Forums

C6713 -> C6727

Started by Bernhard 'Gustl' Bauer June 13, 2005
Hi,

has anybody tried the C6727 already? I've a project on a C6713. I want
increase the speed of this project. I've somewhere read that C6727
should be 20% faster than C6713 (same CPU frequency). Can someone
confirm this?

CCS 3.1 should be capable of generating the C6727 specific code. I've
CCS 2.2 and before I spend a lot of money on a new compiler and a new HW
I want to simulate my project, to be sure that there is an increase in
speed. There was once a 90 day trial version of CCS. I can't find it.
Any idea where to look?

TIA

Gustl



Bhooshan Iyer wrote:

> Gustl-- >>has anybody tried the C6727 already? > There is a C6727? Gee,this is news to me!
>

There is!
http://focus.ti.com/docs/prod/folders/print/tms320c6727.html

>
>>CCS 3.1 should be capable of generating the C6727 specific code. > I have'nt seen a C7627 target in my CCS, but I havent looked either.
>

Could you please have look? What CCS revision do you have?

>
>>There was once a 90 day trial version of CCS. I can't find it.
>>Any idea where to look? > Try this link:
>
> https://focus-webapps.ti.com/general/docs/regapp/showregappresultsaction.tsp?action=showResultsPage&regAppId4
>
> or
>
> https://focus-webapps.ti.com/general/docs/regapp/serveresourcesaction.tsp?regAppId4&resourceId10&actionPerformed=serveResource
>
> Although it is only for CCS3.0, I havent seen CCS3.1 FET yet.

I'm affraid C6727 is supported by CCS 3.1 only.


hello Gustl,
 
Don't get too carried away looking for the "solution to all of your problems".  keep in mind that the 6727 is heavily biased towards audio processing.  You will pickup some improved code density and *could* pickup some performance.  i have seen some audio apps pick up significant improvement [it looked like 15-20% in some cases, but they were my customer's apps and i wan't benchmarking].  Hopefully going from 32 to 64 cpu registers will help performance...
 
To get CCS 3.1 to support c6727, I had to add new simulator, emulation drivers, BIOS, and compiler.  TI should have a patch for this by now...
 
BTW, do you need to simulate only an algorhythm, do you need to simulate I/O also??
 
mikedunn

Bernhard 'Gustl' Bauer <g...@quantec.de> wrote:
Hi,

has anybody tried the C6727 already? I've a project on a C6713. I want
increase the speed of this project. I've somewhere read that C6727
should be 20% faster than C6713 (same CPU frequency). Can someone
confirm this?

CCS 3.1 should be capable of generating the C6727 specific code. I've
CCS 2.2 and before I spend a lot of money on a new compiler and a new HW
I want to simulate my project, to be sure that there is an increase in
speed. There was once a 90 day trial version of CCS. I can't find it.
Any idea where to look?

TIA

Gustl

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/c6x/

<*> To unsubscribe from this group, send an email to:
c...@yahoogroups.com

<*



On 6/14/05, Mike Dunn <mike-dunn@mike...> wrote:
> hello Gustl,
>
> Don't get too carried away looking for the "solution to all of your
> problems". keep in mind that the 6727 is heavily biased towards audio
> processing. You will pickup some improved code density and *could* pickup
> some performance. i have seen some audio apps pick up significant
> improvement [it looked like 15-20% in some cases, but they were my
> customer's apps and i wan't benchmarking]. Hopefully going from 32 to 64
> cpu registers will help performance...
>
> To get CCS 3.1 to support c6727, I had to add new simulator, emulation
> drivers, BIOS, and compiler. TI should have a patch for this by now...
>
> BTW, do you need to simulate only an algorhythm, do you need to simulate I/O
> also??
>
> mikedunn

Mike, you are right about CCS3.1 organically not supporting C6727.I
checked up my CCS3.1 installation and I could not find a C6727
simulator.

It has 64 registers? Thats like an imitation of C64x in floating point then?

--Bhooshan


Hello Mike,

the 6727 isn't at all ideal for my project, which is audio processing. I
would miss a lot from 6713 (Timer, McBsp ...). The first step would be
to simulate only the algorithm. But I have to go over the memory I/O to.
I need to read about 19 Mwords/sec and write about 16 Mwords/sec in
blocks of 8 words.

I've asked TI about an evaluation CCS which supports 6727. I got no
response so far.

Gustl

Mike Dunn wrote:
> hello Gustl,
>
> Don't get too carried away looking for the "solution to all of your
> problems". keep in mind that the 6727 is heavily biased towards audio
> processing. You will pickup some improved code density and *could*
> pickup some performance. i have seen some audio apps pick up
> significant improvement [it looked like 15-20% in some cases, but they
> were my customer's apps and i wan't benchmarking]. Hopefully going from
> 32 to 64 cpu registers will help performance...
>
> To get CCS 3.1 to support c6727, I had to add new simulator, emulation
> drivers, BIOS, and compiler. TI should have a patch for this by now...
>
> BTW, do you need to simulate only an algorhythm, do you need to simulate
> I/O also??
>
> mikedunn
>
> Bernhard 'Gustl' Bauer <gustl@gust...> wrote:
>
> Hi,
>
> has anybody tried the C6727 already? I've a project on a C6713. I want
> increase the speed of this project. I've somewhere read that C6727
> should be 20% faster than C6713 (same CPU frequency). Can someone
> confirm this?
>
> CCS 3.1 should be capable of generating the C6727 specific code. I've
> CCS 2.2 and before I spend a lot of money on a new compiler and a
> new HW
> I want to simulate my project, to be sure that there is an increase in
> speed. There was once a 90 day trial version of CCS. I can't find it.
> Any idea where to look?
>
> TIA
>
> Gustl
>