Erratic hotspot performance -- sometimes the same code is 60% slower

I'm one of the developers on Apache Lucene, trying to run performance
tests of Lucene's new near real-time search feature...
But, I'm getting erratic behaviour out of the hotspot compiler,
whereby the compiled code can run up to ~60% slower depending on what
code has run before it, which really makes my testing hard!
Somehow, hotspot seems to get tricked into compiling the code very
badly, and then never recompiles it.
I'm running with JDK 1.6.0_17-b04, on a CentOS 5.4 (Linux) box. I run
java like this:
java -server -Xms1g -Xmx1g
I've tried various advanced -XX options to try to workaround this,
to no avail.
After whittling down my test, I managed to get a very simple
standalone test (depends only on the Lucene 3.0 JAR) that shows the
problem. The test first runs unrelated "warmup" code, then runs a
fixed search test multiple times, printing the fastest run.
There are 4 options for "warmup", and when I run the test with each of
these 4, it runs as fast as 726 msec and as slow as 1160. The numbers
are very stable (low noise) so I'm pretty sure this is accurately
measuring what hotspot had compiled.
I turned on -XX:+PrintOptoAssembly (downloaded the fastdebug JDK), and
it's really weird -- even very low level methods like readVInt (reads
a variable-length encoded integer, a serious hot spot in Lucene) was
compiled differently, depending on what code ran during "warmup".
Any ideas? For my testing I really just want consistency, but, going
forward I also would like to somehow "bias" hotspot to not produce the
60% slower compiled code...

Thanks!
The "compilation planning", and whole discussion/referenced papers
from there, are very interesting. It sounds like I'm indeed running
up against the complexities in the implicit "performance model" from
Java.
I just wish I had a simple way to "manage", or even understand
the differences. I'll try to study the output of
-XX:+PrintCompilation to see if I can gain any insight.... but this
situation pretty much makes it impossible for me to do the performance
testing I had wanted to do (the 60% perf hit due to different hotspot
decisions completely obliterates what I'm trying to measure!).

Similar Messages

In outlook 2013 Add-In, Adding dynamic menu to splitButton idMso="DialMenu" is working and the same code is not working in outlook 2010 Add-In.

In outlook 2013 Add-In, Adding dynamic menu to <splitButton idMso="DialMenu"> is working and the same code is not working in outlook
2010 Add-In. please let me know, if i am missing something. Below is the xml and screen shot
<contextMenu idMso="ContextMenuFlaggedContactItem">
<splitButton idMso="DialMenu">
<menu>
<dynamicMenu id="CallContactwithFreedomvoice
" label="CallContactwithFreedomvoice"
getContent="OnGetContenttest" insertAfterMso="Call"/>
</menu> </splitButton> </contextMenu>

Hi Narasimha prasad2,
Based on the description, the context menu for the flagged contact doen't work in Outlook. I am tring to rerpoduce this issue however failed.
I suggest that you check the state of the add-in first to see wether the add-in was loaded successfully.
Regards & Fei
We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
Click
HERE to participate the survey.

The same code behaving different in two files.

The same code behaving different in two files.
in pro*c file it is returning rows
in sql file it is not returning any rows.
please suggest me.
regards,
prasad.

please find the code.
sql_stmt := 'SELECT rNum, dpcCode, sID, hDate, fBlock, '||
'lBlock, recQty, fID, fName, ' ||
'tSeqNo, sType, tDate, fStatCode, pDate, recCount, r ' ||
'FROM (SELECT rownum rNum, DPC_CODE_ORIG dpcCode, ' ||
'SENSOR_ID sID, ' ||
'to_char(HEADER_DATE, ''YYYYMMDDHH24MISS'') hDate, ' ||
'FIRST_BLOCK_NUMBER fBlock, ' ||
'LAST_BLOCK_NUMBER lBlock, ' ||
'RECORD_QUANTITY recQty, ' ||
'FILE_ID fID, ' ||
'substr(FILE_NAME,1,30) fName, ' ||
'TRACE_SEQUENCE_NO tSeqNo, ' ||
'SENSOR_TYPE sType, ' ||
'to_char(TRAILER_DATE, ''YYYYMMDDHH24MISS'') tdate, ' ||
'NVL(FILE_STATUS_CODE, ''NL'') fStatCode, ' ||
'to_char(PROCESSED_DATE, ''YYYYMMDDHH24MISS'') pDate, '||
'NVL(RECIRCULATE_COUNT,0) recCount, ' ||
'ROW_NUMBER() ' ||
'OVER(ORDER BY TRAILER_DATE DESC, LAST_BLOCK_NUMBER DESC) r ' ||
'FROM logfc10t ' ||
'WHERE DPC_CODE_ORIG = :IN_DPC_CODE ' ||
'AND SENSOR_ID = :IN_SENSOR_ID ' ||
'AND (((TRAILER_DATE <= TO_DATE(:IN_TRAILER_DATE, ''YYYYMMDDHH24MISS''))' ||
' AND (LAST_BLOCK_NUMBER < :IN_BLOCK_NUMBER)) ' ||
' OR (TRAILER_DATE < TO_DATE(:IN_TRAILER_DATE, ''YYYYMMDDHH24MISS'')))) ' ||
'WHERE r = 1 ';

