Technical discussions related to Analog Devices DSPs (including Blackfin, TigerSHARC, SHARC and ADSP-21xx DSPs).
Hi,
I am using Blackfin BF534 for my application...
I am using asynchronous memory intrface pins to transfer
data to the peripheral device like DAC,LCD...
Questions:
1. Can I Use memory DMA for transferring the data through the Asynchronous memory
interface...For example, I want to transfer the single data(16 bit) to LCD....Can I use
Memory DMA(MDMA) for the transfer........
2. Can I use the asynchronous memory interface pins for both LCD and DAC
3.Can I Control the rate of data transfer using asynchronous memory interface....
4.Are there any other options by which I can transfer data through asynchronous memory
inetrface pins.....
5.The blackfin hardware manual BF534(BF537) :-----It just describes the timing diagram, rather
it does not suggest from where does the data come(data source location).....I mean to say, How
can I define the data source location(address) and dat destination address
6. Suggest me sample code for asynchronous memory data transfers.....
Kind Regards,
Nithin
------------------------------------
Hi
I’ve a problem with enabling SIMD on ADSP 21363 ( VDSP++ 5.0, no update).
I’m trying to compute a very simple fir filter and when compare result in C
and in ASM there is a problem:
Fir Filter written in C:
for (i=0; i<ntap; ++i)
{
yn = yn + w[i] * x[N-1-i];
}
In ASM when written in SISD mode, results are identical but when enabling
SIMD mode, result are different sometimes. Is there somebody who has
encountered same problems?
Thanks in advance
Hi all
=20
I need to implement fir filter but I want to compute output value each time
a new sample (input) is available. In C language my routine is :=20
=20
for (i=3D0; i<ntap; ++i)
{
yn =3D yn + w[i] * x[N-1-i];=20=20
}
=20
New inputs are store in X[n] using =93modulo=94 :=20
X[(index++)%N]=3Dx_new;
=20
I want to optimize performance and enable SIMD, but I don=92t know how to
proceed. I find fir_vec routine in ASM but I va a problem with parameter
passing :=20
float *fir_vec( const float input[],
float output[],
const float pm coefficients[],
float state[],
int samples,
int taps );
=20
Is there anybody who used it, do you have a small code example in order for
me to understand. Fir computation is done each new sample.
So any suggestions will be appreciated.
------------------------------------