Skip to main content

Maxdop and Cost theshold for parallelism SQL Server

 Maxdop 

Maxdop stands for max degree for parallelism. 

Let's say, the maxdop is set to 4, it means during parallel plan execution, SQL server is going to use 4 processors. If you set your Maxdop settings to 0, it means SQL server is going to use as many processors it needs to complete your request. 



So it sounds like keeping maxdop to 0 is good, since more processors mean less processing time, right? It's not true all the time. SO, Let's dig in.

I will be using AdventureWorks2019 Database for the following.

                                  Adventureworks2019

First, I use the following query to find out the size of each of the tables of the database.

SELECT 
    t.NAME AS TableName,
    s.Name AS SchemaName,
    p.rows,
    SUM(a.total_pages) * 8 AS TotalSpaceKB, 
    CAST(ROUND(((SUM(a.total_pages) * 8) / 1024.00), 2) AS NUMERIC(36, 2)) AS TotalSpaceMB,
    SUM(a.used_pages) * 8 AS UsedSpaceKB, 
    CAST(ROUND(((SUM(a.used_pages) * 8) / 1024.00), 2) AS NUMERIC(36, 2)) AS UsedSpaceMB, 
    (SUM(a.total_pages) - SUM(a.used_pages)) * 8 AS UnusedSpaceKB,
    CAST(ROUND(((SUM(a.total_pages) - SUM(a.used_pages)) * 8) / 1024.00, 2) AS NUMERIC(36, 2)) AS UnusedSpaceMB
FROM 
    sys.tables t
INNER JOIN      
    sys.indexes i ON t.OBJECT_ID = i.object_id
INNER JOIN 
    sys.partitions p ON i.object_id = p.OBJECT_ID AND i.index_id = p.index_id
INNER JOIN 
    sys.allocation_units a ON p.partition_id = a.container_id
LEFT OUTER JOIN 
    sys.schemas s ON t.schema_id = s.schema_id
WHERE 
    t.NAME NOT LIKE 'dt%' 
    AND t.is_ms_shipped = 0
    AND i.OBJECT_ID > 255 
GROUP BY 
    t.Name, s.Name, p.Rows
ORDER BY 
    TotalSpaceMB DESC, t.Name

This outputs something like this -

Alternatively, we can use the built reports of SQL Server to find out the size of the tables. For this you need to go to DatabaseName-> Right Click -> Select Report -> Disk Usage By Top Tables 

Disk Usage By Top Tables SQL

This will output something like this 

Disk Usage By Top Tables  SQL

What is the recommended settings for Maxdop?

The rule of thumb is -
  • 8 or less processors ===> 0 to N (where N= no. of processors)
  • 8 More than 8 processors ===> 8
  • NUMA configured ===> MAXDOP should not exceed no of CPU’s assigned to each NUMA node with max value capped to 8
  • Hyper threading Enabled ===> Should not exceed the number of physical processors.
The following query gives you the optimal value for MaxDop Settings -
declare @hyperthreadingRatio bit
declare @logicalCPUs int
declare @HTEnabled int
declare @physicalCPU int
declare @SOCKET int
declare @logicalCPUPerNuma int
declare @NoOfNUMA int

select @logicalCPUs = cpu_count -- [Logical CPU Count]
    ,@hyperthreadingRatio = hyperthread_ratio --  [Hyperthread Ratio]
    ,@physicalCPU = cpu_count / hyperthread_ratio -- [Physical CPU Count]
    ,@HTEnabled = case 
        when cpu_count > hyperthread_ratio
            then 1
        else 0
        end -- HTEnabled
from sys.dm_os_sys_info
option (recompile);

select @logicalCPUPerNuma = COUNT(parent_node_id) -- [NumberOfLogicalProcessorsPerNuma]
from sys.dm_os_schedulers
where [status] = 'VISIBLE ONLINE'
    and parent_node_id < 64
group by parent_node_id
option (recompile);

select @NoOfNUMA = count(distinct parent_node_id)
from sys.dm_os_schedulers -- find NO OF NUMA Nodes 
where [status] = 'VISIBLE ONLINE'
    and parent_node_id < 64

-- Report the recommendations ....
select
    --- 8 or less processors and NO HT enabled
    case 
        when @logicalCPUs < 8
            and @HTEnabled = 0
            then 'MAXDOP setting should be : ' + CAST(@logicalCPUs as varchar(3))
                --- 8 or more processors and NO HT enabled
        when @logicalCPUs >= 8
            and @HTEnabled = 0
            then 'MAXDOP setting should be : 8'
                --- 8 or more processors and HT enabled and NO NUMA
        when @logicalCPUs >= 8
            and @HTEnabled = 1
            and @NoofNUMA = 1
            then 'MaxDop setting should be : ' + CAST(@logicalCPUPerNuma / @physicalCPU as varchar(3))
                --- 8 or more processors and HT enabled and NUMA
        when @logicalCPUs >= 8
            and @HTEnabled = 1
            and @NoofNUMA > 1
            then 'MaxDop setting should be : ' + CAST(@logicalCPUPerNuma / @physicalCPU as varchar(3))
        else ''
        end as Recommendations

For my 2 cores and 4 logical processors, my maxdop setting suggestion is 4.

What can happen if I set Maxdop to 0 (default) ?

You are going to face Threadpool wait - A deadly poison wait for SQL Server sooner or later. 

MAXDOP(1) - What does it imply?

MAXDOP(1) simply allows you to execute SQL requests serially, No parallel processing will be done. These are the following things you can achieve by MAXDOP(1)

  • Serial Processing 
  • Processing requests in a single processor
  • Less chance of Threadpool Wait.

You have to choose MAXDOP(1) intelligently. Since -
  • MAXDOP(1) is going to use more processing time (coz only one processor is used.)
  • As a result, the outputs that were fast before, will be bit slow.

When to set Maxdop(1)

  • For queries that gets the filters of your page or report
  • For queries that cost less (Query Cost)

Comments

Most Loved Posts

Threadpool - A deadly poison wait for SQL Server (The What, When and How)

Introduction  Threadpool is a  poison  wait. Yes, I mean it. Its poison for SQL Server, its poison for the Business and of course, the end-users! The most devastating thing about threadpool is you hardly recognize it because it comes in disguise, meaning you see no memory or cpu pressure in the system, yet you cannot run any query, it seems like your SQL Server is frozen solid. That scary, isn't it?

How to Backup SQL server like Batman: The Ultimate SQL DBA Guide

Intelligent Query Processing in SQL Server 2019 Big Data

SQL Server 2019: Intelligent Query Processing SQL Server 2019 ships with some brand-new features. Many of these features are targeted for Big Data Solutions. No wonder in that, since the world is moving faster towards Big Data and it is absolutely necessary to cope up with that. Today we will discuss one such feature called Approximate Query Processing. Approximate Query Processing SQL Server ships with Intelligent Query Processing out of the box with SQL Server 2019 installation. Approximate Query processing is a part of Intelligent Query Processing. Things we will be covering in this article – Understand the need for Approximation with Case Study Case Study 1: Railway Case Study 2: e-commerce How to use Approximate Query Processing Demo Code for Comparing Performance Results Limitations When to avoid Approximate Query Processing Understand the need for Approximate Query Processing Before using any technological feature, we must understand why we should use it? Should we jus...

Why not use Select * in SQL Server

  Select * We often use the Select *  to fetch data from tables of SQL Server.

SQL Schema Compare with Visual Studio (A complete Guide)

Introduction When you're working on your Dev Database, an urgent issue comes along, and you instantly solve it by changing Scheme in the Staging Database or Production Database :3, few more these type of patching and you're completely out of sync! A lot of paid alternatives are there like SQL Data Compare by RedGate, but my first choice is Visual Studio's SQL Data Tools. In the following article, I tried to image-describe the steps for SQL Data Tool. Like I said before, there are lots of handly DBAtools out there to compare Schema between two DB Sources. I would like to discuss how you can compare two SQL Server DB with Visual Studio. Make sure you have SQL Server Data tools checked while installing Visual Studio.

How to generate C# Class from SQL Server Table

C# Class from SQL Database Table There are multiple ways you can generate a C# class from your Database Table. We will be covering the following topics in today’s article. Generate Class with foreign key relation Generate Class with only entities Generate Class with foreign key relations For this we can simply use Entity Frameworks EDMX update feature which will generate our C# class from Database Tables. The output from EDMX will contain foreign key relations which we can see from the virtual interfaces like this -  public partial class AssetItem { [System.Diagnostics.CodeAnalysis.SuppressMessage("Microsoft.Usage", "CA2214:DoNotCallOverridableMethodsInConstructors")] public AssetItem() { this .AssetItemDepreciations = new HashSet<AssetItemDepreciation>(); this .AssetTaxMappings = new HashSet<AssetTaxMapping>(); this .AssetVatMappings = new HashSet<AssetVatMapping>(); ...

SQL Data Tools - Compare Data

Compare Data between two tables SQL Server Database with the same schema architecture can differ in different environments like Dev, Staging, and Production, especially in configuration tables. Let's see how we can easily sync the data in two different tables.

Slow SQL Server : What we should NOT do

 Try to list the best practices of SQL Server. It will require a heck of a time. Try to list the Bad Practices and it will require more than the best practice list , of course, probably you’ll end up getting frustrated . (seeing all the oops configurations and its effect on SQL Server )

How to configure your Availability Group listener to ASP.NET

SQL Server’s availability group Always On feature is great to have features for your Database. Anytime one of your database nodes goes down, your secondary replica will automatically take over. After a failover, your secondary cluster node becomes the primary cluster. Now the question arises, “Do I need to configure my APP server connectionstring each time I face a failover cluster?”. The answer is NO, you don’t have to configure your app server connectionstring every time. Default ConnectionString By default, your App server connectionstring looks something like this – <connectionStrings>    <add name="ConnStringDb1" connectionString="Data Source=localhost;Initial Catalog=YourDataBaseName;Integrated Security=True;" providerName="System.Data.SqlClient" />   </connectionStrings> ConnectionString for Failover Partner You can manually specify the failover partner in your connectionstring like this <connectionStrings>     <a...