|
List Info
Thread: Speed of FlexBus on the MCF547x
|
|
| Speed of FlexBus on the MCF547x |
|
2006-07-11 06:38:31 |
Hello John,
first of all thanks for the quick answer.
Our intention is to upgrade our current processor (MCF5407)
to a more
powerful one with at least 30% more performance so the
MCF5474 seemed to be
a good candidate.
If there is going to be a new high-end ColdFire device I am
very
interested!
Best regards
Holger Krause.
------------------------------------------------------------
----
SIEB & MEYER AG
Software-Entwicklung CNC / Holger Krause
Auf dem Schmaarkamp 21
D-21339 Lüneburg
Germany
Tel.: +49-(0)4131 203-1505
Fax.: +49-(0)4131 38562
E-Mail: Holger.Krause sieb-meyer.de
http://www.sieb-meyer.de
John Bodnar
<john.bodnarfree
scale.com>
An
holger.krausesieb-meyer.de
10.07.2006 22:14
Kopie
Thema
Re: [ColdFire] Speed
of FlexBus on
the MCF547x
Hello Holger,
I think what you're seeing is an artifact of the MCF547x/8x
design.
We built this device on a PowerPC platform, so the internal
bus is
64-bit XLB. The V4e ColdFire core, however, only has a
32-bit bus
interface and cannot issue burst transactions across the
FlexBus, If I
remember correctly.
Our new devices, the MCF5207/8, MCF532x, and MCF537x are
built on a
different platform with native ColdFire buses throughout and
support
core bursts to the FlexBus.
We're working on a new high-end ColdFire device that also
uses the
FlexBus (it is now our standard external bus interface
module), and you
may want to check it out when it is introduced in a few
months time.
Best regards,
John Bodnar
Systems Engineer
Microcontroller Division
--
holger.krausesieb-meyer.de wrote:
> Hello,
>
> we want to connect SRAM to the FlexBus of our MCF5474
board. So we
> configured the appropriate chip selects and wrote a
simple routine which
> accesses this memory. Everything works fines but the
performance seems to
> be too low: One access is very fast (~25ns) but the
time between two
> accesses is about 130ns! => this is slower than the
MCF5407.
> The FBCSn-signal is active for about <20ns - this is
very fast - but the
> time between the raising edge of FBCSn and the falling
edge of the next
> FBCS is very long (about > 100ns).The R/W signal is
valid for the whole
> time so there are no other accesses. We don't know
what the processor is
> doing in the meantime. Or is it already full speed?
>
> We also changed the configuration of the arbiter but
unfortunantely it
> seems to have no effect.
>
> The CLKIN is 66.66MHz, the CS has no waitstates, IRQs
are off and the
code
> runs in the core SRAM.
>
> Has anyone similar experiences or advices for improving
the speed?
>
> Best regards,
>
> Holger Krause.
>
>
------------------------------------------------------------
----
> SIEB & MEYER AG
> Software-Entwicklung CNC / Holger Krause
> Auf dem Schmaarkamp 21
> D-21339 Lüneburg
> Germany
> Tel.: +49-(0)4131 203-1505
> Fax.: +49-(0)4131 38562
> E-Mail: Holger.Krausesieb-meyer.de
> http://www.sieb-meyer.de
>
>
------------------------------------------------------------
--------
> To Subscribe send a message to: ColdFire-OnLists.Wildrice.com
> To Unsubscribe send a message to: ColdFire-OffLists.Wildrice.com
> For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
>
>
--
POPI Classification
[ ]General Business Information
[x]Freescale Internal Use
[ ]Freescale Confidential Proprietary
------------------------------------------------------------
--------
To Subscribe send a message to: ColdFire-OnLists.Wildrice.com
To Unsubscribe send a message to: ColdFire-OffLists.Wildrice.com
For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
|
|
| Speed of FlexBus on the MCF547x |
|
2006-07-11 08:43:30 |
Hello John, Holger,
we are hooking up a FPGA to the Flexbus (for 2D graphics to
HDTV) on one
of our experimental boards, which has not
been brought up yet.
What is a good ballpark figure of the the data throughput I
can expect?
Regards,
Leon.
holger.krausesieb-meyer.de wrote:
> Hello John,
>
> first of all thanks for the quick answer.
>
> Our intention is to upgrade our current processor
(MCF5407) to a more
> powerful one with at least 30% more performance so the
MCF5474 seemed to be
> a good candidate.
> If there is going to be a new high-end ColdFire device
I am very
> interested!
>
> Best regards
> Holger Krause.
