[PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Stefan Roese]

Performance tests done by AMCC have shown that 256 buffer increase the
performance of the Linux EMAC driver. So let's update the default
values to match this setup.

Signed-off-by: Stefan Roese <sr@denx.de>
---
 drivers/net/ibm_newemac/Kconfig |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ibm_newemac/Kconfig b/drivers/net/ibm_newemac/Kconfig
index 0d3e738..5a06727 100644
--- a/drivers/net/ibm_newemac/Kconfig
+++ b/drivers/net/ibm_newemac/Kconfig
@@ -9,12 +9,12 @@ config IBM_NEW_EMAC
 config IBM_NEW_EMAC_RXB
 	int "Number of receive buffers"
 	depends on IBM_NEW_EMAC
-	default "128"
+	default "256"
 
 config IBM_NEW_EMAC_TXB
 	int "Number of transmit buffers"
 	depends on IBM_NEW_EMAC
-	default "64"
+	default "256"
 
 config IBM_NEW_EMAC_POLL_WEIGHT
 	int "MAL NAPI polling weight"
-- 
1.5.4.rc2

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Benjamin Herrenschmidt]

On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> Performance tests done by AMCC have shown that 256 buffer increase the
> performance of the Linux EMAC driver. So let's update the default
> values to match this setup.
> 
> Signed-off-by: Stefan Roese <sr@denx.de>
> ---

Do we have the numbers ? Did they also measure latency ?

Ben.

>  drivers/net/ibm_newemac/Kconfig |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ibm_newemac/Kconfig b/drivers/net/ibm_newemac/Kconfig
> index 0d3e738..5a06727 100644
> --- a/drivers/net/ibm_newemac/Kconfig
> +++ b/drivers/net/ibm_newemac/Kconfig
> @@ -9,12 +9,12 @@ config IBM_NEW_EMAC
>  config IBM_NEW_EMAC_RXB
>  	int "Number of receive buffers"
>  	depends on IBM_NEW_EMAC
> -	default "128"
> +	default "256"
>  
>  config IBM_NEW_EMAC_TXB
>  	int "Number of transmit buffers"
>  	depends on IBM_NEW_EMAC
> -	default "64"
> +	default "256"
>  
>  config IBM_NEW_EMAC_POLL_WEIGHT
>  	int "MAL NAPI polling weight"

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Stefan Roese]

On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > Performance tests done by AMCC have shown that 256 buffer increase the
> > performance of the Linux EMAC driver. So let's update the default
> > values to match this setup.
> >
> > Signed-off-by: Stefan Roese <sr@denx.de>
> > ---
>
> Do we have the numbers ? Did they also measure latency ?

I hoped this question would not come. ;) No, unfortunately I don't have any 
numbers. Just the recommendation from AMCC to always use 256 buffers.

Best regards,
Stefan

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Benjamin Herrenschmidt]

On Sat, 2008-01-05 at 13:38 +0100, Stefan Roese wrote:
> On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> > On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > > Performance tests done by AMCC have shown that 256 buffer increase the
> > > performance of the Linux EMAC driver. So let's update the default
> > > values to match this setup.
> > >
> > > Signed-off-by: Stefan Roese <sr@denx.de>
> > > ---
> >
> > Do we have the numbers ? Did they also measure latency ?
> 
> I hoped this question would not come. ;) No, unfortunately I don't have any 
> numbers. Just the recommendation from AMCC to always use 256 buffers.

Ok. Well, it's just a .config option so I suppose the patch is fine. Can
you also update the defconfigs (at least for the recent AMCC boards) ?

Ben.

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Josh Boyer]

On Sun, 06 Jan 2008 07:53:06 +1100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> 
> On Sat, 2008-01-05 at 13:38 +0100, Stefan Roese wrote:
> > On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> > > On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > > > Performance tests done by AMCC have shown that 256 buffer increase the
> > > > performance of the Linux EMAC driver. So let's update the default
> > > > values to match this setup.
> > > >
> > > > Signed-off-by: Stefan Roese <sr@denx.de>
> > > > ---
> > >
> > > Do we have the numbers ? Did they also measure latency ?
> > 
> > I hoped this question would not come. ;) No, unfortunately I don't have any 
> > numbers. Just the recommendation from AMCC to always use 256 buffers.
> 
> Ok. Well, it's just a .config option so I suppose the patch is fine. Can
> you also update the defconfigs (at least for the recent AMCC boards) ?

No need for a defconfig update patch.  Paul or I usually do a general
defconfig update for most boards before the next kernel version.  This
will get picked up then.

josh

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Benjamin Herrenschmidt]

On Sat, 2008-01-05 at 15:48 -0600, Josh Boyer wrote:
> No need for a defconfig update patch.  Paul or I usually do a general
> defconfig update for most boards before the next kernel version.  This
> will get picked up then.

Will it ? I think the defconfigs will stick to the old value.

Ben.

