Process "parallel."

Similar Questions

• in parallel to the max, process, process servers parallel slave: relationship

  Hello
  
  Need help to understand the relationship of what follows.
  
  I have a database with the following in the init ora.
  
  test > view parallel_max_servers setting
  
  Value of parameter TYPE
  ------------------------------ ----------- -----------------------------------
  PARALLEL_MAX_SERVERS integer 400
  
  
  test > view the process of setting
  
  Value of parameter TYPE
  ------------------------------ ----------- -----------------------------------
  whole process 500
  
  Select * from v$ resource_limit
  2 where resource_name = '' process. ''
  
  RESOURCE_NAME CURRENT_UTILIZATION MAX_UTILIZATION
  ------------------------------ ------------------- ---------------
  INITIAL_ALLOCATION LIMIT_VALUE
  ------------------------------ ------------------------------
  process 452 500
  500 500
  
  
  
  
  Question:
  
  Linux machine, I see the PX server with ps - ef | grep 'database_name ' | WC-l-> value is to achieve a greater it 400, but less than 500. My understanding is that parallel_max_servers specifies the maximum number of processes parallel slave that an instance is allowed to have both. Why is that process parallel slave of the ps - ef | grep 'database_name ' | WC-l reached above 400. In addition, this has been verified in v$ resource_limit. The value of the column process reached above 400.
  
  
  Thank you.
  
  Published by: user1006412 on February 24, 2011 22:09

  I said "processes_ Server _all (Oracle), including both background processes.

  There are 3 types of processes:
  1 background (DWBR, LGWR, PMON etc.)
  2 (dedicated or shared server) process for the user sessions
  3 Parallels operators (also called process a ParallelQuery or ParallelSlave)

  You have 26 other methods for user sessions.

  Hemant K Collette

• Process synchronization strategy

  Trying not to reinvent the wheel...

  I started to rethink an older data acquisition/processing application using as much parallelism as possible.

  UI/data acquisition / Data Saving/computer/display of data are all handled in separate loops or screw that communicate and Exchange data via queues or the authors of the notifications. It works great so far.

  Then, I realized that the way I have provided the data processed by the data processing (DP) for tasks of the display (D) wasn't going to work in all circumstances.

  I use a notification utility, as well as several tasks D can access data and such that there is not much cost memory (no queue) if tasks D cannot hold. If the data are generated faster D tasks can get it, they miss some but get a chance to catch up more late or at least get some data from time to time. That's what I had in mind of origin.

  Then I realized that I I want to display all the data (for example in offline mode). This architecture will not work. If I had to keep the approach of notifying data, I would need a way to tell the task of DP to send data to the system to alert only when all D tasks are done with their previous data segment.

  What is the best way to achieve this? I need to be able to dynamically change the number of clients (task D), so I thought Rendezvous could work. But then looking at the example "Wait for all Notifications Demo.vi", it seemed that this might be an approach more natural but a bit tedious to handle since notifiers would be created (and destroyed() dynamically saw separate...

  The fact is I want DP about how many customers it is supposed to wait (at any time) to send a new notification (with the new data).

  Thanks for your comments.

  X.

  In any case, since LV 9 SP1 crashed and threw all my work today (what about auto save every 5 minutes? Is it not working the 15th of each month or something?), I'll take a break and to describe the solution I came with:

  -J' created a process parallel to the VI of DP, what I call "Consumer Response Monitor" (CRM). This process asks a queue (waiting for messages to arrive using Dequeue, so no waste of resources) which is populated by the DP and D

  -DP warns CRM that new data has been sent (and thus to refocus its attention on the fate of the new data.)

  D the first register with CRM (and unregister when they left) and then tell CRM when they received data (who and who they are).

  -CRM monitors the number of registrants D receiving the new data and tells DP when all have (or there is a timeout). She did this by using a different queue.

  -If I am in 'direct mode' and don't care not if the D to time to process all the data, I just ignore the traffic with CRM.

