Hi
Has anyone had any SQL performance issues when Hyper Threading is enabled on
BIOS? Do you have any comments regarding to HT?
I know this is a bit of a grey area for Citrix - but i have not seen any
comments about this for SQL
Regards
James
On our SQL boxes, we keep HT enabled and tweak the SQL Server max degree of
parallelism config option to specify no more than the number of processor
cores.
Every application is different so, if performance is important to you,
consider taking the time to run benchmarks with your application under
various configurations.
Hope this helps.
Dan Guzman
SQL Server MVP
"James" <hushdontspamme@.hotmail.com> wrote in message
news:eU68yba9FHA.1148@.tk2msftngp13.phx.gbl...
> Hi
> Has anyone had any SQL performance issues when Hyper Threading is enabled
> on BIOS? Do you have any comments regarding to HT?
> I know this is a bit of a grey area for Citrix - but i have not seen any
> comments about this for SQL
> Regards
> James
>
|||Have a look here:
http://blogs.msdn.com/slavao/archive...12/492119.aspx
Markus
|||<Excerpt>
So does it mean you have to disable HT when using SQL Server? The answer is
it really depends on the load and hardware you are using.
You have to test your application with HT on and off under heavy loads to
understand HT's implications.
</Excerpt>
There is no substitute for thorough performance testing when you have a
demanding application and/or expect to tax the hardware. I ran a benchmark
with our app on a 4-way (8 logical) proc server and we got 15-20% percent
improvement with HT enabled at the bios level. It would have been a pity to
throw this away.
Hope this helps.
Dan Guzman
SQL Server MVP
"MarkusB" <m.bohse@.quest-consultants.com> wrote in message
news:1133359063.147112.317730@.z14g2000cwz.googlegr oups.com...
> Have a look here:
> http://blogs.msdn.com/slavao/archive...12/492119.aspx
> Markus
>
|||Hi
Hi I wonder which tools (benchmarks) are you using to get the
performace infomation.
Thanks
Yoel
Dan Guzman wrote:
> <Excerpt>
> So does it mean you have to disable HT when using SQL Server? The answer is
> it really depends on the load and hardware you are using.
> You have to test your application with HT on and off under heavy loads to
> understand HT's implications.
> </Excerpt>
> There is no substitute for thorough performance testing when you have a
> demanding application and/or expect to tax the hardware. I ran a benchmark
> with our app on a 4-way (8 logical) proc server and we got 15-20% percent
> improvement with HT enabled at the bios level. It would have been a pity to
> throw this away.
>
|||I performed a controlled test using our production application code and
data. The performance I reported was based on the actual elapsed time
difference. IMHO, this is the most meaningful type of test since the
objective is to optimize the server for production application processing.
Importantly, the application I tested was typical OLTP and highly optimized
with very few scans. Results could be different with an OLAP/reporting
application profile, a different OLTP application or different application
mix. I mentioned this when I posted to Slava's blog.
It can take considerable work to develop and run application benchmarks like
this. Such effort can be justified when you have demanding mission critical
applications that will fully tax your hardware but perhaps not justified
when hardware resources are less utilized. In my case, performance testing
and tuning was required anyway so that we could perform an application
migration in the shortest possible time.
Hope this helps.
Dan Guzman
SQL Server MVP
"Yoel Zumbado" <yzumbado@.hotmail.com> wrote in message
news:OK8Cfe0AGHA.3984@.TK2MSFTNGP14.phx.gbl...[vbcol=seagreen]
> Hi
> Hi I wonder which tools (benchmarks) are you using to get the performace
> infomation.
> Thanks
> Yoel
> Dan Guzman wrote:
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment