mpi-1sided Mailing List Archive by subject
Starting: Thu 01 Jun 1995 - 17:31:37 PDT
Ending: Mon 19 May 1997 - 14:14:12 PDT
Messages: 356
- "PUT/ACCEPT" proposal (ascii)
- (no subject)
- -No Subject-
- 1-sided proposals comparison
- 1sided and better post/start
- 1sided chapter comments
- 2-phase win_barrier
- 3rd-party comm in PUT/ACCEPT proposal using custom agents
- 5.6 Semantics and correctness
- a (radical ?) alternative to current chapter 4
- A new version of the chapter is at the ftp server at Wisconsin
- a proposal of Lyndon to replace post/start/complete/wait
- A proposal to change the direction of 1-sided comm
- alt. proposal
- alternate proposal
- another non-operative amendment
- are arguments _really_ passed by reference in fortran?
- CDS1
- chapter 4 -- draft for next meeting (comments/corrections are welcome)
- Chapter 4 -- new version
- chapter 4 postscript
- chapter 5
- chapter on one-sided communication for next meeting
- clarifications
- comments
- comments on 4/8 1-sided version
- Comments on current chapter 4 draft (reply to Lloyd)
- Comments on current draft
- Comments on latest one-sided
- Comments on March 18 draft
- Comparisons and additions
- current chapter version (changes in handlers and few addional
- current proposal
- darft of chapter four
- Detailed proposal for Handlers & implications for MPI2
- draft of chapter 4
- draft of chapter 4 for next meeting -- pls comment
- draft of one sided section for next meeting
- draft, one sided for next meeting
- Enquiries on MPI_Window
- Epoch
- Extended proposal: point-to-point 1-sided
- Fairness versus efficiency
- Handlers
- Hrecv versus MPI_POST_HANDLER
- inconsistency in MPI_WINWOW_IN
- info, hints, commandline in dynamic/io/1sided
- info, hints, commandline in dynamic/io/1sided (fwd)
- introduction to 1-sided chapter
- Invitation for implementation-specific mechanisms
- latest one-sided chapter
- Latest version of P/G/O one-sided alternate proposal
- LINDA and MPI2
- lock/unlock does not synchronize!!!
- Minor comments on MPI-2
- more fun with proposals
- More lock stuff
- MPI Archive Test
- MPI Archive Testing
- MPI handlers need preemption
- MPI-1, volatile, MPI-1sided
- MPI-2 PUT/GET datatypes
- MPI-2 volatile and PUT/GET/HANDLER
- MPI_ACCUMULATE
- MPI_GET & async. producer/consumer & volatile
- MPI_Iconsume_counter Request completion
- MPI_LOCK semantics P.S.
- MPI_NOPRECEDE/MPI_NOSUCCEED Semantics
- MPI_POST_HANDLER()
- MPI_RMA_INIT
- MPI_RWM
- MPI_SWAP -> MPI_PUT
- MPI_Type_send/recv
- mpi_win_Ibarrier -- non-blocking
- mpi_win_test
- mpi_win_test & mpi_Iwin_barrier
- my application model
- new chapter 5
- New one-sided "P/G/O" proposal
- new version
- New version of 1sided chapter (postscript) is on ftp server at
- New version of 1sided chapter (sent the old way, as the server at
- new version of chapter 4
- New version of Chapter 4 draft
- new version of chapter 5
- new version of chapter four
- New version of P/G/O proposal
- New version of the chapter
- nonrecursive locks -- pros and cons
- One more thought on 1sided
- Onesided -- comments & corrections
- OOps -- previous note had wrong file.
- Optimization hints
- P/G/O proposal
- Plea for simplicity
- points of chapter
- polling is less efficient with alternate proposal
- possible race condition (functional bug)
- possible race condition in post-wait-start-complete
- post-start-complete-wait
- postscript chapter 4
- postscript for 1sided chapter is on the ftp server at Wisconsin
- Postscript for P/G/O one-sided proposal
- Problem with multiple IPUTs to same target
- Progress rule vs. MPI_DELIVER
- Proposal for Handlers (postscript)
- Proposal for two processes to "share" a buffer (TeX)
- Proposed amendment
- PUT on an unreliable interconnect
- put/get implementations
- question on threads proposal
- question when multiple windows
- Questions prompted by Chicago
- quick question about 1-sided
- Raja's questions
- RE: 3rd-party comm in PUT/ACCEPT proposal using custom agents
- RE: 5.6 Semantics and correctness
- RE: Plea for simplicity
- RE: Problems with Register Optimization and Caching
- RE: Progress rule vs. MPI_DELIVER
- RE: put/get implementations
- RE: relaxed 1-sided model
- RE: suggestions of Lloyd
- RE: we should wait for 1sided implementations
- Reasons for negative votes on 1-sided at Sept. MPI mtg
- Reduce 1SIDED functionality
- relaxed 1-sided model
- Request Handler semantics
- Request Handlers, please
- semantics of lock/unlock
- Strart, post, wait, complete
- suggestions of Lloyd
- test of mpi-1sided archive
- to lock or not to lock
- Two "count" arguments on MPI_PUT
- Updated "PUT/ACCEPT" proposal (ascii)
- version of 1sided chapter for next meeting
- version of chapter 4 for next meeting
- Volatile & Caching section
- Volatile & Caching section (1.draft)
- we should wait for 1sided implementations
- Welcome to mpi-1sided
- Window counters
- window lock functions
- YASM
Last message date: Mon 19 May 1997 - 14:14:12 PDT
Archived on: Mon May 19 1997 - 14:14:19 PDT
This archive was generated by hypermail 1.02.