Dear Prof. Roberto,
I am working with Parma Pplyhedra Library. I realize that PPL does not wok well with mutilthreading and it leads to segmentation fault. I read in the ChangeLog and found that you has discussed about this:
"2015-11-26 Enea Zaffanella zaffanella@cs.unipr.it
* src/Temp_defs.hh, src/Temp_inlines.hh, src/Temp_templates.hh: Avoid a memory leak due to the global list of dirty temporaries. The memory leak was not causing a real problem, in that the list of dirty temporaries is meant to survive up to the end of the process. However, it could cause a problem if/when the library is extended to become multi-thread safe, since in such a scenario it is meaningful to have a list of dirty temporaries for each thread (to be released on thread exit)."
Has you fixed this problem. Could you give me a suggestion to solve this problem (it could be temporary solution) please?
Best regards,
The Anh
Hello.
The library is not thread safe (yet), so such an error is to be expected; we are going to add thread safety in an experimental branch. Your multi-threaded application can still use the library, as long as only one thread is used to perform all PPL-related computations.
Enea.
On 08/17/2016 10:53 AM, the anh pham wrote:
Dear Prof. Roberto,
I am working with Parma Pplyhedra Library. I realize that PPL does not wok well with mutilthreading and it leads to segmentation fault. I read in the ChangeLog and found that you has discussed about this:
"2015-11-26 Enea Zaffanella <zaffanella@cs.unipr.it mailto:zaffanella@cs.unipr.it>
- src/Temp_defs.hh, src/Temp_inlines.hh, src/Temp_templates.hh: Avoid
a memory leak due to the global list of dirty temporaries. The memory leak was not causing a real problem, in that the list of dirty temporaries is meant to survive up to the end of the process. However, it could cause a problem if/when the library is extended to become multi-thread safe, since in such a scenario it is meaningful to have a list of dirty temporaries for each thread (to be released on thread exit)." Has you fixed this problem. Could you give me a suggestion to solve this problem (it could be temporary solution) please? Best regards, The Anh
PPL-devel mailing list PPL-devel@cs.unipr.it http://www.cs.unipr.it/mailman/listinfo/ppl-devel
participants (2)
-
Enea Zaffanella -
the anh pham