Hi Martin,
I finally found the solution. I the method *set_align(int ), * to tell the
scheduler output is multiple of a number but this is different than
set_output_multiple(int ). The set_align will decrease the amount of buffer
if the output number of items is less than the actual number of written
output until the output buffer aligned with the number of items will
written on it!

This is interesting.

best,


On Wed, May 14, 2014 at 1:59 PM, Mostafa Alizadeh <m.alizade...@gmail.com>wrote:

> Hi Martin,
> Yes you're right, 75000 items isn't so large! But after passing these
> items through an encoder, the number of items becomes about 227000!! That's
> a big number!
>
> However, I skewed over the problem and till now I've reached to this point
> that I may use 3 or 4 inputs-ouputs ports per block (instead of actually
> one input-ouput port) and divide the 227000 items into 4 parallel streams
> and read all of them through 3/4 different input ports and put them out to
> the 3/4 output ports ! That seems work but I'm not sure that's the best
> one.
>
> If you have any idea I'll be appreciate you so much.
>
> best,
>
>
>
> On Wed, May 14, 2014 at 1:04 PM, Martin Braun <martin.br...@ettus.com>wrote:
>
>> On 13.05.2014 19:27, Mostafa Alizadeh wrote:
>>
>>> Hi,
>>>
>>> I recently encountered a problem with the large number of items. I wanna
>>> pass about 75000 items from one block to the another. I thought that I
>>> could do this with setting the min of output items in the constructor of
>>> the block but I got the following error by runtime:
>>>
>>> thread[thread-per-block[1]: <block crc (1)>]: Buffer too small for
>>> min_noutput_items
>>>
>>> I searched around but I couldn't find a cogent response. please help me!
>>>
>>
>> You could use your own version of tagged streams to indicate
>> boundaries... regular tagged stream blocks also suffer from buffer
>> limitations.
>>
>> That said, 75000 items doesn't seem all that large. What's your item
>> size, is it sizeof(gr_complex)? Did you set your kernel.shmmax = 2147483648?
>>
>> M
>>
>>
>> _______________________________________________
>> Discuss-gnuradio mailing list
>> Discuss-gnuradio@gnu.org
>> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>>
>
>
_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Reply via email to