fortran also thank
you guys for watching and see you guys
next time
bye