NetTalk Central

Author Topic: IMDD and DFD  (Read 2723 times)

kevin plummer

  • Hero Member
  • *****
  • Posts: 1195
    • View Profile
    • Production Accounting and Software Payroll
IMDD and DFD
« on: January 18, 2017, 05:20:14 PM »
Hi All,

just wondering if it is possible to use the combination of these file drivers (dyn.cachefile() ) to cache some data on a thread to make my app go a bit faster? I would just cache the tables where I'm just doing reads (not writes) in the thread. I'm using MSSQL so hoping this will also ease the traffic/workload on that server. Any examples of doing this on forms, browses, source code etc would be great.

Cheers,

Kevin


Bruce

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 11191
    • View Profile
Re: IMDD and DFD
« Reply #1 on: January 19, 2017, 02:25:54 AM »
Given that a thread only lasts for a very short time, presumably you'd need to be doing a lot of random reads _on that thread_ to make this useful? Do you have instances where that is the case?

Cheers
Bruce

kevin plummer

  • Hero Member
  • *****
  • Posts: 1195
    • View Profile
    • Production Accounting and Software Payroll
Re: IMDD and DFD
« Reply #2 on: January 19, 2017, 03:28:55 AM »
On one app I'm looking to speed up reports and some of my source procedures that do a lot of calculations for tax, fringes and hours to gross. I still get a couple of SQL locks a day on this app so hoping caching some data will take pressure off MSSQL

On another app on an older server its just running slow. 60% requests > .5 The same app on a new highly specked server in Europe runs fine with 86% requests < .5

I'm not in the position to upgrade the slower server yet so just trying to work out how to make the app go faster.

Bruce

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 11191
    • View Profile
Re: IMDD and DFD
« Reply #3 on: January 19, 2017, 10:50:54 PM »
>> On one app I'm looking to speed up reports and some of my source procedures that do a lot of calculations for tax, fringes and hours to gross.

yes, lots of random reads on a single table can definitely be sped up by caching some data.
Also remember to make user of the BUFFER statement when manually reading a set of records.

cheers
Bruce