SQL Server performance very very LOW in UPDATE/INSERT operations (but not always)

Question

I have a very simple database with this table

This table is used to generate counters for some operations (Order N°, Invoice Num ecc.)

The problem is that on my developer machine the generation of 1000 counters needs 0.3s instead on a Virtual machine with SQL Server on Windows Azure needs 8 seconds more than 20 times slower.

The script I used is this:

    SET STATISTICS IO ON; 
    SET STATISTICS TIME ON;

    CREATE DATABASE mytest;
    GO

    USE mytest;

    CREATE TABLE [dbo].[Counters]
    (
        [SequenceName] [varchar](100) NOT NULL,
        [LastUsed] [bigint] NULL,
      CONSTRAINT [PK_Counters] PRIMARY KEY CLUSTERED ([SequenceName] ASC)
      WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]) ON [PRIMARY]
      GO

        Declare @counter int
        declare @LastUsed as BIGINT
        declare @SequenceName as varchar(100)

        Set @counter = 1

        set @SequenceName = 'AA00TEST-DELETE-THIS'

        set Nocount  on;

        DELETE FROM Counters

        INSERT INTO [dbo].[Counters] ([SequenceName] ,[LastUsed]) VALUES ('AA00TEST-DELETE-THIS', 0)


        While @counter <= 1000
        Begin

             UPDATE Counters
                     SET LastUsed = LastUsed + 1,
                            @LastUsed = LastUsed + 1
                     WHERE SequenceName=@SequenceName

            Set @counter = @counter + 1

        End

        SELECT LastUsed From Counters WHERE SequenceName=@SequenceName

DROP TABLE Counters
GO

USE master
GO

DROP DATABASE mytest
GO

Note tha if I replace the UPDATE istruction with this one the execution time is 0.3s:

SELECT @LastUsed = LastUsed + 1 FROM Counters WHERE SequenceName=@SequenceName

Note that I test this behaviour on many servers and in all I got the same performance.

I test on Azure VM SQL Server with 8 CPU, SQL 2014... 8 Sec.

I test on Azure DB... 8 Sec.

I test on a VMWare virtual machine, SQL 2005 8 sec.

I test on a phisical SQL Server server with 16 CPU, SQL 2008... 8 Sec.

On my MacBook pro with Windows 8 Pro SQL 2012... 0,3sec On my PC with Windows 8 Pro SQL 2012... 0,3sec

The script is executed directly on the server with Managemant Studio connected to localhost. No network roundtrip is present and no application present, I'm directly on the server and on SQL Server core.

This are statistics on my PC (1 iteration):

SQL Server parse and compile time: 
   CPU time = 0 ms, elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 2 ms.
Table 'Counters'. Scan count 0, logical reads 3, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.
Table 'Counters'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

This are statistics on Server (1 iteration):

SQL Server parse and compile time: 
   CPU time = 0 ms, elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 11 ms.
Table 'Counters'. Scan count 0, logical reads 3, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 18 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.
Table 'Counters'. Scan count 0, logical reads 2, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 8 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

 SQL Server Execution Times:
   CPU time = 0 ms,  elapsed time = 0 ms.

All PC And server have the same "Recovery model"

I have found no way to boost it up

Thank You!

Answer 1

I try many many configurations and settings on azure disks and blobs. Finally I found the problem!

The problem is the OBSCENE performance of Windows Azure Disks in Write.

Creating the database on the Temporary Disk of the Azure Server where the virtual machine is hosted the time for execution go to 0sec.

CREATE DATABASE [mytest] ON  PRIMARY 
( NAME = N'mytest_Data', FILENAME = N'D:\Test\mytest_Data.mdf')
    LOG ON 
( NAME = N'mytest_Log', FILENAME = N'D:\Test\mytest_Log.ldf')
GO

Finally there are NO Solution.

I try to use SQL Azure but the performance in UPDATE or INSERT are similar to the SQL Inside the virtual machine. Little bit better.