16 citations · 57 across the 26 of their papers we have counts for
32 papers
Effective MPI: User-defined Datatypes and Cartesian Communicators for Zero-copy All-to-all Communication in Multidimensional Tori
Jesper Larsson Träff
We present and show how to implement a non-trivial all-to-all communication algorithm for arbitrary -dimensional tori effectively in MPI. Given a factorization of the number of…
Communication Round and Computation Efficient Exclusive Prefix-Sums Algorithms (for MPI_Exscan)
Jesper Larsson Träff
Parallel scan primitives compute element-wise inclusive or exclusive prefix sums of input vectors contributed by consecutively ranked processors under an associative, binary op…
Optimal, Non-pipelined Reduce-scatter and Allreduce Algorithms
Jesper Larsson Träff
The reduce-scatter collective operation in which processors in a network of processors collectively reduce input vectors into a result vector that is partitioned over the p…
Optimal Broadcast Schedules in Logarithmic Time with Applications to Broadcast, All-Broadcast, Reduction and All-Reduction
Jesper Larsson Träff
We give optimally fast time (per processor) algorithms for computing round-optimal broadcast schedules for message-passing parallel computing systems. This affirmativel…
Lectures on Parallel Computing
Jesper Larsson Träff
These lecture notes are designed to accompany an imaginary, virtual, undergraduate, one or two semester course on fundamentals of Parallel Computing as well as to serve as backgrou…
Round-optimal -Block Broadcast Schedules in Logarithmic Time
Jesper Larsson Träff
We give optimally fast time (per processor) algorithms for computing round-optimal broadcast schedules for message-passing parallel computing systems. This affirmativel…