replica nfl jerseysreplica nfl jerseyssoccer jerseyreplica nfl jerseys forums.silverfrost.com :: View topic - error with /opt with derived type
forums.silverfrost.com Forum Index forums.silverfrost.com
Welcome to the Silverfrost forums
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

error with /opt with derived type

 
Post new topic   Reply to topic    forums.silverfrost.com Forum Index -> Support
View previous topic :: View next topic  
Author Message
JohnCampbell



Joined: 16 Feb 2006
Posts: 2615
Location: Sydney
PostPosted: Thu Jul 11, 2013 2:49 am    Post subject: error with /opt with derived type Reply with quote
Paul,

To check my program, I am compiling most of my program with /check, but trying to use /opt for a few routines that are compute intensive. With these compile options mix, the following code example does not appear to work correctly with /opt. the module was compiled with /check

Code:

      count_best = 0
      count_last = 0
      sum_ri     = 0
!
      do id = 1,num_node
         if (mod (id,500000) == 0) write (*,2001) 'Checking sector allocation',id
!
         x  = (node_data(id)%ix - 1) * grid_dx + xmin_g
         y  = (node_data(id)%iy - 1) * grid_dx + ymin_g
!
!      find nearest segment centre
         ri = 1.e30
         ik = 0
         do k = 1,channel_no+1
            rk = (x - sector_data(k)%x1)**2 + (y - sector_data(k)%y1)**2
            if ( rk >= ri ) cycle
            ri = rk
            ik = k
         end do
!
         count_best(ik) = count_best(ik) + 1
         count_last(k)  = count_last(k)  + 1
         sum_ri         = sum_ri + sqrt(ri)


I have included count_best, count_last and sum_ri stats to confirm the behaviour.
with /opt, ik is always 1, implying rk is not being updated (correctly calculated) for each itteration of k. average of ri is 30,498 (metres)
I have checked that the itteration goes for 730 cycles (channel_no+1)

with /debug used for the module and routine, ik is being correctly calculated in the range 1:730 and ri averages 160 metres

This is a large program with a 2.5gb data set so not easy to send.
num_node = 7,548,638

Any suggestions, as I would like to be able to use /opt for some key loops.

John
Back to top
View user's profile Send private message
PaulLaidler
Site Admin


Joined: 21 Feb 2005
Posts: 8210
Location: Salford, UK
PostPosted: Thu Jul 11, 2013 7:22 am    Post subject: Reply with quote
John

I don't have any suggestions about what is going wrong. /opt can only optimise locally so should have no effect on other subprograms.

At a quick glance, I don't think that I can reproduce the effect without further information. If you are able to provide a working sample then I would be happy to investigate further.
Back to top
View user's profile Send private message AIM Address
JohnCampbell



Joined: 16 Feb 2006
Posts: 2615
Location: Sydney
PostPosted: Thu Jul 11, 2013 7:30 am    Post subject: Reply with quote
Paul,

the problem is when using /opt for the following loop
Code:
!      find nearest segment centre
         ri = 1.e30
         ik = 0
         do k = 1,channel_no+1
            rk = (x - sector_data(k)%x1)**2 + (y - sector_data(k)%y1)**2
            if ( rk >= ri ) cycle
            ri = rk
            ik = k
         end do


I suspect that /opt is optimising out the recalculation of rk at each step of loop k, as with /opt, at the end of the loop; ik=1 and k = 731

John
Back to top
View user's profile Send private message
PaulLaidler
Site Admin


Joined: 21 Feb 2005
Posts: 8210
Location: Salford, UK
PostPosted: Thu Jul 11, 2013 8:22 am    Post subject: Reply with quote
Try doing the calculation in different ways.
Does the /explist help in this situation?
Back to top
View user's profile Send private message AIM Address
JohnCampbell



Joined: 16 Feb 2006
Posts: 2615
Location: Sydney
PostPosted: Fri Jul 12, 2013 4:05 am    Post subject: Reply with quote
Paul,
I can confirm that rk is not being updated for each itteration of k.
I have changed the code for this loop to:
Code:
         print_this = in_list ( id, print_list, 10)
!
         do k = 1,channel_no+1
            rk = (x - sector_data(k)%x1)**2 + (y - sector_data(k)%y1)**2
            if (print_this) write (98,*) id, k, ik, ri, rk
            if ( rk >= ri ) cycle
!
            if (kd > 0) num_better = num_better + 1    ! count changes to existing definitions
            ri = rk
            ik = k
         end do
         if (print_this) write (98,*) id, k, ik, ri, rk

For a selected node, this gives a report like:
Code:

       255565           1          22          7114407.26500              4393344.85000
       255565           2           1          4393344.85000              4393344.85000
       255565           3           1          4393344.85000              4393344.85000
       255565           4           1          4393344.85000              4393344.85000
       255565           5           1          4393344.85000              4393344.85000
....
       255565         730           1          4393344.85000              4393344.85000
       255565         731           1          4393344.85000              4393344.85000

If this routine is compiled with /debug, I get:
Code:
       255565           1          22          7114407.26500              4393344.85000
       255565           2           1          4393344.85000              3986640.05000
       255565           3           2          3986640.05000              3599935.25000
       255565           4           3          3599935.25000              3233230.45000
       255565           5           4          3233230.45000              2886456.78500
...
       255565          22          21          31932.6849000              22340.5961000
       255565          23          22          22340.5961000              32751.1042000
       255565          24          22          22340.5961000              63159.6970000
...
       255565         730          22          22340.5961000              2676905821.09
       255565         731          22          22340.5961000              2676905821.09


I can overcome the problem by not using /opt, but the real problem is that /opt does not work for this situation, which I think is derived types.
Can some of these /opt short-comings be investigated?

John
Back to top
View user's profile Send private message
PaulLaidler
Site Admin


Joined: 21 Feb 2005
Posts: 8210
Location: Salford, UK
PostPosted: Fri Jul 12, 2013 4:49 pm    Post subject: Reply with quote
I will make a note of this for investigation.
Back to top
View user's profile Send private message AIM Address
JohnCampbell



Joined: 16 Feb 2006
Posts: 2615
Location: Sydney
PostPosted: Mon Jul 15, 2013 12:57 pm    Post subject: Reply with quote
Paul,

I have sent an email with the code for this problem. It includes the module required to define the derived types.

John
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    forums.silverfrost.com Forum Index -> Support All times are GMT + 1 Hour
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group