While compiling openacc code, i am getting following warnings
215, Scalar last value needed after loop for x at line 239
Scalar last value needed after loop for y at line 239
Scalar last value needed after loop for x at line 240
Scalar last value needed after loop for y at line 240
Scalar last value needed after loop for x at line 242
Scalar last value needed after loop for y at line 242
Scalar last value needed after loop for x at line 246,248
Scalar last value needed after loop for y at line 248,252
Does such warnings makes program to run sequentially how to use lastvalue clause in OpenACC ?
Yes, these warnings can cause the code to be run sequentially (see the compiler feedback output from -Minfo=accel assuming you're using PGI). By default, scalars are first private. However, if the scalar's value is used outside the OpenACC compute region, the compiler can't automatically privatize the scalar since it doesn't know which value to use.
This can occur when the scalar variable is used after the compute region, for example
#pragma acc parallel loop
for (int i=...
x = <expr>
...
}
printf("Final X=%d\n",x);
You can work around this my putting "x" inside a private clause, but the value printed for "x" will remain unchanged from it's value before the loop.
Another scenario in which this can occur, is when updating a global scalar inside a compute region. In this case, you would want to put the variable in a data clause (such as "copy") to have the variable be shared by all threads, and then use the OpenACC "atomic" directive when updating the value.
A third scenario is when a scalar is passed by address to a device subroutine. In this case, the compiler must assume that other references are made the scalar. While most likely you're not going to assign a global pointer to the scalar, it's possible and since the compiler does not have visibility into the subroutine, it must assume it. To fix, add the scalar to an OpenACC "private" clause, or change the code to pass the scalar by value. Note that by default Fortran passes arguments by address. To pass by value, use the F2003 "value" attribute on the argument's declaration.
If you have a different scenario than what's listed above, please provide example code illustrating the problem.
Related
Every single time I load my program, even for the fist time, it says
file.rb:9: warning: already initialized constant W_mum
file.rb:6: warning: previous definition of W_mum was here.
a little help here?
W_mum = gets.to_i
elsif (W_mum = 1)
Ruby uses two different "storage bins" for data: variables and constants. In your source code, you can identify them y their first letter: constants always have a capital letter at the start of their name, variables a lower-case letter.
In your case, you thus have a constant named W_mum. Now, when you first set a value to a constant and then later set a different value to it, Ruby will show a warning (as such: you can set new values to constants, but you should not).
Now, as for why Ruby warns here: in your elsif, you are actually assigning the constant the value 1. This might be a bug though. Instead of an assignment with =, you likely intended to use a comparison here, using the == operator.
I am working on a large Fortran code, where parts are written in FORTRAN77.
There is a piece of code, which causes debugger to raise errors like:
Fortran runtime error:
Index '2' of dimension 1 of array 'trigs' above upper bound of 1
but when compiled without debugging options runs and does not crash the program. Debugging options used:
-g -ggdb -w -fstack-check -fbounds-check\
-fdec -fmem-report -fstack-usage
The logic of the problematic piece of code is following: in file variables.cmn I declare
implicit none
integer factors,n
real*8 triggers
parameter (n=32)
common /fft/ factors(19), triggers(6*n)
Variables factors and triggers are initialized in procedure initialize:
include 'variables.cmn'
...
CALL FFTFAX(n,factors,triggers)
...
FFTFAX is declared in another procedure as:
SUBROUTINE FFTFAX(N,IFAX,TRIGS)
implicit real*8(a-h,o-z)
DIMENSION IFAX(13),TRIGS(1)
CALL FAX (IFAX, N, 3)
CALL FFTRIG (TRIGS, N, 3)
RETURN
END
and lets look at procedure FFTRIG:
SUBROUTINE FFTRIG(TRIGS,N,MODE)
implicit real*8(a-h,o-z)
DIMENSION TRIGS(1)
PI=2.0d0*ASIN(1.0d0)
NN=N/2
DEL=(PI+PI)/dFLOAT(NN)
L=NN+NN
DO 10 I=1,L,2
ANGLE=0.5*FLOAT(I-1)*DEL
TRIGS(I)=COS(ANGLE)
TRIGS(I+1)=SIN(ANGLE)
10 CONTINUE
DEL=0.5*DEL
NH=(NN+1)/2
L=NH+NH
LA=NN+NN
DO 20 I=1,L,2
ANGLE=0.5*FLOAT(I-1)*DEL
TRIGS(LA+I)=COS(ANGLE)
TRIGS(LA+I+1)=SIN(ANGLE)
20 CONTINUE
In both FFTFAX and FFTRIG procedures there are different bounds for dimensions of arguments than the actual input array size (for TRIGS it is 1 and 19, respectively).
I printed out TRIGS after calling FFTFAX in no-debugger compilation setup:
trigs: 1.0000000000000000 0.0000000000000000\
0.99144486137381038 0.13052619222005157 0.96592582628906831\
0.25881904510252074 0.92387953251128674 0.38268343236508978\
...
My questions are:
Is notation :
DIMENSION TRIGS(1)
something more than setting bound of an array?
Why is the program even working in no-debugger mode?
Is setting:
DIMENSION TRIGS(*)
a good fix if I want variable trigs be a result of the procedure?
In f77 statements like the DIMENSION TRIGS(1) or similar or ..(*) with any number, if pertaining an argument of the procedure just tells the compiler
the rank of the array, the length in memory must be assigned to the array which is given in the call of the subroutine, normally f77 does not check this!
My recommendation either use (*) or better reformat (if necessary) the f77 sources to f90 (the bits shown would compile without change...).
and use dimension computed using n in the declaration within the subroutines/procedures.
Fortan passes arguments by address (i.e. trigs(i) in the subroutine just
will refer on the memory location, which corresponds to the address of trigs(1) + i*size(real*8).
A more consisted way to write the subroutine code could be:
SUBROUTINE FFTRIG(TRIGS,N,MODE)
! implicit real*8(a-h,o-z)
integer, intent(in) :: n
real(kind=8) :: trigs(6*n)
integer :: mode
! DIMENSION TRIGS(1)
.....
PI=2.0d0*ASIN(1.0d0)
.....
or with less ability for the compiler to check
SUBROUTINE FFTRIG(TRIGS,N,MODE)
! implicit real*8(a-h,o-z)
integer, intent(in) :: n
real(kind=8) :: trigs(:)
integer :: mode
! DIMENSION TRIGS(1)
.....
PI=2.0d0*ASIN(1.0d0)
.....
To answer your question, I would change TRIGS(1) to TRIGS(*), only to more clearly identify array TRIGS as not having it's dimension provided. TRIGS(1) is a carry over from pre F77 for how to identify this.
Using TRIGS(:) is incorrect, as defining array TRIGS in this way requires any routine calling FFTRIG to have an INTERFACE definition. This change would lead to other errors.
Your question is mixing the debugger's need for the array size vs the syntax excluding the size being provided. To overcome this you could pass the array TRIGS's declared dimension, as an extra declared argument, for the debugger to check. When using "debugger" mode, some compilers do provide hidden properties including the declared size of all arrays.
I'm trying to run a FOR loop on robot framework depending of the status of another variable.
${STATUS1}= Run Keyword And Return Status Should Be Equal As Strings ${CELLVALUE} ${EXPECTEDVALUE}
\ ${COUNT}= Set Variable If '${STATUS1}' == 'True' ${COUNT}+1
\ ... '${STATUS1}' == 'False' ${COUNT}+0
But all I get is '''0'+0'+0'+1 or similar, even if I use Run keyword If and Evaluate instead of set var, I tried to convert to integer but nothing happens and I cannot convert it to integer or number. Any suggestions? thanks in advance!
It looks like you're simply wanting to increment ${COUNT} if ${CELLVALUE} equals ${EXPECTEDVALUE}. That can be done pretty easily with Set Variable if
If you know that ${CELLVALUE} and ${EXPECTEDVALUE} are of the same internal type (eg: strings or ints), and you're using robot framework 2.9 or greater, you can write it like this:
${COUNT}= Set variable if $CELLVALUE == $EXPECTEDVALUE
... ${COUNT+1} ${COUNT}
This assumes that ${COUNT} is initialized to an integer value, which you can do by assigning it the value ${0}
If you don't know the type, can't guarantee the type, or are using an older version of robot, you can use triple-quoted strings to coerce the values to strings:
${COUNT}= Set variable if '''${CELLVALUE}''' == '''${EXPECTEDVALUE}'''
... ${COUNT+1} ${COUNT}
Of course, you could use Run Keyword and Return Status like in your example, and then compare the status. That seems like an unnecessary extra step, but it might make sense in your actual test.
The point being, you can use Set variable if and extended variable syntax to solve this problem.
Note 1: With Set variable if, two values are provided. The first value is assigned if the expression is true, the second one is assigned if the value is false. The second value is the original variable, meaning it won't be changed. If you don't provide the second value, the variable will be set to None.
Note 2: Putting an expression inside curly braces (eg: ${COUNT+1} is documented in rule 4 of extended variable syntax.
Note 3: Starting with robot framework 2.9, variables are available in the evaluation namespace with the simplified syntax $varname. So, the robot variable ${CELLVALUE} can be used in python expressions as $CELLVALUE. This is documented in the section Evaluating Expressions in the BuiltIn library documentation.
When using the OpenFile function in Go's os package, what exactly is the purpose of the pipe character?
Example:
os.OpenFile("foo.txt", os.O_RDWR|os.O_APPEND, 0660)
Does it serve as a logical OR? If so, does Go choose the first one that is "truthy"?? Being that the constants those flags represent, at the heart of them are just integers written in hexadecimal, when compiled how does Go choose which flag to apply?
After all, if the function call were to go by the largest number, os.O_APPEND would take precedence over all other flags passed in as seen below:
os.O_RDWR == syscall.O_RDWR == 0x2 == 2
os.O_APPEND == syscall.O_APPEND == 0x400 == 1024
os.O_CREATE == syscall.O_CREAT == 0x40 == 64
UPDATE 1
To follow up on the comment below, if I have a bitwise operator calculation using os.O_APPEND|os.O_CREATE will that error if the file exists, or simply create/append as needed?
UPDATE 2
My question is two fold, one to understand the purpose of the bitwise operator, which I understand now is being used more as a bitmask operation; and two, how to use the os.OpenFile() function as a create or append operation. In my playing around I have found the following combination to work best:
file, _ := os.OpenFile("foo.txt", os.O_RDWR|os.O_CREATE|os.O_APPEND, 0660)
file.WriteString("Hello World\n")
file.Sync()
Is this the correct way or is there a more succinct way to do this?
It is a bitwise, not a logical OR.
If you write out the numbers in binary, and assign each a truth value 0/1, and apply the logical OR to each of the bits in place i between the arguments, and then reassemble the result into an integer by binary expansion - that's the | operator.
It is often used in a way that is commonly described as a "bitmask" - you use a bitmask when you want a single int value to represent a (small) set of switches that could be turned on or off. One bit per switch.
You should see in this context, A | B means "all the switches in A that are on, as well as all the switches in B that are on". In your case, the switches define the exact behavior of the file open/creation function, as described by the Go manual. (And probably more in detail by the Unix manpage I linked above).
In a bitmask, constants are typically defined that represent each switch - that's how those O_* constants are determined. Each is an int with exactly one bit set and represents a particular switch. (though, be careful, because sometimes they represent combinations of switches!).
Also:
^A // All of the "switches" not currently on in A
A&^B // All of the "switches" on in A but not on in B
A^B // All of the "switches" on in exactly one of A or B
, etc.
The operator | itself is described in the Go manual here.
It is a bitwise OR operator. Its purpose being used here is to allow for multiple values to be passed as a bitmask. Thus you can combine flags to create a desired result such as using the OpenFile() function to create a file if it does not exist or append to it if it does.
os.Openfile("foo.txt", os.O_RDWR|os.O_CREATE|os.O_APPEND, 0660
The constants being passed as arguments from the os package are assigned values from the syscall package. This package contains low-level operating system independent values.
Package syscall contains an interface to the low-level operating system primitives. The details vary depending on the underlying system, and by default, godoc will display the syscall documentation for the current system. If you want godoc to display syscall documentation for another system, set $GOOS and $GOARCH to the desired system. For example, if you want to view documentation for freebsd/arm on linux/amd64, set $GOOS to freebsd and $GOARCH to arm. The primary use of syscall is inside other packages that provide a more portable interface to the system, such as "os", "time" and "net".
https://golang.org/pkg/syscall/
As noted by #BadZen, a bitwise OR operator, in this case the '|' character, acts at the binary level changing any 0 values to 1's that are not already ones.
You should see in this context, A | B means "all the switches in A that are on, as well as all the switches in B that are on".
By doing this as the function above displays, you are manipulating the behavior of the function to create a file (os.O_CREATE) with the given name of foo.txtor open the file for reading/writing (os.O_RDWR) and any value written to it will be appended (os.O_APPEND). Alternatively you could pass along os.O_TRUNC in order to truncate the file before writing.
The bitwise OR operator allows you a powerful solution to combining different behaviors in order to get the result from the function that you are desiring.
Is it possible in Visual Foxpro to have 2 variables that point to the same address in memory. Such that if the value of one of the variables is changed then the other is also changed. I understand that when passing arguments to functions they can be passed by value or reference but I want to know if this is possible in straight code. I think in other languages such as C this is called a pointer but I don't believe VFP has pointers. So if one writes the following code it will output the number 4.
a=4
b=a
a=6
? b && answer 4
But could one write code such as the following where the answer could be 6?
a=4
b=*a && note the inclusion of the asterisk (pointer?) here which won't compile in VFP
a=6
? b
No. There are no pointers or references in Foxpro; as you note, the closest thing to it is passing parameters by reference to functions. You might be able to try to kludge something together (as Jerry mentions) with objects using Access/Assign methods, but even then, all that gets passed to the Assign method is the value being assigned - nothing about whether it was originally another variable, a literal value, an object's property, etc.
You could simulate it by using an array or a table. The variables would contain only the array index or record number (or other index) as a reference, and you'd have to get the actual value from the array or table.
Take a look at the Visual Foxpro Access and Assign Methods. These methods can be used to execute code when querying a property or trying to change the value of a property. Below is a link that shows an example:
Access and Assign Example
You could do something like this:
a=4
b='a'
a=6
?&b