>
>
------------------------------------------------------------
----
> SIEB & MEYER AG
> Software-Entwicklung CNC / Holger Krause
> Auf dem Schmaarkamp 21
> D-21339 Lüneburg
> Germany
> Tel.: +49-(0)4131 203-1505
> Fax.: +49-(0)4131 38562
> E-Mail: Holger.Krausesieb-meyer.de
> http://www.sieb-meyer.de
>
>
>
>
> John Bodnar
> <john.bodnarfree
> scale.com>
An
>
holger.krausesieb-meyer.de
> 10.07.2006 22:14
Kopie
>
>
Thema
> Re: [ColdFire]
Speed of FlexBus on
> the MCF547x
>
>
>
>
>
>
>
>
>
>
>
> Hello Holger,
>
> I think what you're seeing is an artifact of the
MCF547x/8x design.
>
> We built this device on a PowerPC platform, so the
internal bus is
> 64-bit XLB. The V4e ColdFire core, however, only has a
32-bit bus
> interface and cannot issue burst transactions across
the FlexBus, If I
> remember correctly.
>
> Our new devices, the MCF5207/8, MCF532x, and MCF537x
are built on a
> different platform with native ColdFire buses
throughout and support
> core bursts to the FlexBus.
>
> We're working on a new high-end ColdFire device that
also uses the
> FlexBus (it is now our standard external bus interface
module), and you
> may want to check it out when it is introduced in a few
months time.
>
> Best regards,
>
> John Bodnar
> Systems Engineer
> Microcontroller Division
>
> --
>
> holger.krausesieb-meyer.de wrote:
>
>> Hello,
>>
>> we want to connect SRAM to the FlexBus of our
MCF5474 board. So we
>> configured the appropriate chip selects and wrote a
simple routine which
>> accesses this memory. Everything works fines but
the performance seems to
>> be too low: One access is very fast (~25ns) but the
time between two
>> accesses is about 130ns! => this is slower than
the MCF5407.
>> The FBCSn-signal is active for about <20ns -
this is very fast - but the
>> time between the raising edge of FBCSn and the
falling edge of the next
>> FBCS is very long (about > 100ns).The R/W signal
is valid for the whole
>> time so there are no other accesses. We don't
know what the processor is
>> doing in the meantime. Or is it already full speed?
>>
>> We also changed the configuration of the arbiter
but unfortunantely it
>> seems to have no effect.
>>
>> The CLKIN is 66.66MHz, the CS has no waitstates,
IRQs are off and the
>>
> code
>
>> runs in the core SRAM.
>>
>> Has anyone similar experiences or advices for
improving the speed?
>>
>> Best regards,
>>
>> Holger Krause.
>>
>>
------------------------------------------------------------
----
>> SIEB & MEYER AG
>> Software-Entwicklung CNC / Holger Krause
>> Auf dem Schmaarkamp 21
>> D-21339 Lüneburg
>> Germany
>> Tel.: +49-(0)4131 203-1505
>> Fax.: +49-(0)4131 38562
>> E-Mail: Holger.Krausesieb-meyer.de
>> http://www.sieb-meyer.de
>>
>>
------------------------------------------------------------
--------
>> To Subscribe send a message to:
ColdFire-OnLists.Wildrice.com
>> To Unsubscribe send a message to:
ColdFire-OffLists.Wildrice.com
>> For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
>>
>>
>>
>
> --
>
> POPI Classification
> [ ]General Business Information
> [x]Freescale Internal Use
> [ ]Freescale Confidential Proprietary
>
>
>
>
------------------------------------------------------------
--------
> To Subscribe send a message to: ColdFire-OnLists.Wildrice.com
> To Unsubscribe send a message to: ColdFire-OffLists.Wildrice.com
> For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
>
>
------------------------------------------------------------
--------
To Subscribe send a message to: ColdFire-OnLists.Wildrice.com
To Unsubscribe send a message to: ColdFire-OffLists.Wildrice.com
For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
|
|
| Speed of FlexBus on the MCF547x |
|
2006-07-11 16:43:15 |
Holger,
I think you'll see improved performance from the 547x over
the 5407 with
the increased core and SDRAM bus frequencies, the larger
caches, and the
dedicated SDRAM bus. Furthermore, the DMA module can take
advantage of
the 64-bit internal buses to deliver fairly high performance
for the
Ethernets.
As for the new high-end part, I'm not a liberty to say much
about it,
but it is on our publicly disclosed roadmap here:
http://www.freescale.com/files/archives/doc/ro
admap/COLDFIRERD.pdf
Needless to say, I think y'all will be pretty excited when
it comes out
 I
certainly think it's pretty cool, but that's just me --
I've been
up to my ears working on it
Regards,
John
--
holger.krausesieb-meyer.de wrote:
> Hello John,
>
> first of all thanks for the quick answer.
