FPGA Intellectual Property
PCI Express*, Networking and Connectivity, Memory Interfaces, DSP IP, and Video IP
6462 ディスカッション

Avalon ST to FFT megafunction variable speed

Altera_Forum
名誉コントリビューター II
1,500件の閲覧回数

Hi everyone, 

 

I have an implemenation of the FFT megacore, currently under the Avalon ST interface 

In front of this I have a fifo 

the data coming in to the system is relatively slow, possibly irregular time periods  

Is it safe to use clock gating on the FFT/Avalon interface as a means to hold it back (even though this could be mid packet) while collecting more data ? 

Or is it better to de-assert the valid? or both? 

 

Hope you can point me the right way, 

 

Regards 

 

Pete B
0 件の賞賛
4 返答(返信)
Altera_Forum
名誉コントリビューター II
808件の閲覧回数

clock gating isn't really safe in FPGAs. It is better to use the valid signal and deassert it when you don't have any new data to give to the Avalon sink.

Altera_Forum
名誉コントリビューター II
808件の閲覧回数

 

--- Quote Start ---  

clock gating isn't really safe in FPGAs. It is better to use the valid signal and deassert it when you don't have any new data to give to the Avalon sink. 

--- Quote End ---  

 

Ok, thanks, 

 

so deasserting Valid mid packet doesnt do any harm though? 

 

Thanks again
Altera_Forum
名誉コントリビューター II
808件の閲覧回数

no it shouldn't. The sink will just wait until there is valid data again. 

Be sure to check the ready signal too. A data sample is only transferred from a source to a sink when both the valid and ready signals are asserted.
Altera_Forum
名誉コントリビューター II
808件の閲覧回数

That's great, thanks so much for that, very helpful and clear

返信