• Control of the event: how to stop the race while loop within a structure of the event

  Hello

  I have some problems with control of a while loop inside a structure of the event (see annex VI).

  I have 3 buttons (the "Start measurement, stop, exit the program"). When a measure is running, it should be possible to stop the measure by clicking on "Stop the measure", but it does not work.

  Someone at - it an idea?

  Thank you very much and best regards,

  Michael

  Hi Michael,

  create a process parallel to your measurement. You can use the design model "producer/consumer with vents. You'll find it with «File-> new...» ».

  Another possibility would be to remove the "lock frontpanel" sign, in the case of event start measurement.

  Mike

• DBMS_STATS. SET_TABLE_PREFS does not work as expected

  Hello!
  
  I have a rather strange behavior of the DBMS_STATS. SET__PREFS proceduree.
  
  I want to realize that statistics on the table are collected with degrre = 4 in parallel, and I want to achieve this by defining the table preferences and not explicitly mention the degree in the call to the collection of statistics on the table.
  
  That's what I did:
  
  -first of all check the global preferences:
  
  Select dbms_stats.get_prefs ('DEGREE', 'GLOBAL') of double;
  ======================================
  NULL VALUE
  
  Then get the table preferences:
  
  Select
  DBMS_STATS.get_prefs ("DEGREE", "DARL', 'EIGNUNGSEINZELERGEBNISSE'")
  Double;
  ================================================
  NULL VALUE
  
  Then I try to define the DEGREE of 4:
  
  Start
  DBMS_STATS.set_table_prefs ('DARL', 'EIGNUNGSEINZELERGEBNISSE', 'LEVEL', '4');
  end;
  
  and after that I question the preferences:
  
  Select
  DBMS_STATS.get_prefs ("DEGREE", "DARL', 'EIGNUNGSEINZELERGEBNISSE'")
  Double;
  ====================================================
  4
  
  But when I run
  
  Start
  DBMS_STATS.gather_table_stats ('DARL', 'EIGNUNGSEINZELERGEBNISSE');
  end;
  
  Statistics is not picked up at the same time...
  
  I just need to explicitly mention degree - and it runs in parallel:
  
  Start
  DBMS_STATS.gather_table_stats ('DARL', 'EIGNUNGSEINZELERGEBNISSE' = > '4');
  end;
  
  
  Now - I checked just the prosesses each order product... in the last example you could clearly see the process parallels 4 th.
  
  To be honest, I didn't ' t now in what display catalog could I look up if the statistics resulting, collecting the operation takes place in parallel or not.
  
  
  And the table is partitioned and 5.4 G... It takes time to collect statistics, with all the other settings as they are.
  
  What is happening here... I forgot something. Is there another setting that should be set for the preferences to function properly?
  
  Database version is 11.2
  
  Published by: Robert on November 26, 2012 14:25
  
  Published by: Robert on November 27, 2012 08:18

  I just checked and got the same result in the 11.2 database.

  At first, I noticed the following in the description of the procedure:

  ... DBMS_STATS can use serial execution if the size of the object does not guarantee parallel execution.

  I checked on the tables of 4G and 250G with the same result.

  The second thing I noticed was how default degree is calculated in the procedure:

  ...
    degree           NUMBER   DEFAULT to_degree_type(get_param('DEGREE')),
  ...
  

  Please note the use of DBMS_STATS currently obsolete. GET_PARAM procedure which is replaced by GET_PREFS. The first steps pick up preferences defined at the table level. If you set the global preferences using DBMS_STATS. SET_GLOBAL_PREFS it would be used as the default value (I checked).

  Cause depending on your situation you may find this unacceptable solution. I have this case you always explicitly specify the degree as you did before.

  Looks like a bug to me, but cause I have no confirmation of that yet.

  It will be useful.

• What broadcast and Round Robin at the same time of the modes of distribution?

