From: RCIngham on
>On 12 Apr., 18:59, "RCIngham"
><robert.ingham(a)n_o_s_p_a_m.n_o_s_p_a_m.gmail.com> wrote:
>> >On 4/12/2010 5:14 PM, RCIngham wrote:
>> >> Greetings,
>>
>> >> The synthesis report on my design tells me that I have caused 3
latche=
>s
>> to
>> >> be created: 1 LDC and 2 LDP. I didn't intend to create any...
>>
>> >> Rather frustratingly, the reports aren't telling me which signals
are
>> >> associated with these latches, so that I can fix my code! My
colleague=
>s
>> >> can't remember what the trick is. Please can someone enlighten me?
>>
>> >> Thanks in advance,
>> >> R.
>>
>> >Search through the EDIF file?
>>
>> >HTH, Syms.
>>
>> What EDIF file? Or perhaps, which EDIF file?
>> And don't suggest the post-Synthesis gate-level netlist, either. Been
>> through that and failed to find and LD* components...
>>
>> --------------------------------------- =A0 =A0 =A0 =A0
>> Posted throughhttp://www.FPGARelated.com
>
>Hi,
>the synthesis report tells you at least in which module of your design
>the latch is found.
>The reasons are mostly cases without when others and ifs withut else
>in combinatorical processes.
>(or conditional assignments outside of processes)
>
>These should be easy to identify.
>You shouldn't have too many combinatorical processes anyway.
>
>Have a nice synthesis
> Eilert
>

I've been through the .SYR file several times, and there is no such report.
Is there a missing command switch in the script?

Perhaps they are in the .NGC file relating to the MicroBlaze processor. As
this is a reused item, why were they not spotted before? I'll try to
synthesize some of the blocks separately...

Grrrr!




---------------------------------------
Posted through http://www.FPGARelated.com
From: whygee on
RCIngham wrote:
> Perhaps they are in the .NGC file relating to the MicroBlaze processor. As
> this is a reused item, why were they not spotted before? I'll try to
> synthesize some of the blocks separately...

I think that I have read somewhere (where ? from another thread on
latch inference ?) that the Microblaze core uses a few latches.
So this could be a valid theory.

You could test it by replacing your MB core with a dumb one,
and see if the latches are still there.

> Grrrr!
don't worry... it wouldn't change anything.

yg
--
http://ygdes.com / http://yasep.org
From: RCIngham on
>RCIngham wrote:
>> Perhaps they are in the .NGC file relating to the MicroBlaze processor.
As
>> this is a reused item, why were they not spotted before? I'll try to
>> synthesize some of the blocks separately...
>
>I think that I have read somewhere (where ? from another thread on
>latch inference ?) that the Microblaze core uses a few latches.
>So this could be a valid theory.
>
>You could test it by replacing your MB core with a dumb one,
>and see if the latches are still there.
>

A separate synthesis of the MB core yields the 3 latches, as does a
synthesis of one of the other designs using the same MB.

Probably explains why no useful warnings in the Synthesis Report.

"Onward & Upward"


---------------------------------------
Posted through http://www.FPGARelated.com
From: Martin Thompson on
"RCIngham" <robert.ingham(a)n_o_s_p_a_m.n_o_s_p_a_m.gmail.com> writes:

> Greetings,
>
> The synthesis report on my design tells me that I have caused 3 latches to
> be created: 1 LDC and 2 LDP. I didn't intend to create any...
>
> Rather frustratingly, the reports aren't telling me which signals are
> associated with these latches, so that I can fix my code! My colleagues
> can't remember what the trick is. Please can someone enlighten me?

If you haven't found them yet (or for anyone later on finding this
thread), you could try our (free) tool which will narrow things down to
which module the latches are in.

http://www.conekt.net/fpgaoptim.html

It's how I first noticed that Microblaze has a few latches in it :)

Cheers,
Martin

--
martin.j.thompson(a)trw.com
TRW Conekt - Consultancy in Engineering, Knowledge and Technology
http://www.conekt.net/electronics.html
From: whygee on
RCIngham wrote:
> A separate synthesis of the MB core yields the 3 latches, as does a
> synthesis of one of the other designs using the same MB.
>
> Probably explains why no useful warnings in the Synthesis Report.

I'm happy that I have helped you !

yg
--
http://ygdes.com / http://yasep.org
First  |  Prev  | 
Pages: 1 2
Prev: MPEG Reading material
Next: Virtex-5 FPGA PCIe card