Augustsson L and Johnsson T, "Parallel graph reduction with the ! ;G?-machine", Proc IFIP Conference on Functional Programming Languages and Computer Architecture, London, (September 1989). Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Processing Transactions On Grip, A Parallel Graph Reducer - Akerholt, Hammond, Jones, .. (1993) (1 citation) (Correct)
....These include: 1) One PE, with varying numbers of virtual disks to measure the potential concurrency gain for a sequential version exploiting fwif. 2) Other architectures, for example a shared memory machine such as the Sequent Symmetry, using the Chalmers ;G machine compiler [AJ89]. 3) Eager pre fetch of normal form packets from the IMUs. 4) Local sparking [HP92] to reduce communication overhead. 5) An improved implementation of fwif with indirection chaining via the IMUs. ....
Augustsson L and Johnsson T, "Parallel graph reduction with the ! ;G?-machine", Proc IFIP Conference on Functional Programming Languages and Computer Architecture, London, (September 1989).
GUM: a portable parallel implementation of Haskell - Trinder, Hammond, Mattson.. (1996) (39 citations) (Correct)
....with a single shared heap, and a two level task queue, which aimed to reduce memory contention. To avoid duplicating work, each thunk was locked when it was entered, an expensive operation on a sharedmemory machine. Relative speedups of 6 to 10 were achieved with 12 processors. The ; G Machine [3] was based on the sequential Chalmers G Machine compiler, and ran on a 16 processor Sequent Symmetry. There was no stack, but instead thunks were built with enough space to hold all necessary arguments plus some local workspace for temporary variables. As with Buckwheat, a single shared heap was ....
Augustsson L, and Johnsson T, "Parallel Graph Reduction with the !; G?-Machine", Proc. FPCA '89, London, UK, (1989), pp. 202--213.
GUM: a portable parallel implementation of Haskell.. - Trinder, Hammond.. (1996) (39 citations) (Correct)
....To avoid duplicating work, each thunk was locked when it was entered. This is a potential weakness of the design: locking can be costly, and consumes significant shared memory bandwidth, generally unnecessarily. Relative speedups of 6 to 10 were achieved with 12 processors. The ; G Machine [3] was based on the sequential Chalmers G Machine compiler, and ran on a 16 processor Sequent Symmetry. There was no stack, but instead thunks were built with enough space to hold all necessary arguments plus some local workspace for temporary variables. As with Buckwheat, a single shared heap was ....
Augustsson L, and Johnsson T, "Parallel Graph Reduction with the ! ; G ?-Machine", Proc. FPCA '89, London, UK, (1989), pp. 202--213.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC