A Faster/Better way?

Hi All,

    I have an unformatted file, and I open it for reading in big
endian mode. The file has a repetitive structure (is a time step
re****t for a simulation): at the beginning of every time step, there
is this pseudo-declaration:

'PARAMS'
keyword number kind

where "keyword" is an 8 character string, "number" is an integer which
tells me how much data this time step contains, and "kind" is a 4
character string which tells me the data type (i.e., 'CHAR', 'REAL'
and so on). After this declaration, the data for the current time step
follows. As an example:

'PARAMS  '  4200 'REAL'
0.5372 3.7393 0.3273 .... (4200 real numbers)

With the file I am working on, this 4200 stays constant for every time
step in the simulation.
The numerical arrays are divided into blocks of up to 1000 items each:
so, in my previous example, I suppose I can not read the whole 4200
real numbers in one shot, I have to loop 4 times and read 1000 items
every time, plus read the remaining 200 as last step.

I have written the following routine to read and store the data from
this file. Is there anything I could do to speed up the reading/
storing of the data?

Thank you very much for your suggestions and comments. I apologize for
the long post.

Andrea.

! Begin Code

subroutine ReadSMSPEC(fileName, miniSteps, dimens, matrix)

	! PARAMETERS:
	! fileName: the file containing the data
	! miniSteps: the number of time steps in the simulation
	! dimens: the number of data in each time step
	! matrix: a dimens x miniStep 2D array of real data

	character*1000, intent(in) :: fileName
	integer, intent(in) :: miniSteps, dimens
	real(4), intent(out) :: matrix(dimens, miniSteps)

	integer keywordNumber, internalCount, readBuffer, matrixIndex

	character*8 keywordName
	character*4 keyworType

	logical feof

	! Open the file
	open(unit=1, file=fileName, form='UNFORMATTED', convert='BIG_ENDIAN')

	feof = .false.

	! The numerical arrays are divided into blocks of up to 1000 items
each
	readBuffer = int(dimens/1000)

	do while (.not.feof)

		! Read the keyword, number and kind
	    read(1, end=40) keywordName, keywordNumber, keywordType

		! I look for the 'PARAMS' keyword
	    if (keywordName == 'PARAMS') then

			! A new time step has been found
			! Increase the index in the matrix
			matrixIndex = matrixIndex + 1

		    internalCount = 0

			! Loop by reading a block of 1000 items
			! at a time

			do while (internalCount <= readBuffer)

				! Read a block of 1000 items until
				! you reach the end of this time step

				if (internalCount < readBuffer) then

					! We are still far from the last block,
					! Read 1000 items of data

		            read(1, end=40) matrix(1000*internalCount
+1:1000*(internalCount+1), matrixIndex)

		        else

					! We reached the last block of data:
					! Read the remaining data, which can be
					! less than 1000 items.

					read(1, end=40) matrix(1000*internalCount+1:dimens, matrixIndex)

				endif

				internalCount = internalCount + 1

			enddo

		endif

	enddo

40	continue

	close(1)
	return

end subroutine readSMSPEC

! End Code