RTDX latency

Hi everybody.

I could finally restructure my code so that both data and code fits in L1 cache.
My software pipeline is working perfectly. Thank you all for your help last
time.

But now i'm figthing against RTDX. Could someone tell me what's the
aprox. latency for each message sent from a VB host app. to the target?

Thanks.

Charles
	

Hello Charles,

--- carlosmarchiaro@carl... wrote:

> Hi everybody.
> 
> I could finally restructure my code so that both
> data and code fits in L1 cache. My software pipeline
> is working perfectly. Thank you all for your help
> last time.
> 
> But now i'm figthing against RTDX. Could someone
> tell me what's the aprox. latency for each message
> sent from a VB host app. to the target?
You have considerable variables in your question.  In
my opinion, they are something like the list below.
1. Windows [like when you try to do something and
Windows just 'pauses' for a brief time].
2. Target activity - how much 'work' the target is
performing
3. CCS activity - the more data that you display, the
greater the 'update display overhead' will be.
Other factors that are 'relatively constant', but
affect speed are:
1. PC speed
2. VB performance
3. Hardware emulator used - If you are sending small
amounts of data and using a USB emulator, your
effective throughput rate will be reduced.  If you are
using an XDS560 emulator [vs. XDS510 class], your
performance will be greater.
4. Type of RTDX [JTAG vs. HS-RTDX] - HS-RTDX [XDS560
only - if your HW supports it] is significantly faster
5. Target type and speed - c620x, c671x, c64x devices
have different RTDX 'performance profiles'.

For your setup, you can use some of the CCS RTDX
examples to characterize your particular environment.

I realize that I didn't give you a specific answer,
but maybe you now have enough info to help yourself.

,mikedunn
> 
> Thanks.
> 
> Charles
> 
> 
> 
> 
> 
> 
> 
>     c6x-unsubscribe@c6x-...
> 
>  
> 
> 
> 
>
	

Charles,

--- Carlos Marchiaro <carlosmarchiaro@carl...>
wrote:

> Thanks Mike. Now it's more clear to me. I
disabled
> the messages log and also
> i'm using less but bigger ammounts of data, and it's
> working better. The
> target is a 6713DSK, it doesn't support HS-RTDX,
> does it?
Yes it does.  HS-RTDX is supported by c671x/c621x [not
c670x/620x or 672x] and c64x devices [except for the
obsolete version 1.0 devices - all DSKs have 1.1 or
newer devices].  Of course you will need an XDS560
emulator.

,mikedunn
> 
> Charles
> 
> 2006/3/8, Mike Dunn <mike-dunn@mike...>:
> >
> > Hello Charles,
> >
> > --- carlosmarchiaro@carl... wrote:
> >
> > > Hi everybody.
> > >
> > > I could finally restructure my code so that both
> > > data and code fits in L1 cache. My software
> pipeline
> > > is working perfectly. Thank you all for your
> help
> > > last time.
> > >
> > > But now i'm figthing against RTDX. Could someone
> > > tell me what's the aprox. latency for each
> message
> > > sent from a VB host app. to the target?
> > You have considerable variables in your question. 
> In
> > my opinion, they are something like the list
> below.
> > 1. Windows [like when you try to do something and
> > Windows just 'pauses' for a brief time].
> > 2. Target activity - how much 'work' the target is
> > performing
> > 3. CCS activity - the more data that you display,
> the
> > greater the 'update display overhead' will be.
> > Other factors that are 'relatively constant', but
> > affect speed are:
> > 1. PC speed
> > 2. VB performance
> > 3. Hardware emulator used - If you are sending
> small
> > amounts of data and using a USB emulator, your
> > effective throughput rate will be reduced.  If you
> are
> > using an XDS560 emulator [vs. XDS510 class], your
> > performance will be greater.
> > 4. Type of RTDX [JTAG vs. HS-RTDX] - HS-RTDX
> [XDS560
> > only - if your HW supports it] is significantly
> faster
> > 5. Target type and speed - c620x, c671x, c64x
> devices
> > have different RTDX 'performance profiles'.
> >
> > For your setup, you can use some of the CCS RTDX
> > examples to characterize your particular
> environment.
> >
> > I realize that I didn't give you a specific
> answer,
> > but maybe you now have enough info to help
> yourself.
> >
> > ,mikedunn
> > >
> > > Thanks.
> > >
> > > Charles
> > >
> > >
> > >
> > >
> > > 
> > >
> > >
> > > to:
> > >     c6x-unsubscribe@c6x-...
> > >
> > >
> > >
> > >
> > >
> > >
> >
> >
>
	

> Thanks Mike. Now it's more clear to me. I disabled the messages log and
> also i'm using less but bigger ammounts of data, and it's working better.

That Mike can list those factors off the top of his head is amazing.

> The target is a 6713DSK, it doesn't support HS-RTDX, does it?

C6713 DSK support high-speed RTDX?  No way.  You can get a maximum of about 30k to 40 kbps across the JTAG controller-USB interface on that board.

-Jeff
 

2006/3/8, Mike Dunn <m...@sbcglobal.net>:

Hello Charles,

--- c...@gmail.com wrote:

> Hi everybody.
>
> I could finally restructure my code so that both
> data and code fits in L1 cache. My software pipeline
> is working perfectly. Thank you all for your help
> last time.
>
> But now i'm figthing against RTDX. Could someone
> tell me what's the aprox. latency for each message
> sent from a VB host app. to the target?
You have considerable variables in your question.  In
my opinion, they are something like the list below.
1. Windows [like when you try to do something and
Windows just 'pauses' for a brief time].
2. Target activity - how much 'work' the target is
performing
3. CCS activity - the more data that you display, the
greater the 'update display overhead' will be.
Other factors that are 'relatively constant', but
affect speed are:
1. PC speed
2. VB performance
3. Hardware emulator used - If you are sending small
amounts of data and using a USB emulator, your
effective throughput rate will be reduced.  If you are
using an XDS560 emulator [vs. XDS510 class], your
performance will be greater.
4. Type of RTDX [JTAG vs. HS-RTDX] - HS-RTDX [XDS560
only - if your HW supports it] is significantly faster
5. Target type and speed - c620x, c671x, c64x devices
have different RTDX 'performance profiles'.

For your setup, you can use some of the CCS RTDX
examples to characterize your particular environment.

I realize that I didn't give you a specific answer,
but maybe you now have enough info to help yourself.

,mikedunn
>
> Thanks.
>
> Charles
>

Charles,

--- Jeff Brower <jbrower@jbro...> wrote:

> Carlos-
> 
> > Thanks Mike. Now it's more clear to me. I disabled
> the messages log and
> > also i'm using less but bigger ammounts of data,
> and it's working better.
> 
> That Mike can list those factors off the top of his
> head is amazing.
> 
> > The target is a 6713DSK, it doesn't support
> HS-RTDX, does it?
> 
> C6713 DSK support high-speed RTDX?  No way.  You can
> get a maximum of about 30k to 40
> kbps across the JTAG controller-USB interface on
> that board.
Let me clarify my previous response [at the same time
as Jeff's].  I answered the question that was asked,
but probably not the one that was 'thought'.  The 6713
DSK target board does support HS-RTDX when connected
to an external emulator.  As Jeff mentioned, it does
not support HS-RTDX when connected through the
built-in USB emulator [it is an XDS510 class
emulator].

mikedunn
> 
> -Jeff
> 
> 
> > 2006/3/8, Mike Dunn <mike-dunn@mike...>:
> >
> >      Hello Charles,
> >
> >      --- carlosmarchiaro@carl... wrote:
> >
> >      > Hi everybody.
> >      >
> >      > I could finally restructure my code so that
> both
> >      > data and code fits in L1 cache. My software
> pipeline
> >      > is working perfectly. Thank you all for
> your help
> >      > last time.
> >      >
> >      > But now i'm figthing against RTDX. Could
> someone
> >      > tell me what's the aprox. latency for each
> message
> >      > sent from a VB host app. to the target?
> >      You have considerable variables in your
> question.  In
> >      my opinion, they are something like the list
> below.
> >      1. Windows [like when you try to do something
> and
> >      Windows just 'pauses' for a brief time].
> >      2. Target activity - how much 'work' the
> target is
> >      performing
> >      3. CCS activity - the more data that you
> display, the
> >      greater the 'update display overhead' will
> be.
> >      Other factors that are 'relatively constant',
> but
> >      affect speed are:
> >      1. PC speed
> >      2. VB performance
> >      3. Hardware emulator used - If you are
> sending small
> >      amounts of data and using a USB emulator,
> your
> >      effective throughput rate will be reduced. 
> If you are
> >      using an XDS560 emulator [vs. XDS510 class],
> your
> >      performance will be greater.
> >      4. Type of RTDX [JTAG vs. HS-RTDX] - HS-RTDX
> [XDS560
> >      only - if your HW supports it] is
> significantly faster
> >      5. Target type and speed - c620x, c671x, c64x
> devices
> >      have different RTDX 'performance profiles'.
> >
> >      For your setup, you can use some of the CCS
> RTDX
> >      examples to characterize your particular
> environment.
> >
> >      I realize that I didn't give you a specific
> answer,
> >      but maybe you now have enough info to help
> yourself.
> >
> >      ,mikedunn
> >      >
> >      > Thanks.
> >      >
> >      > Charles
> >      >
> >
>