>
> Our intention is to upgrade our current processor
(MCF5407) to a more
> powerful one with at least 30% more performance so the
MCF5474 seemed to be
> a good candidate.
> If there is going to be a new high-end ColdFire device
I am very
> interested!
>
> Best regards
> Holger Krause.
>
>
------------------------------------------------------------
----
> SIEB & MEYER AG
> Software-Entwicklung CNC / Holger Krause
> Auf dem Schmaarkamp 21
> D-21339 Lüneburg
> Germany
> Tel.: +49-(0)4131 203-1505
> Fax.: +49-(0)4131 38562
> E-Mail: Holger.Krausesieb-meyer.de
> http://www.sieb-meyer.de
>
>
>
>
> John Bodnar
> <john.bodnarfree
> scale.com>
An
>
holger.krausesieb-meyer.de
> 10.07.2006 22:14
Kopie
>
>
Thema
> Re: [ColdFire]
Speed of FlexBus on
> the MCF547x
>
>
>
>
>
>
>
>
>
>
>
> Hello Holger,
>
> I think what you're seeing is an artifact of the
MCF547x/8x design.
>
> We built this device on a PowerPC platform, so the
internal bus is
> 64-bit XLB. The V4e ColdFire core, however, only has a
32-bit bus
> interface and cannot issue burst transactions across
the FlexBus, If I
> remember correctly.
>
> Our new devices, the MCF5207/8, MCF532x, and MCF537x
are built on a
> different platform with native ColdFire buses
throughout and support
> core bursts to the FlexBus.
>
> We're working on a new high-end ColdFire device that
also uses the
> FlexBus (it is now our standard external bus interface
module), and you
> may want to check it out when it is introduced in a few
months time.
>
> Best regards,
>
> John Bodnar
> Systems Engineer
> Microcontroller Division
>
> --
>
> holger.krausesieb-meyer.de wrote:
>
>> Hello,
>>
>> we want to connect SRAM to the FlexBus of our
MCF5474 board. So we
>> configured the appropriate chip selects and wrote a
simple routine which
>> accesses this memory. Everything works fines but
the performance seems to
>> be too low: One access is very fast (~25ns) but the
time between two
>> accesses is about 130ns! => this is slower than
the MCF5407.
>> The FBCSn-signal is active for about <20ns -
this is very fast - but the
>> time between the raising edge of FBCSn and the
falling edge of the next
>> FBCS is very long (about > 100ns).The R/W signal
is valid for the whole
>> time so there are no other accesses. We don't
know what the processor is
>> doing in the meantime. Or is it already full speed?
>>
>> We also changed the configuration of the arbiter
but unfortunantely it
>> seems to have no effect.
>>
>> The CLKIN is 66.66MHz, the CS has no waitstates,
IRQs are off and the
>>
> code
>
>> runs in the core SRAM.
>>
>> Has anyone similar experiences or advices for
improving the speed?
>>
>> Best regards,
>>
>> Holger Krause.
>>
>>
------------------------------------------------------------
----
>> SIEB & MEYER AG
>> Software-Entwicklung CNC / Holger Krause
>> Auf dem Schmaarkamp 21
>> D-21339 Lüneburg
>> Germany
>> Tel.: +49-(0)4131 203-1505
>> Fax.: +49-(0)4131 38562
>> E-Mail: Holger.Krausesieb-meyer.de
>> http://www.sieb-meyer.de
>>
>>
------------------------------------------------------------
--------
>> To Subscribe send a message to:
ColdFire-OnLists.Wildrice.com
>> To Unsubscribe send a message to:
ColdFire-OffLists.Wildrice.com
>> For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
>>
>>
>>
>
> --
>
> POPI Classification
> [ ]General Business Information
> [x]Freescale Internal Use
> [ ]Freescale Confidential Proprietary
>
>
>
>
------------------------------------------------------------
--------
> To Subscribe send a message to: ColdFire-OnLists.Wildrice.com
> To Unsubscribe send a message to: ColdFire-OffLists.Wildrice.com
> For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
>
>
--
POPI Classification
[ ]General Business Information
[x]Freescale Internal Use
[ ]Freescale Confidential Proprietary
------------------------------------------------------------
--------
To Subscribe send a message to: ColdFire-OnLists.Wildrice.com
To Unsubscribe send a message to: ColdFire-OffLists.Wildrice.com
For further information, visit: <http://www.Wild
Rice.com/ColdFire/>
|
|
[1-3]
|
|