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Josh Boyer]

On Sun, 06 Jan 2008 08:54:28 +1100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> 
> On Sat, 2008-01-05 at 15:48 -0600, Josh Boyer wrote:
> > No need for a defconfig update patch.  Paul or I usually do a general
> > defconfig update for most boards before the next kernel version.  This
> > will get picked up then.
> 
> Will it ? I think the defconfigs will stick to the old value.

Put another way, if Kconfig doesn't update it automagically, I'll be
sure to do it myself. :)

josh

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Eugene Surovegin]

On Sat, Jan 05, 2008 at 01:38:17PM +0100, Stefan Roese wrote:
> On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> > On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > > Performance tests done by AMCC have shown that 256 buffer increase the
> > > performance of the Linux EMAC driver. So let's update the default
> > > values to match this setup.
> > >
> > > Signed-off-by: Stefan Roese <sr@denx.de>
> > > ---
> >
> > Do we have the numbers ? Did they also measure latency ?
> 
> I hoped this question would not come. ;) No, unfortunately I don't have any 
> numbers. Just the recommendation from AMCC to always use 256 buffers.

This cannot be true for all chips. Default numbers I selected weren't 
random. In particular, 256 for Tx doesn't make a lot of sense for 405. 
You just gonna waste memory.

I'd be quite reluctant to follow such advices from AMCC without actual 
details. 

-- 
Eugene

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Benjamin Herrenschmidt]

On Fri, 2008-01-11 at 09:48 -0800, Eugene Surovegin wrote:
> On Sat, Jan 05, 2008 at 01:38:17PM +0100, Stefan Roese wrote:
> > On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> > > On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > > > Performance tests done by AMCC have shown that 256 buffer increase the
> > > > performance of the Linux EMAC driver. So let's update the default
> > > > values to match this setup.
> > > >
> > > > Signed-off-by: Stefan Roese <sr@denx.de>
> > > > ---
> > >
> > > Do we have the numbers ? Did they also measure latency ?
> > 
> > I hoped this question would not come. ;) No, unfortunately I don't have any 
> > numbers. Just the recommendation from AMCC to always use 256 buffers.
> 
> This cannot be true for all chips. Default numbers I selected weren't 
> random. In particular, 256 for Tx doesn't make a lot of sense for 405. 
> You just gonna waste memory.
> 
> I'd be quite reluctant to follow such advices from AMCC without actual 
> details. 

I think we can make defaults based on other config options nowadays. Not
very nice but we could do things like

	default 128 if PPC_40x
	default 256

Or even more detailed.

Cheers,
Ben.

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Stefan Roese]

On Friday 11 January 2008, Benjamin Herrenschmidt wrote:
> On Fri, 2008-01-11 at 09:48 -0800, Eugene Surovegin wrote:
> > On Sat, Jan 05, 2008 at 01:38:17PM +0100, Stefan Roese wrote:
> > > On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> > > > On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > > > > Performance tests done by AMCC have shown that 256 buffer increase
> > > > > the performance of the Linux EMAC driver. So let's update the
> > > > > default values to match this setup.
> > > > >
> > > > > Signed-off-by: Stefan Roese <sr@denx.de>
> > > > > ---
> > > >
> > > > Do we have the numbers ? Did they also measure latency ?
> > >
> > > I hoped this question would not come. ;) No, unfortunately I don't have
> > > any numbers. Just the recommendation from AMCC to always use 256
> > > buffers.
> >
> > This cannot be true for all chips. Default numbers I selected weren't
> > random. In particular, 256 for Tx doesn't make a lot of sense for 405.
> > You just gonna waste memory.

This may be the case with the "old" 405 PPC's. But with the new ones coming 
out right now, like the up to 666MHz 405EX with GBit support, 256 could be an 
improvement. I still owe you figures though. Will try to do some testing in a 
short while.

> > I'd be quite reluctant to follow such advices from AMCC without actual
> > details.
>
> I think we can make defaults based on other config options nowadays. Not
> very nice but we could do things like
>
> 	default 128 if PPC_40x
> 	default 256
>
> Or even more detailed.

We shouldn't make it too complicated. We can always select different settings 
in the defconfig file. My thinking here is to better wast a little memory 
with a potential performance improvement. Just me 0.02$

Best regards,
Stefan

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Benjamin Herrenschmidt]

On Sat, 2008-01-12 at 08:26 +0100, Stefan Roese wrote:
> 
> We shouldn't make it too complicated. We can always select different
> settings 
> in the defconfig file. My thinking here is to better wast a little
> memory 
> with a potential performance improvement. Just me 0.02$

If it gets really critical, then we can move those settings to the
device-tree.

Ben.

Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers

[Eugene Surovegin]

On Sat, Jan 12, 2008 at 06:57:24PM +1100, Benjamin Herrenschmidt wrote:
> 
> On Sat, 2008-01-12 at 08:26 +0100, Stefan Roese wrote:
> > 
> > We shouldn't make it too complicated. We can always select different
> > settings 
> > in the defconfig file. My thinking here is to better wast a little
> > memory 
> > with a potential performance improvement. Just me 0.02$
> 
> If it gets really critical, then we can move those settings to the
> device-tree.

Come on guys, it's not that critical. I guess I just don't trust a 
certain company :)

-- 
Eugene