March 2010 - Ad-dev - mailman3.cels.anl.gov

[openAD] #200: introduce scope per CFG node
by OpenAD 18 Nov '14

18 Nov '14

#200: introduce scope per CFG node -------------------------+-------------------------------------------------- Reporter: utke | Owner: snarayan Type: enhancement | Status: new Priority: major | Milestone: Component: xaif | Keywords: -------------------------+-------------------------------------------------- for C++, with default scope being the same scope as the top level cfg scope for C/fortran -- Ticket URL: <https://trac.mcs.anl.gov/projects/openAD/ticket/200> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 5

[openAD] #202: insufficient parsing of rewind statements
by OpenAD 09 Sep '10

09 Sep '10

#202: insufficient parsing of rewind statements ------------------------------+--------------------------------------------- Reporter: lyonsam | Owner: lyonsam Type: defect | Status: new Priority: minor | Milestone: Component: SourceProcessing | Keywords: ------------------------------+--------------------------------------------- occuring in centrm/pxlib_read_M.f90 (we were able to get around this for now because we dont really need to transform this file): rewind ['(', 'linpxs', ')'] -- Ticket URL: <https://trac.mcs.anl.gov/projects/openAD/ticket/202> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 4

[openAD] #208: value selector added in error to RHS of active pointer assignment
by OpenAD 17 Apr '10

17 Apr '10

#208: value selector added in error to RHS of active pointer assignment --------------------------+------------------------------------------------- Reporter: lyonsam | Owner: utke Type: defect | Status: new Priority: major | Milestone: Component: OpenADFortTk | Keywords: --------------------------+------------------------------------------------- Example ------------ declarations: XN is active and exists in an external module (centrm/data_centrm_C.f90): {{{ type(active) :: XN(:,:) ALLOCATABLE XN TARGET XN }}} pointer declaration and troublesome statements (in centrm/accel_I.f90): {{{ type(active) :: XNTN(:,:) POINTER XNTN }}} troublesome statement: {{{ XNTN => XN(:,IFTG:)%v }}} should be {{{ XNTN => XN(:,IFTG:) }}} -- Ticket URL: <http://trac.mcs.anl.gov/projects/openAD/ticket/208> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 3

[openAD] #209: assumed-size arrays
by OpenAD 16 Apr '10

16 Apr '10

#209: assumed-size arrays ---------------------------------+------------------------------------------ Reporter: utke | Owner: utke Type: task | Status: new Priority: minor | Milestone: Component: multiple components | Keywords: ---------------------------------+------------------------------------------ An assumed-size array (F77 idiom) is a dummy argument that has a '*' in the dimension specification. The dummy argument will assume the size (but not necessarily the shape or extent) of the actual argument. No lbound/ubound/size info can be extracted from the assumed-size array dimension. Consequently it is impossible to create any temporaries for such arrays. HOWEVER, for many practical uses the actual size of the array will still be known at call time, because it is passed in another dummy argument. In that case the suggested workaround is to replace the "*" in the dimension with the dummy argument that contains the size (or a permitted expression thereof) The alternative in current Fortran are assumed-shape arrays. They do require an interface and consequently matching shape. -- Ticket URL: <http://trac.mcs.anl.gov/projects/openAD/ticket/209> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 1

[openAD] #206: activity analysis fails for pointers
by OpenAD 26 Mar '10

26 Mar '10

#206: activity analysis fails for pointers --------------------------------------+------------------------------------- Reporter: lyonsam | Owner: utke Type: defect | Status: new Priority: major | Milestone: scale1 Component: OpenAnalyis (old branch) | Keywords: --------------------------------------+------------------------------------- very small example: C$openad INDEPENDENT(x) p => x y = x + p C$openad DEPENDENT(y) -- Ticket URL: <https://trac.mcs.anl.gov/projects/openAD/ticket/206> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 2

[openAD] #207: value/deriv component selectors not added for active pointers
by OpenAD 26 Mar '10

26 Mar '10

#207: value/deriv component selectors not added for active pointers --------------------------+------------------------------------------------- Reporter: lyonsam | Owner: utke Type: defect | Status: new Priority: major | Milestone: scale1 Component: OpenADFortTk | Keywords: --------------------------+------------------------------------------------- Manifests in Scale; hard to re-create in a small example without ability to induce an active pointer. example from centrm/accel_I.f90: active array pointer XNTN needs value selector in a number of places. Here is the declaration: {{{ type(active) :: XNTN(:,:) POINTER XNTN }}} -- Ticket URL: <https://trac.mcs.anl.gov/projects/openAD/ticket/207> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 6

Re: [Ad-dev] [openAD] #179: need logic for explicit initialization of derivative components
by OpenAD 24 Mar '10

24 Mar '10

#179: need logic for explicit initialization of derivative components -------------------------+-------------------------------------------------- Reporter: utke | Owner: hlcm Type: enhancement | Status: assigned Priority: major | Milestone: MATWS adjoint Component: xaifBooster | Keywords: -------------------------+-------------------------------------------------- Comment(by utke): more changesets: 1149:8cd158fde9b3 1150:6f09a7237ff6 1151:f7d1ff5cd420 1152:c9f0d81e51a3 1161:679ce59503f7 -- Ticket URL: <https://trac.mcs.anl.gov/projects/openAD/ticket/179#comment:5> OpenAD <http://www.mcs.anl.gov/openad> Bug tracking for OpenAD components

1 0

Re: [Ad-dev] [ADIC] #12: malloc sizeof function calls
by ADIC 22 Mar '10

22 Mar '10

#12: malloc sizeof function calls ------------------------------+--------------------------------------------- Reporter: snarayan | Owner: snarayan Type: defect | Status: new Priority: major | Milestone: Update to ROSE version 0.9.x Component: Canonicalization | Version: 2.0 Keywords: | ------------------------------+--------------------------------------------- Comment(by hyjuh): Also, there is a race problem with ARexx scripts, the MAILSTATUS command, and the feature. Last comment every day - [http://putana003.co.cc] [http://putana006.co.cc] [http://putana008.co.cc] [http://putana026.co.cc] GOOOOOD -- Ticket URL: <http://trac.mcs.anl.gov/projects/ADIC/ticket/12#comment:5> ADIC <http://trac.mcs.anl.gov/projects/ADIC> ADIC: Automatic Differentiation for C and C++

1 0

Re: [Ad-dev] [ADIC] #5: Global Array Inititialization
by ADIC 22 Mar '10

22 Mar '10

#5: Global Array Inititialization ------------------------------+--------------------------------------------- Reporter: snarayan | Owner: snarayan Type: defect | Status: new Priority: major | Milestone: Update to ROSE version 0.9.x Component: Canonicalization | Version: 2.0 Keywords: Array | ------------------------------+--------------------------------------------- Comment(by vengh): Cheap Independents Yuzhno-Sakhalinsk [http://putana004.co.cc] [http://putana007.co.cc] [http://putana009.co.cc] [http://putana013.co.cc] Fucking -- Ticket URL: <http://trac.mcs.anl.gov/projects/ADIC/ticket/5#comment:3> ADIC <http://trac.mcs.anl.gov/projects/ADIC> ADIC: Automatic Differentiation for C and C++

1 0

Re: [Ad-dev] [ADIC] #4: Local Array Init Function Names
by ADIC 22 Mar '10

22 Mar '10

#4: Local Array Init Function Names -------------------------+-------------------------------------------------- Reporter: snarayan | Owner: Sri Hari Krishna Narayanan Type: defect | Status: closed Priority: major | Milestone: Update to ROSE version 0.9.x Component: component1 | Version: 2.0 Resolution: fixed | Keywords: Array -------------------------+-------------------------------------------------- Comment(by NN7Y): As far as I know, in our department no vacancies. In what department you're going to hire? [http://putana014.co.cc] [http://putana020.co.cc] [http://putana010.co.cc] [http://putana012.co.cc] Fucking -- Ticket URL: <http://trac.mcs.anl.gov/projects/ADIC/ticket/4#comment:3> ADIC <http://trac.mcs.anl.gov/projects/ADIC> ADIC: Automatic Differentiation for C and C++

1 0