Re: Hyperthreading Options in 2.4.19

Re: Hyperthreading Options in 2.4.19

[Renato]

I read a while back that the kernel boot parameter "idle=poll" helps to improve the wake up latency of processes in high 
performance systems.

On Wed, 07 Aug 2002 14:22:05 -0400, "Reed, Timothy A" <timothy.a.reed@lmco.com> escreveu :

> De: "Reed, Timothy A" <timothy.a.reed@lmco.com>
> Data: Wed, 07 Aug 2002 14:22:05 -0400
> Para: "Linux Kernel ML (E-mail)" <linux-kernel@vger.kernel.org>
> Assunto: Hyperthreading Options in 2.4.19
> 
> Hello All,
> 	I am going rounds with a sub-contractor of ours about what options
> should and should not be compiled into the kernel in order for
> Hyperthreading to work.  Can anyone make any suggestions and comments to the
> options (below)  that I am planning on enforcing:
> 	MSR
> 	MTRR
> 	CPUID
> 
> 	Lilo.conf : acpismp=force?? 
> 
> 	Are the following worth any thing of value to Hyperthreading:
> 	Microcode
> 	ACPI
> 
> TIA
> 
> Timothy Reed
> Software Engineer/Systems Administrator
> Lockheed Martin - NE & SS Syracuse
> Email: timothy.a.reed@lmco.com
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 
> 
> 

Hyperthreading Options in 2.4.19

[Reed, Timothy A]

Hello All,
	I am going rounds with a sub-contractor of ours about what options
should and should not be compiled into the kernel in order for
Hyperthreading to work.  Can anyone make any suggestions and comments to the
options (below)  that I am planning on enforcing:
	MSR
	MTRR
	CPUID

	Lilo.conf : acpismp=force?? 

	Are the following worth any thing of value to Hyperthreading:
	Microcode
	ACPI

TIA

Timothy Reed
Software Engineer/Systems Administrator
Lockheed Martin - NE & SS Syracuse
Email: timothy.a.reed@lmco.com

Re: Hyperthreading Options in 2.4.19

[J.A. Magallon]

On 2002.08.07 "Reed, Timothy A" wrote:
>Hello All,
>	I am going rounds with a sub-contractor of ours about what options
>should and should not be compiled into the kernel in order for
>Hyperthreading to work.  Can anyone make any suggestions and comments to the
>options (below)  that I am planning on enforcing:
>	MSR
>	MTRR
>	CPUID

I thikn none is needed for ht. Of course, mtrr raises performance.
The other are not needed, afaik.

>
>	Lilo.conf : acpismp=force?? 
>

True for old kernels, not needed anymore.

>	Are the following worth any thing of value to Hyperthreading:
>	Microcode
>	ACPI

No.

In summary, with 2.4.19 you do not have to do nothing to have hyperthreading.
Other useful options are 'noht' to disable ht, and 'idle=poll', that I think
improves latency.

by

-- 
J.A. Magallon <jamagallon@able.es>      \                 Software is like sex:
werewolf.able.es                         \           It's better when it's free
Mandrake Linux release 9.0 (Cooker) for i586
Linux 2.4.20-pre1-jam1 (gcc 3.2 (Mandrake Linux 9.0 3.2-0.2mdk))

RE: Hyperthreading Options in 2.4.19

[Reed, Timothy A]

Guys,
	Thanks for the answers!  In doing research on those options I too
found that they were not need for Hyperthreading, but I needed something
more to back up opinion.

Thanks Again.

Timothy Reed
Software Engineer\Systems Administrator
Lockheed Martin - NE & SS Syracuse
Email: timothy.a.reed@lmco.com


-----Original Message-----
From: J.A. Magallon [mailto:jamagallon@able.es]
Sent: Wednesday, August 07, 2002 6:37 PM
To: Reed, Timothy A
Cc: Linux Kernel ML (E-mail)
Subject: Re: Hyperthreading Options in 2.4.19



On 2002.08.07 "Reed, Timothy A" wrote:
>Hello All,
>	I am going rounds with a sub-contractor of ours about what options
>should and should not be compiled into the kernel in order for
>Hyperthreading to work.  Can anyone make any suggestions and comments to
the
>options (below)  that I am planning on enforcing:
>	MSR
>	MTRR
>	CPUID

I thikn none is needed for ht. Of course, mtrr raises performance.
The other are not needed, afaik.

>
>	Lilo.conf : acpismp=force?? 
>

True for old kernels, not needed anymore.

>	Are the following worth any thing of value to Hyperthreading:
>	Microcode
>	ACPI

No.

In summary, with 2.4.19 you do not have to do nothing to have
hyperthreading.
Other useful options are 'noht' to disable ht, and 'idle=poll', that I think
improves latency.

by

-- 
J.A. Magallon <jamagallon@able.es>      \                 Software is like
sex:
werewolf.able.es                         \           It's better when it's
free
Mandrake Linux release 9.0 (Cooker) for i586
Linux 2.4.20-pre1-jam1 (gcc 3.2 (Mandrake Linux 9.0 3.2-0.2mdk))