  Hi all
  
  as you know it some methods so that they were distributing data between consumer and producer of games. for ex: partition, hash, range...
  
  If you use one order by statement at the same time a statement that you will probably see the range method because the sets of consumption creates a range of order the lines and every time an execution in parallel to producer (parallel) server wants to send a row, he look for parallel processes of consumer who has an appropriate range of selection.
  
  but on the methods of dissemination and Round robin, I couldn't understand what they do and how to work Hey... I searched the documentation, but I couldn't find a document that explains their good. for ex if I write a select statement like this
  

  select /*+ parallel(sales, 4) */ * from sh.sales
  union
  select * from sh.sales2

  then I can see alternating in the execution plan.
  
  So what's "Broadcast" and what do do "Round robin".
  
  Thanks a lot for the answers.
  
  Edited by: elcaro on Sep 25, 2011 11:00

  in the method of Robin, a set of parallel processes to manage your data and process series send all data (she) up to the parallel treats each as a game of poker. in your example, manages the sales data in the table by the (process) parallel execution servers and sales2 table is managed by a process series (one).

  disseminate the method, a line sends to all the processes of consumption.

• several errors during the bulk add (ORA-13199, ORA-12801, ORA-01652-ORA-01653)

  Hello
  
  In a batch process, I use a big parallel add to insert millions of triples in an RDF model.
  At one point, an insert failed with this message:
  
  ORA-13199: DDL_same_rdf_indexes call failed: ORA-13199: step_num = 6 = ORA-12801 SQLERRM: error reported in the parallel query P020 Server
  ORA-01652: unable to extend segment temp of 128 in tablespace RDF dss = ALTER INDEX REBUILD PARTITION of 'RDF_LNK_SCPM_IDX' MODEL_63 NOLOGGING PARALLEL in LINE ORA-06512: at the 'MDSYS. MD", line 1723
  ORA-06512: at the 'MDSYS. MDERR", line 17
  ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL', line 9223
  ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL', line 9428
  ORA-06512: at the 'MDSYS. MD", line 1723
  ORA-06512: at the 'MDSYS. MDERR", line 17
  ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL', line 9441
  ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL', line 4255
  ORA-06512: at the 'MDSYS. SDO_RDF', line 276
  ORA-06512: at the 'MDSYS. RDF_APIS', line 693
  ORA-06512: at line 1
  
  Given that the batch continued and tried to insert later, I got this error twice, then another several times:
  
  ORA-13199: during the LBV:ORA - 01653: unable to extend table MDSYS. RDF$ TV_3F_1501FA303B1C4_56 of 128 in tablespace RDF ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL', line 1319
  ORA-06512: at the 'MDSYS. MD", line 1723
  ORA-06512: at the 'MDSYS. MDERR", line 17
  ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL", line 1414
  ORA-06512: at the 'MDSYS. SDO_RDF_INTERNAL', line 3786
  ORA-06512: at the 'MDSYS. SDO_RDF', line 276
  ORA-06512: at the 'MDSYS. RDF_APIS', line 693
  ORA-06512: at line 1
  
  Finally, later, inserting block worked without any errors. I don't have any sort of similar since errors.
  
  Any idea of what could have caused this?
  
  Thank you
  
  Kind regards
  Julien

  Indeed, it is a little uncertain. But here's a possibility (unless the tablepace has been extended before the retry):

  When several parallel processes are working on a query, each requires a space of sort. If N of them is active at a time and using their sorting max space requirements, the total space of necessary sorting can be beyond what is available in the temp tablespace. But, if we ever reach this point because some processes parallel either end early or have not reached the point where they need their sorting max spaces, then we can't hit it out of space error.

  Once again, an easy way to avoid this problem would be to use BIGFILE tablespaces or allocate enough space for relevant tablespaces.

• How to improve integration with to and not

  There is a previous process that fills data doc_t.id. Lines normally about 10,000. Subsequently the main process is running:
  

  insert into doc_t
  (SELECT DISTINCT no2, NULL, NULL, NULL 
  FROM document_t
  WHERE  no1 IN (SELECT id FROM doc_t) 
  AND    no2 NOT IN (SELECT id FROM doc_t) 
  AND    msg IN (33, 2) 
  AND    doc_time <= '31.12.2009')
  ; 

  Post-processing transfers the data of doc_t and finally truncates. doc_t Table document_t contains lines of 300Mio.
  
  All the work runs fine most of the time. But if him preprocess inserts 10000 s of lines in doc_t, the work goes on for hours.
  
  How can we improve the insertion / design? Some options: we think:
  -document_t of partition and run processes parallel (permit required)
  -transfer doc_t into a temporary table
  -calculate the statistics on the table for doc_t before insertion
  
  What other options should be considered? Is there a way to avoid these / not in constructions (= redesign)?

  This should be faster, check yourself.

  insert into doc_t
  (SELECT DISTINCT no2, NULL, NULL, NULL
  FROM document_t a inner join doc_t b on (a.no1 = b.id)
  WHERE  no2 NOT IN (SELECT id FROM doc_t)
  AND    msg IN (33, 2)
  AND    doc_time <= '31.12.2009')
  ;
  
  create index doct_id_index on doc_t(id);
  dbms_stats.gather_index_stats(user,'DOCT_ID_INDEX');
  

  It is important that you gather statistics during the doc_t table is in a typical/filled State.

  Published by: René Jeruschkat, January 15, 2010 08:20

  Published by: René Jeruschkat, January 15, 2010 08:24

• A single button to launch several parallel processes

  When I change my template to the parallel process, I see that each process has its own button 'OK '. I would like to have a button to start the process. How should I do this?

  Batch processing model can no longer make sense for your situation then.

  -Doug

• parallel to the process and return on investment?

  Hello

  I have two questions:

  1 how I can run two parallel processes in labview, which a single process is webcam that stream videos, and the other process is instant of the video image

  (I want a snapshot image after selecting a KING in the live camera view and keep the live camera view).

  2. How can I choose the KING will be part of the whole picture?

  Hello

  Below, I have attached a sample program that implements something similar.  It uses a structure of the event to listen to when trigger and records and image while performing a grab IMAQdx.  I'm not sure if you try to extract the KING, however, you should be able to change the structure of the event to perform processing on the image instead of save to file.  If you are looking to extract a return on investment, I recommend that you look at from Example.vi which comes with the Vision Development module.  "" "" You can find by going on Help "find examples... then navigate to Modules and Toolkits ' Vision ' functions ' Example.vi extracted.  It uses a node to invoke to get a last event to listen to when a KING is placed then passes the IMAQ Image extract.  For your application, you could implement something similar to my attachment.  Have one end grasping images then place the node to invoke to get a last event to display the images in a parallel loop to perform the extraction of KING and post-processing.  Probably, you wouldn't need a structure for this event.  Notice a 2nd IMAQ Create.VI is used in both examples to create a space of memory to hold the image of treatment/record.

  If you want to programmatically return on investment to the whole image set, it may be easier to simply identify the KING with the help of a property node and just pass the image as usual for post-processing.  Treatment that you ever want to performwill automatically be applied to the entire image.

• Implementation of the synchronization steps without using batch or parallel processing sequences

  We use a model of customized integrated process based on sequential process model.

  I have to call the Synchronization Manager in order to use the steps of synchronization when you do not use batch or models of parallel process?

  All operations (locks, Rendezvous, Notifications, queues, etc.) and the steps of synchronization works in any model or a model, with the exception of the synchronization of the lot.

  If you need batch synchronization, your model must use synchronization > advanced > step type specification of batch processing to specify what threads are in the batch. Since you are not from the model of batch processing, I doubt that you have to worry about this.

• Some performances are not completed to help put an end to everything in the model of parallel processes

  Hello

  This question confued me recently, I would use a parallel process of model and to finish running at the same time.