Performance for the below code

Can any one help me in improving the performance for the below code.
FORM RETRIEVE_DATA .
CLEAR WA_TERRINFO.
CLEAR WA_KNA1.
CLEAR WA_ADRC.
CLEAR SORT2.
*To retrieve the territory information from ZPSDSALREP
SELECT ZZTERRMG
       ZZSALESREP
       NAME1
       ZREP_PROFILE
       ZTEAM
     INTO TABLE GT_TERRINFO
     FROM ZPSDSALREP.
*Preparing Corporate ID from KNA1 & ADRC and storing it in SORT2 field
LOOP AT GT_TERRINFO INTO WA_TERRINFO.
SELECT SINGLE * FROM KNA1 INTO WA_KNA1
                  WHERE KUNNR = WA_TERRINFO-SALESREP.
SELECT SINGLE * FROM ADRC INTO WA_ADRC
                  WHERE ADDRNUMBER = WA_KNA1-ADRNR.
IF NOT WA_ADRC-SORT2 IS INITIAL.
CONCATENATE 'U' WA_ADRC-SORT2 INTO SORT2.
MOVE SORT2 TO WA_TERRINFO-SORT2.
MODIFY GT_TERRINFO1 FROM WA_TERRINFO.
APPEND WA_TERRINFO TO GT_TERRINFO1.
CLEAR WA_TERRINFO.
ENDIF.
CLEAR WA_KNA1.
CLEAR WA_ADRC.
ENDLOOP.
ENDFORM.                    " RETRIEVE_DATA

Hi
The code is easy so I don't think you can do nothing, only u can try to limit the reading of KNA1:
FORM RETRIEVE_DATA .
CLEAR WA_TERRINFO.
CLEAR WA_KNA1.
CLEAR WA_ADRC.
CLEAR SORT2.
*To retrieve the territory information from ZPSDSALREP
SELECT ZZTERRMG
ZZSALESREP
NAME1
ZREP_PROFILE
ZTEAM
INTO TABLE GT_TERRINFO
FROM ZPSDSALREP.
SORT GT_TERRINFO BY SALESREP.
*Preparing Corporate ID from KNA1 & ADRC and storing it in SORT2 field
LOOP AT GT_TERRINFO INTO WA_TERRINFO.
    IF KNA1-KUNNR <> WA_KNA1-KUNNR.
      SELECT SINGLE * FROM KNA1 INTO WA_KNA1
           WHERE KUNNR = WA_TERRINFO-SALESREP.
      IF SY-SUBRC <> 0.
        CLEAR: WA_KNA1, WA_ADRC.
      ELSE.
        SELECT SINGLE * FROM ADRC INTO WA_ADRC
                                 WHERE ADDRNUMBER = WA_KNA1-ADRNR.
        IF SY-SUBRC <> 0. WA_ADRC. ENDIF.
      ENDIF.
    ENDIF.
    IF NOT WA_ADRC-SORT2 IS INITIAL.
      CONCATENATE 'U' WA_ADRC-SORT2 INTO SORT2.
      MOVE SORT2 TO WA_TERRINFO-SORT2.
* MODIFY GT_TERRINFO1 FROM WA_TERRINFO.
      APPEND WA_TERRINFO TO GT_TERRINFO1.
      CLEAR WA_TERRINFO.
    ENDIF.
ENDLOOP.
ENDFORM. " RETRIEVE_DATA
If program takes many times to upload the data from ZPSDSALREP, you can try to split in sevaral packages:
SELECT ZZTERRMG ZZSALESREP NAME1 ZREP_PROFILE ZTEAM
INTO TABLE GT_TERRINFO PACKAGE SIZE <...>
FROM ZPSDSALREP.
SORT GT_TERRINFO BY SALESREP.
*Preparing Corporate ID from KNA1 & ADRC and storing it in SORT2 field
LOOP AT GT_TERRINFO INTO WA_TERRINFO.
    IF KNA1-KUNNR <> WA_KNA1-KUNNR.
      SELECT SINGLE * FROM KNA1 INTO WA_KNA1
           WHERE KUNNR = WA_TERRINFO-SALESREP.
      IF SY-SUBRC <> 0.
        CLEAR: WA_KNA1, WA_ADRC.
      ELSE.
        SELECT SINGLE * FROM ADRC INTO WA_ADRC
                                 WHERE ADDRNUMBER = WA_KNA1-ADRNR.
        IF SY-SUBRC <> 0. WA_ADRC. ENDIF.
      ENDIF.
    ENDIF.
    IF NOT WA_ADRC-SORT2 IS INITIAL.
      CONCATENATE 'U' WA_ADRC-SORT2 INTO SORT2.
      MOVE SORT2 TO WA_TERRINFO-SORT2.
* MODIFY GT_TERRINFO1 FROM WA_TERRINFO.
      APPEND WA_TERRINFO TO GT_TERRINFO1.
      CLEAR WA_TERRINFO.
    ENDIF.
ENDLOOP.
ENDSELECT.
Max

How to handle multiple exceptions by the same code?

Hi, all:
In my situation I want AException and BException handled by the same code, while CException and DException handled by another code. How can I write my try-catch code in a simple way? Of course I can do the following:
public void TheMainFunction() {
    try {
    } catch (AException e) {
        Handle_AB();
    } catch (BException e) {
        Handle_AB();
    } catch (CException e) {
        Handle_CD();
    } catch (DException e) {
        Handle_CD();
private void Handle_AB() {
private void Handle_CD() {
}But is there a simpler way?
Thanks in advance.

If you have one or two places in your code that need multiple exceptions, just do it with multiple catch statements. Unless you are trying to write the most compact Programming 101 homework program, inventing tricks to remove two lines of code is not good use of your time.
If you have multiple catches all over your code it could be a code smell. You may have too much stuff happening inside one try statement. It becomes hard to know what method call throws one of those exceptions, and you end up handling an exception from some else piece of code than what you intended. E.g. you mention NumberFormatException -- only process one user input inside that try/catch so it is easy to see what error message is given if that particular input is gunk. The next step of processing goes inside its own try/catch.
In my case, the ArrayIndexOutOfBoundsException and
NumberFormatException should be handled by the same way.Why?
I don't think I have ever seen an ArrayIndexOutOfBoundsException that didn't indicate a bug in the code. Instead of an AIOOBE perhaps there should be an if statement somewhere that prevents it, or the algorithm logic should prevent it automatically.

Use the same code module for several steps in the sequence

Hello All,
I have been trying to set up a sequence that uses the same code module for all of the steps in the sequence, but am having trouble referring back to it when I need to send it commands. I have gotten as far as calling the VI in a new thread so that it can be run asynchronously. I can run the simple sequence and it will indeed open the VI, and move on to the next step. When I close the VI manually from the front panel, the sequence in TestStand completes, as expected. So it appears that I have that much working.
My question is how to call the separate thread from the main sequence or other sub sequences when I need to edit the parameters. If I insert an Action step, I am required to select a VI file, but from what I can tell, it opens a different instance of the file, and does not provide an interface with the other instance running asycronously. My next guess was to use a Statement step, but I was not able to figure out how to configure the lookup string to call the VI parameters. Beyond that, I'm not sure how to proceed. Please advise.
My intention is to start the code module VI (asynchronously), run several different subsequences within the main sequence that call that same VI and edit it's parameters, close everything and report on the results. If I am misunderstanding how TestStand is supposed to work, please let me know.
Thanks,
GSinMN
Solved!
Go to Solution.

What I do is use a Queue to send data to the asynchronous VI. So it can run and do whatever, but also recieve the commands from the queue. I use an Action Engine that holds the queue reference and sends the commands. So you really just have to call the Action Engine from your sequences.
There are only two ways to tell somebody thanks: Kudos and Marked Solutions
Unofficial Forum Rules and Guidelines

Region Monitoring iOS 7 : didEnterRegion method is not calling when app is killed by user or by OS in iOS 7 only. It is working fine when it is in background. and the same code is working fine with iOS 6 for both app in suspended mode and background mode.

Region Monitoring iOS 7 : didEnterRegion method is not calling when app is killed by user or by OS in iOS 7 only. It is working fine when it is in background. and the same code is working fine with iOS 6 for both app in suspended mode and background mode. What changes I have to made to work great in iOS 7 also.

I rewrote code for debugging purpose and tried to catch error using GetLastError();  method,
but it only printed 0. Below is code snippet; I think Create() throw an exception
and code goes to catch block.
LONG ConnectTS(CString strIP, UINT n_Port)
try{
          ErrorLog(0,0,"ConnectTS is calling Create [is going to call]","");
          if(!Create())
// Exception Line
n_Err = GetLastError();
return NET_INIT;
catch(...)
                   DWORD errorCode = GetLastError();
                   CString errorMessage
                   errorMessage.Format("%lu",errorCode);
                   ErrorLog (0, 0, "Image
System", (LPTSTR)(LPCTSTR)errorMessage);
                   return  IS_ERR_WINDOWS;
Output: -
ConnectTS is calling Create [is going to call]
Image System
0

Drawing images gives 2 different results with the same code.

Im adding my code on here, and its sort of long but ill try to explain it. Most of it you can ignore.
Anyway i'm making a tree menu with only 1 level of expansion. So you have a TreeMenu object, and this object has some Branch classes. These Branch classes have Leaf classes. That is all. On the menu there is also a little box where you can click to expand and collapse each branch. if a branch is expanded, all of its leaves are shown. if it is collapsed only the branch name is shown.
So far what I've described above works, there is a small scale compilable working program, that is alright. Theres just a few bugs to fix.
A little feature of the tree is if a branch or a leaf is selected (has been clicked on in the past) it has a blue background. This sort of works. The branch names seem to work alright, and some of the leaves. However other if you run the example i provided you can click on Leaf 1, and it will be ok, but if you click on Leaf 2 then there blue background is twice as big as it should be. But for the rest, it works ok.
This is weird to me because the same code handles any of these same situations. So it seems that i get 2 diferent results with the same code. If anyone would like to help heres some stuff you should know:
You can ignore Leaf class, i dont think theres any errors in there.
You can ignore just about all of Branch class except the update() method.
You can ignore all of TreeMenu class except the inner MouseListener class might be useful.
You can ignore TreeTester. Just the JFrame tester class.
to run the code:
javac TreeMenu.java
java TreeMenu
http://cs.ucsb.edu/~jsterling/treemenu.zip
thanks.

You make a good point Stefan but unfortunately I already checked that. I have even copied the mapping (right click copy) and pasted it from the working Mapping to the bad Map (Paste) to make sure the context values were all the same.
I also opened all the queues and compared the values in each. All values match exactly except for the result(out) queue.
In both tests I am using the same IDOC XML file.
These are the only differences I see.
1. The IDOC is exactly the same as far as I can tell but they come from two different source SWCV. SAP APPL 4.7 and Steelcase_Procurement (ECC 6.0).
2. The working map uses a local UDF. The bad map calls the UDF from a Function Group. Is there a Java version difference between the two?
A better sample of the result queues look like this:
Correct Result:
Thickness of Glass: 20
Height: 10
Width: 5
Length: 18
Thickness of Glass: 10
Height: 15
Width: 8
Length: 14
Bad Result:
Thickness of Glass: 20
<null>
<null>
<null>
Width: 5
<null>
<null>
<null>
I took this chance to redesign my Variant Config output to separate the characteristics/values, but I'm still bothered that I was unable to determine what would cause the differing values.
Any ideas?
Thanks,
Matt

Why differences in the way different machines handle the same code??

OK, I've posted some other threads re some difficulties in getting the same code to function identically on, not only different platforms, but even on different machines running Windows!
Here's another, and I would really appreciate an explanation:
I have a tree control. Per the user's request, I made the right button function the same way it does on MS File Explorer: When they right click, the currently selected nodes are compared to the node they clicked over. If they clicked over a selected node, then the context menu for the nodes selected is displayed. If they clicked over a different node, then the selection is changed to the new node and the context menu displayed. If they right click over the tree control, but not a node in the tree, then the selection is cleared and no context menu is displayed.
Works great on my machine.
However, on some machines, no context menu is displayed, nor is the node selected... UNLESS THEY ALREADY HAVE A NODE SELECTED. I.e., if they have no nodes in the tree selected, they cannot display the context menu by right clicking on a node, nor does the selection change to the node they right click over. It only works if they already have something selected.
My app is distributed as an executable jar file and I am distributing the JRE with it, so there are no class path issues or differences in the JDK. Basically, my problem can be paraphrased as: Why doesnt' Swing behave the same way on different machines...? At least why not the same on the eame machine.
It's driving me nuts to think that I've resolved some UI glitch only to find out that, no, it's only fixed on some machines, not others.
Thanks.

Well, I normally develop on OSX and then deploy to Windows. But, this last build I have switched to Windows XP. The kinds of differences I am seeing are subtle: what triggers a repaint of the UI and now, it seems, what events are getting triggered by mouse clicks.
My deployment is an executable jar file, therefore it uses the manifest in the jar, not any classpath variables. I deploy an entire JRE along with the app, and the bat file that launches the app ensures that is the JAVA_HOME I'm using.
I can only guess that it's some underlying difference in versions of the platform and OS that causes this.
Here's the entire code for this latest issue re the popups:
        private void maybeShowPopup(NodeInfoTree tree, MouseEvent e) {
            if (e.isPopupTrigger()) {
                // They clicked the platform dependent popup trigger...
                int x = e.getX();
                int y = e.getY();
                TreePath[] selectedPaths = tree.getSelectionPaths();
                if (selectedPaths == null) {
                    selectedPaths = new TreePath[0];
                TreePath clickedPath = tree.getPathForLocation(x,y);
                if (clickedPath == null) {
                    // they right clicked, but not on a node, so clear the selection and return
                    tree.clearSelection();
                    return;
                } else {
                    // They clicked on a path, so let's see if it is a selected
                    // node. Search the selected array of paths, short circuiting
                    // out if it is found...
                    boolean foundInSelection = false;
                    for (int i=0; (!foundInSelection && (i < selectedPaths.length)); i++) {
                        foundInSelection = clickedPath.equals(selectedPaths);
if (!foundInSelection) {
// They right clicked elsewhere than the selection, so
// reset the selection to the new path...
tree.setSelectionPath(clickedPath);
final TreePath menuPath;
selectedPaths = tree.getSelectionPaths();
if (selectedPaths == null) {
return;
} else {
menuPath = selectedPaths[selectedPaths.length-1];
// Since they did click on a node, now we need to determine
// what popup menu should be displayed...
tree.scrollPathToVisible(menuPath);
NodeInfo ni = (NodeInfo) menuPath.getLastPathComponent();
// If the last path component is null, then they don't have
// anything selected, so we do nothing
if (ni == null) return;
// See if the Node is overriding our default popup menu...
com.harcourt.applications.tgen.browser.nodes.NodeMenu nodeMenu = ni.getMenu();
nodeMenu.activate(tree.getView(),ni);
JPopupMenu popup = (JPopupMenu) nodeMenu;
if (popup != null) {
popup.show(e.getComponent(),
e.getX(), e.getY());
The nodeMenu.activate call is just used to check whether or not certain menu options need to be enabled or disabled before displaying the menu. My theory right now is that, on some of these machines (all windows), a right click of the mouse is not generating a mouse event unless they have already something selected. I don't think it's a focus issue, because if they left click to select a node, and then ctrl-left click to de-select it, the right button then no longer generates a popup menu. They must select a node, then right click. Maybe there's no mouse event generated, or maybe this line is failing on some machines?:
TreePath clickedPath = tree.getPathForLocation(x,y);

I dw file.dwt files may insert server side code supposed all pages depended are in the same code eg

I dw file.dwt (exact like this ext) files may insert server side code supposed all pages depended are in the same code eg cfml?

I'm sorry, but your English is pretty difficult for me to figure. I'm assuming your asking how to have a Dreameraver template support ColdFusion. It hat is what you are asking, the template file extension should be: YourTemplateName.dwt.cfm - this tells Dreamweaver is a ColdFusion template.
I hope I understood you and that this helps.
Lawrence Cramer - *Adobe Community Professional*
http://www.Cartweaver.com
Shopping Cart for Adobe Dreamweaver
available in PHP, ColdFusion, and ASP
Stay updated - http://blog.cartweaver.com

OIM11 event handlers: How to avoid firing the same code in both pre+post

Hi everyone,
I have a question around event handlers. My experiments so far have gleaned the following:
Manual updates of user form in admin interface -> Fires: pre-insert + Post-insert
Reconciliation trusted creates -> Fires: Post-insert
Reconciliation trusted updates -> Fires: pre-insert + Post-insert
Now I have various handlers that trigger to update fields coming in from trusted recon. Formatting telephone numbers, setting custom user attributes.. etc. I am using the same code in both the post and pre handlers as wherever the update comes from it needs to be processed in the same way. Problem is if a handler fires twice I can't be sure exactly how the system is going to behave (updating the same field again etc), never mind it is unneeded processing.
We have to keep the pre handlers because otherwise changes completed via the admin interface won't be seen until you refresh.
Can anyone please advice how to go about ensuring a handler is only fired once? i.e. if pre fires don't fire post. Have I missed something key here?
Edit: I know I've worded this badly. They will always fire as that's how OIM behaves, what I want is some way to work out in the underlying code if a field has already been modified in pre process... or something like it.
Thanks,
Wayne.
Edited by: wblacklock on 27-Feb-2012 05:38

I am sorry but I am not agree with your design.
However there is no chance to avoid this according to my knowledge. As both class has separate thread under different process, so there is no way to implement thread lock on entity operation.
Alternate way to acheive your requirement:
You can have hidden UDFs. Update hidden UDFs in prepost handler with some data (hardcoded).
"You can have the value this hidden filed like :- "NAME_UPDATE|EMAIL_UPDATED|MANGER_UPDATE"
Now in Post - Process handler check the value of hidden UDF.
Get the value of hidden UDF, tokenize with String Tokenizer. Check which filed is not updated in Pre handler.
*If it is already updated - do not update.* else Update.
Thanks,
Kuldeep

Running the same code multiple times with different paramters automatically

Hi guys,
I want to run the same code multiple times with different paramters automatically. Actually, I am doing some sample scans which takes around 2 hours n then I have to be there to change the paramters and run it again. Mostly I do use folowing paramters only
X_Intial, X_Final, X-StepSize
Y_Intial, Y_Final, Y-StepSize
Thanks,
Dushyant

All you have to di is put all of the parameters for each run into a cluster array. Surround your main program with a for loop and wire the cluster array through the for loop. A for loop will autoindex an input array so inside the for loop you just have to unbundle the cluster to get the parameters for each run.
Message Edited by Dennis Knutson on 07-13-2006 07:50 AM
Attachments:
Cluster Array.JPG ‏9 KB

When dragging a thumbnail from one place to another, the origional thumbmail ends up in other locations. Sometimes the same thumbnail covers up other thimbmails.

When dragging a thumbnail from one place to another, the origional thumbmail ends up in other locations. Sometimes the same thumbnail covers up other thimbmails.

Hi christophers37215878
Please elaborate your issue.
What exactly are you trying to achieve?
Which Adobe Product or service are you referring to?
Can you please provide the screenshot?
Regards,
Rahul

How to make the same code that works only?

how to make the same code that works only?
code:
Form2.Close = Me.Enabled = true
effect such as this code

I have this code:
Public Class Form1
Private Sub Button4_Click(ByVal sender As System.Object, ByVal e As System.EventArgs) Handles Button4.Click
Form2.Show()
Me.Enabled = False
End Sub
Private Sub Button5_Click(ByVal sender As System.Object, ByVal e As System.EventArgs) Handles Button5.Click
Form3.Show()
Me.Enabled = False
End Sub
End Class
he wants to after closing Form2 and Form3 unblocked the whole form1
it means that:
when I click the button, displays the himself form2 and blocks himself form1" and after closing Form2 , unlocks himself Form1"
the same with Form3

What happens when two computers call the same code stored in network storage?

We have our code on network storage and have only been running code on one computer. We have now duplicated our setup and would like to use the same code to run the new setup from a different computer. We would like to run both setups simultaneously (both using same code from network storage). Will there be conflicts here or do both computers load their own copy of the vi to run from? I should say that we have no global variables.
I'm interested in the answer to this question anyway but I would also be happy to hear any suggestions (or even criticisms) on how to do what we are doing.
Solved!
Go to Solution.

You should seriously consider a source control system, which is the best solution here. That gives you one master copy of the code on the network. Each computer retrieves the latest revision from source control, making a copy on the local disk. Source control helps maintain a known good backup, and allows you to track changes and compare versions or revert to older code when something breaks.
I'm fairly certain you can open the same VIs stored on a network drive on multiple computers concurrently. Global variables are irrelevant here. Each computer will independently load the VIs into memory. The problem is if you edit and save the VI on one computer while it's open on another, and if you make edits on both simultaneously, you'll definitely lose some of the edits and could possibly corrupt your VI and lose all of them, which is why you should be using a source control system instead.
If you have the application builder, I recommend building your code into an application. You can store the application the network if you want, and run it on multiple computers.

Erratic hotspot performance -- sometimes the same code is 60% slower

Similar Messages

Maybe you are looking for