| 1 |
C $Header$ |
C $Header$ |
| 2 |
C $Name$ |
C $Name$ |
| 3 |
C |
CBOP |
| 4 |
C /==========================================================\ |
C !ROUTINE: GLOBAL_MAX.h |
| 5 |
C | GLOBAL_MAX.h | |
C !INTERFACE: |
| 6 |
C | o Globals used by Fortran global max routine. | |
C include "GLOBAL_MAX.h" |
| 7 |
C |==========================================================| |
C !DESCRIPTION: |
| 8 |
C | The global max shared memory scheme uses global heap data| |
C *==========================================================* |
| 9 |
C | structures (.i.e COMMON blocks ). Each thread writes to | |
C | GLOBAL_MAX.h |
| 10 |
C | an its own element of the shared memory array and then | |
C | o Globals used by Fortran global max routine. |
| 11 |
C | one thread reads all the entries and maxs them. The max | |
C *==========================================================* |
| 12 |
C | result is then read by all threads. | |
C | The global max shared memory scheme uses global heap data |
| 13 |
C | Remember - you are working with regions of memory that | |
C | structures (.i.e COMMON blocks ). Each thread writes to |
| 14 |
C | are being updated concurrently by different threads. | |
C | an its own element of the shared memory array and then |
| 15 |
C | What happens, when it happens and who gets to see what | |
C | one thread reads all the entries and maxs them. The max |
| 16 |
C | happens at what stage depends on the computer systems | |
C | result is then read by all threads. |
| 17 |
C | memory model. Every computer has a different memory model| |
C | Remember - you are working with regions of memory that |
| 18 |
C | and they are never simple. In all current platforms it is| |
C | are being updated concurrently by different threads. |
| 19 |
C | possible for one thread to see events happening in a | |
C | What happens, when it happens and who gets to see what |
| 20 |
C | different order from the order they are written in the | |
C | happens at what stage depends on the computer systems |
| 21 |
C | code. | |
C | memory model. Every computer has a different memory model |
| 22 |
C | Unless you understand this it is not a good idea to | |
C | and they are never simple. In all current platforms it is |
| 23 |
C | make modifications te way these header files are setup or| |
C | possible for one thread to see events happening in a |
| 24 |
C | the way the global sum routines work. | |
C | different order from the order they are written in the |
| 25 |
C \==========================================================/ |
C | code. |
| 26 |
|
C | Unless you understand this it is not a good idea to |
| 27 |
|
C | make modifications te way these header files are setup or |
| 28 |
|
C | the way the global sum routines work. |
| 29 |
|
C *==========================================================* |
| 30 |
|
CEOP |
| 31 |
COMMON / GMAX_COMMON_R8 / phiGMRL |
COMMON / GMAX_COMMON_R8 / phiGMRL |
| 32 |
Real*8 phiGMRL(lShare8, MAX_NO_THREADS ) |
Real*8 phiGMRL(lShare8, MAX_NO_THREADS ) |
| 33 |
|
|
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch