Re: [Gems-users] LogTM: assertion failure in hitCallBack()

[Jayaram Bobba <bobba@xxxxxxxxxxx>]

In the released version of the code, if a log store misses in the cache,
the corresponding transactional store does not stall. When the miss is
serviced, you could run into these assert failures. Commenting them out
should be a safe temporary option...

Jayaram

Zeshan Chishti wrote:
> Jayaram and Kevin! Thanks for your replies.
>
> I am using Gems version 1.3. I have also tried using  version 1.2 but 
> have encountered the same error.
>   
> > Does the assertion fail in the vicinity of a transaction abort? Could 
> > you comment out the assertion and report the
> > result of the execution?
> >     
> The assertion fails during or after the transaction commit. Transaction 
> abort doesn't seem to be a pre-condition as I have encountered this 
> error even when running a single large transaction.
> I tried to comment out the assertion, but then received another 
> assertion failure:
> failed assertion 'data.getAddress() == m_lingering_request_address' at 
> fn void SimicsProcessor::hitCallback(SubBlock&) in 
> simics/SimicsProcessor.C:249
> When I commented out the above assertion as well, the simulation 
> completed successfully. However, occasionally, I encounter the following 
> assertion failure upon running large number of transactions concurrently:
> failed assertion 'isReady(request)' at fn void 
> Sequencer::makeRequest(const CacheMsg&) in system/Sequencer.C:487
> Thanks for your help
>
> -zeshan
>
>
> _______________________________________________
> Gems-users mailing list
> Gems-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your search.
>