Hardware Extension Support in SQL Server 2016

Microsoft’s Bob Dorr has recently consolidated a series of CSS posts about many low-level performance improvements in SQL Server 2016. One of these, How It Works: SQL Server 2016 SSE/AVX Support talks about how SQL Server 2016 added the use of SSE and AVX instructions in various code paths to improve performance when your processor supports one of those instruction sets.

Unless you are a hardware enthusiast, you might be wondering what these instructions are and how you can determine whether your processor supports them or not. First, we have some basic definitions:

Advanced Vector Extensions (AVX) were introduced in the server space with the Sandy Bridge-EP architecture, which was released in Q1 2012. This was the Intel Xeon E5-2600 Product Family. This means that if you have a server with Intel Xeon E5-2600 (Sandy Bridge-EP), Intel Xeon E5-2600 v2 (Ivy Bridge-EP), Intel Xeon E5-2600 v3 (Haswell-EP), or Intel Xeon E5-2600 v4 (Broadwell-EP) processors, you will have AVX support which SQL Server 2016 will use for better performance.

You will also have AVX support with the Intel Xeon E7-4800/8800 v2 (Ivy Bridge-EX), Intel Xeon E7-4800/8800 v3 (Haswell_EX), and Intel Xeon E7-4800/8800 v4 (Broadwell-EX) product families. Anything older than these families will not have AVX support.

This is yet another reason to use a new server when you upgrade to SQL Server 2016 (which is a good idea anyway for lots of other reasons).

Advanced Vector Extensions 2 (AVX2), is a newer version of AVX that was introduced with the Haswell-EP/EX architecture. If you have a Haswell-EP/EX or Broadwell-EP/EX processor, then you will have AVX2 support.

Streaming SIMD Extensions (SSE) are a much older instruction set that were introduced back in 1999. Since then, there have been several newer versions such as SSE2, SSE3, SSE4.1, and SSE4.2. The latest SSE4.2 version was introduced in the Nehalem-EP/EX architecture, which came out in Q1 2009 and Q1 2010 respectively. This means that the Intel Xeon 5500 (Nehalem-EP) and Intel Xeon 7500 (Nehalem-EX) Product families are the oldest models that will have SSE4.2 support. From Bob’s post, it appears that the same code path in SQL Server 2016 will be used if you have any version of SSE (but you don’t have AVX or AVX2 support), so even much older Xeon processors will have some level of SSE support. Just to be clear, AVX and AVX2 are newer and better than any version of SSE.

The CPU-Z utility shows the various extension instructions supported by your processor. As you can see in Figure 1, the Intel Xeon E5-2690 supports the AVX instruction set.


Figure 1: Intel Xeon E5-2690 CPU Details

A current model Intel Core i7-6700K Skylake desktop processor supports the newer FMA, and Transactional Synchronization Extensions (TSX) instructions, as shown in Figure 2. These were both introduced with the Haswell microarchitecture, and they might be useful in a future version of SQL Server, if Microsoft decides to use them.


 Figure 2: Intel Core i7-6700K CPU Details

SQL Server Diagnostic Information Queries for June 2016

This month, I have several improvements in the SQL Server 2014 and 2016 sets, along with additional comments and documentation in the SQL Server 2012, 2014 and 2016 sets. One new feature is a series of six dashes after each query to make them easier to parse by people who use PowerShell to run the queries and collect the results. I have gotten quite a bit of interest about making a special version of these queries for SQL Database in Microsoft Azure. So, I will be doing that pretty soon.

Rather than having a separate blog post for each version, I have just put the links for all six major versions in this single post. There are two separate links for each version. The first one on the top left is the actual diagnostic query script, and the one below on the right is the matching blank results spreadsheet, with labeled tabs that correspond to each query in the set. 

Here are links to the latest versions of these queries for SQL Server 2016, 2014 and 2012:

SQL Server 2016 Diagnostic Information Queries (June 2016)

SQL Server 2016 Blank Results

SQL Server 2014 Diagnostic Information Queries (June 2016)

SQL Server 2014 Blank Results

SQL Server 2012 Diagnostic Information Queries (June 2016)

SQL Server 2012 Blank Results

Here are links to the most recent versions of these scripts for SQL Server 2008 R2 and older:

Since SQL Server 2008 R2 and older are out of Mainstream support from Microsoft (and because fewer of my customers are using these old versions of SQL Server), I am not going to be updating the scripts for these older versions of SQL Server every single month going forward.  I started this policy a while ago, and so far, I have not heard any complaints. I did update these queries slightly in January 2016 though.

SQL Server 2008 R2 Diagnostic Information Queries (CY 2016)

SQL Server 2008 R2 Blank Results

SQL Server 2008 Diagnostic Information Queries (CY 2016)

SQL Server 2008 Blank Results

SQL Server 2005 Diagnostic Information Queries (CY 2016)

SQL Server 2005 Blank Results

The basic instructions for using these queries is that you should run each query in the set, one at a time (after reading the directions for that query). It is not really a good idea to simply run the entire batch in one shot, especially the first time you run these queries on a particular server, since some of these queries can take some time to run, depending on your workload and hardware. I also think it is very helpful to run each query, look at the results (and my comments on how to interpret the results) and think about the emerging picture of what is happening on your server as you go through the complete set. I have some comments in the script on how to interpret the results after each query.

After running each query, you need to click on the top left square of the results grid in SQL Server Management Studio (SSMS) to select all of the results, and then right-click and select “Copy with Headers” to copy all of the results, including the column headers to the Windows clipboard. Then you paste the results into the matching tab in the blank results spreadsheet.

About half of the queries are instance specific and about half are database specific, so you will want to make sure you are connected to a database that you are concerned about instead of the master system database. Running the database-specific queries while being connected to the master database is a very common mistake that I see people making when they run these queries.

Note: These queries are stored on Dropbox. I occasionally get reports that the links to the queries and blank results spreadsheets do not work, which is most likely because Dropbox is blocked wherever people are trying to connect.

I also occasionally get reports that some of the queries simply don’t work. This usually turns out to be an issue where people have some of their user databases in 80 compatibility mode, which breaks many DMV queries, or that someone is running an incorrect version of the script for their version of SQL Server.

It is very important that you are running the correct version of the script that matches the major version of SQL Server that you are running. There is an initial query in each script that tries to confirm that you are using the correct version of the script for your version of SQL Server. If you are not using the correct version of these queries for your version of SQL Server, some of the queries are not going to work correctly.

If you want to understand how to better run and interpret these queries, you should consider listening to my three latest Pluralsight courses, which are SQL Server 2014 DMV Diagnostic Queries – Part 1SQL Server 2014 DMV Diagnostic Queries – Part 2 and SQL Server 2014 DMV Diagnostic Queries – Part 3. All three of these courses are pretty short and to the point, at 67, 77, and 68 minutes respectively. Listening to these three courses is really the best way to thank me for maintaining and improving these scripts…

Please let me know what you think of these queries, and whether you have any suggestions for improvements. Thanks!