Hi all
We are experiencing better execution 2nd time we run a program in the day, especially with large amount of records. The program read through a big file much faster second time. Has anybody else experienced anything like this
Of course, this can happen by using the computer's cache. But if the computer is not closed, how can the same happen the next day. First execution takes a lot longer time that the second. Is there a way to control this cache? First of all make it as big as possible. Second, how do I avoid the cache is deleted during the night. I have been told that Acu only makes cache for programs when using acuserve. Is it possible to see what is stored in the cache?
The programs is executed with Wrun32 directly on the server, version 7.2 but it is also seen in 9.2.1
Best Regards
steen
Is it possible that your file start statement is causing this. Perhaps when the application is run ,a file start sets the file in a certain order. When the application is run the 2nd time, that file order does takes as long as it did when running the app the first time.
Hi all
We are experiencing better execution 2nd time we run a program in the day, especially with large amount of records. The program read through a big file much faster second time. Has anybody else experienced anything like this
Of course, this can happen by using the computer's cache. But if the computer is not closed, how can the same happen the next day. First execution takes a lot longer time that the second. Is there a way to control this cache? First of all make it as big as possible. Second, how do I avoid the cache is deleted during the night. I have been told that Acu only makes cache for programs when using acuserve. Is it possible to see what is stored in the cache?
The programs is executed with Wrun32 directly on the server, version 7.2 but it is also seen in 9.2.1
Best Regards
steen
I've seen the same caching behaviour with linux. This is operating system behavior and nothing to do with the extend runtime. Having said that, some thoughts:
* depending on what your program does, thin client should help. It's more efficient than peer-to-peer file IO (even faster than acuserver in my opinion)
* The manual mentions "V-BUFFERS" configuration variable that default to 64, but can go up to 2 million. (512 bytes and a bit per buffer). This can help, it's unclear how much.
* You could write a little utility program that just reads the same files and same records as your main program, and schedule it to run in the early morning before anyone starts using your main program. This would reload/refresh the server cache memory. It's a cheat, but it would probably help.
Hi all
We are experiencing better execution 2nd time we run a program in the day, especially with large amount of records. The program read through a big file much faster second time. Has anybody else experienced anything like this
Of course, this can happen by using the computer's cache. But if the computer is not closed, how can the same happen the next day. First execution takes a lot longer time that the second. Is there a way to control this cache? First of all make it as big as possible. Second, how do I avoid the cache is deleted during the night. I have been told that Acu only makes cache for programs when using acuserve. Is it possible to see what is stored in the cache?
The programs is executed with Wrun32 directly on the server, version 7.2 but it is also seen in 9.2.1
Best Regards
steen
@shjerpe - sort of off-topic but related...
Does a start statement actually re-order/sort a file? I was under the impression that it modified the current file position and set the current key being used. I have noticed the behavior indicated in OP's question as well.