  If there are action steps in order, it seemed that will put an end to the execution of all.

  However, I put a wait instead of steps of the action function, there are a few executions not interrupted while I pressed the button finish everything in the toolbar.

  Hope someone can give me help. Thank you very much.

  PS. the number of testsockets has the value 9 in my teststand.

  I ran your sequence again, and I have reproduced the issue.

  It seems the action of the polls Test for success/failure of the VI at full speed.

  It will consume your CPU resource to 100%.

  So I add some delay in the VI to prevent election full speed.

  Please see the following block diagram.

  

  I tried and it seems resolved the problem.

  Ting

• Press the button on parallel port signal to trigger a process

  I just need to connect a pushbutton on the parallel port to set it as a signal of the beginning of a process.

  When the button is pressed, labview runs a process.

  Kindly help me how to do this, the prayer of wiring

  Lines of data to use as input Some parallel ports can be configured for use as inputs data lines. This depends a lot on how the manufacturer has designed the parallel port. With some models the data lines can be read in the same way that we read the lines of control, leading them to logic high, so they will take the value of an external signal. However, most parallel ports require that you set the bit of sense for entry. It is bit 5 of the control register (base + 2). If the port is capable of it, setting the high bit for effect to sort of lines-State so it can be driven on the outside. Sometimes it is also necessary activate the 6 bit high or low. However, it is noted that some manufacturers actually block those bits so that the software can edit them. An example is shown below in Figure 7.

• Process vs asynchronous sequence in parallel model

  I've been studying the TestStand features and learn how to use it for about a month, so still very new for the environment (although I have been using Labview and Veristand quite strongly for about a year).  I wanted to get some clarification on the use of different process models, because I think I can be lack of understanding of certain terms.

  Here's a little history of my project:

  I have a Labview VI that I created to interface with a remote target (emulator).  Previously, I used the VI to run the tests manually, and I would like to use it as a code module in TestStand so that I can run automated tests.  I intend to use the same VI several times throughout the test sequence.  The functionality of the system is dependent on maintaining constant communication with the emulator, so I can't be opening and closure of the code module several times.  Once opened, it should stay open and communicate continuously (I hope I won't have to create modules of code to be through with my current VI 'wrapper').  Break the communication would result in most of the results of the tests to become invalid.  For these reasons, I chose to call the VI as a module of code into a sequence of void so that it can be run asynchronously, outside the main sequence.

  Now, as I learn more about the details of TestStand, I introduced the notion of "process models.  I was initially using the default sequential process template, but I would like to know if I should come parallel process model.  From what I can tell, the parallel process model is used when multiple Test DUT tests or running in parallel.  Is this correct?  To clarify my situation, I'll test only 1 USE, I use only 1 code module and I run several test with this module of 1 code steps.  I continually have to pass data back and forth with the code module as it runs parallel to the main sequence and there are probably several sequences of void called during the process, so that I can maintain the modularity with my tests.

  So the question is, can I take the parallel process model, or should I continue with the sequential process model and asynchronous sequence for my code module run in parallel?  Thank you very much.

  GSinMN

  I would definitely recommend using the sequential model with a sequence of asynchronous way running for the special code module, in this case. Let us know if you have any other questions on how to architect the system! It is certainly a good idea to plan these kinds of things before we get too deep in development.

• What happens when two parallel process call the same VI in an FPGA?

  Hi all

  I was wondering, what happens when say two parallel processes call the same Subvi in an FPGA? I think that the FPGA would create two instances of the Subvi, not sure. Any point of view?

  Thank you

  Altras

  I don't know if the FPGA adds any ordinary special, but in LabVIEW rule and in LabVIEW Real-time, a VI by default is not reentrant.  Therefore, a process that is called when it is already in use is blocked and must wait.

  If the VI is reentrant, however, then you have two (or more) for the VI dataspaces, so that each can run independently.