discuss-gnuradio
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: How can you force the Scheduler to deliver appropriate vector size i


From: Ivan Iudice
Subject: Re: How can you force the Scheduler to deliver appropriate vector size input
Date: Wed, 26 Oct 2022 08:30:17 +0200

Hello George,
the input vector size is determined (more or less) by the forecast method.
If the size is greater than you need, this is not a problem, you need to 
consume only the number of items you need. You will find the unconsumed items 
in the next buffer.
The problem arises if you need a minimum output vector size, but this is 
another story.
Regards.

Ivan

> Il giorno 26 ott 2022, alle ore 04:23, George Edwards 
> <gedwards.eng@gmail.com> ha scritto:
> 
> 
> Hello GNURadio Community, 
> 
> I designed an OOT block to accept vectors of size 448 samples. In the 
> flowgraph, my block is preceded a Stream-to-Vector block with the vector size 
> set to 448 samples. My OOT block expects vector data that are of length 448 
> or multiple of 448 samples for signal processing. I dumped the output of my 
> block to a file and noticed that the first 1000 values were correct and the 
> rest garbage. I put a print statement in the code to see the length of the 
> input samples that the Scheduler provided on each iteration. I found out that 
> the inputs were 496 and multiples of 496 (obviously not integer multiple of 
> my vector length of 448 samples). Is there a way to force the Scheduler to 
> deliver input data samples that are the exactly 448 and multiple of 448 
> samples (which I thought I would be getting by setting the vector size 
> parameter to 448)? 
> 
> I am aware that the alternative may be to pad the data entering the 
> Stream-to-Vector block which precedes my OOT Block with 48 dummy samples and 
> in my signal processing remove them.
> 
> I will appreciate any response that leads to me keeping the vector size at 
> 448 and having the Scheduler provide input data that are of size 448 or 
> integer multiples of 448. 
> 
> Regards,
> George    




reply via email to

[Prev in Thread] Current Thread [Next in Thread]