Manuel Pöter

dblp:154/6339 · DBLP profile ↗
← Back
3ranked-venue papers
2as first author
1since 2021 · last 2021
—ORCID · none

Domains — the database's venue-derived domains; a paper can count in several

Systems, architecture and hardware · 3 · 2 first-author · 1 since 2021

Expertise — from the expertise taxonomy: the topics of the expert's papers under the CCF categories. A weight counts papers with recency: 1 for a paper about the topic, 0.3 when the topic is its context, halved every five years.

Software engineering, system software, and programming languages
2 papers
Concurrent programming · 96% Operating systems · 4%

Topics — the 7 heaviest of 7, each with the papers that count most for it

TopicWeightPapersLastEvidence papers
Concurrent programming
concurrent data structures
0.512021
A more pragmatic implementation of the lock-free, ordered, linked list · PPoPP 2021
Concurrent programming › concurrent data structures
linked list
0.512021
A more pragmatic implementation of the lock-free, ordered, linked list · PPoPP 2021
Concurrent programming › non-blocking algorithms
lock-freedom
0.512021
A more pragmatic implementation of the lock-free, ordered, linked list · PPoPP 2021
Concurrent programming › synchronization
non-blocking synchronization
0.512021
A more pragmatic implementation of the lock-free, ordered, linked list · PPoPP 2021
Concurrent programming
memory reclamation
0.312018
Stamp-it, amortized constant-time memory reclamation in comparison to five other schemes · PPoPP 2018
Concurrent programming › atomicity
linearizability
0.112021
A more pragmatic implementation of the lock-free, ordered, linked list · PPoPP 2021
Operating systems › resource management
memory management
0.112018
Stamp-it, amortized constant-time memory reclamation in comparison to five other schemes · PPoPP 2018

Methods — techniques the papers use, named apart from their topics

compare-and-swap · 0.5
YearPublicationVenuePosition
2021 A more pragmatic implementation of the lock-free, ordered, linked list
abstract
The lock-free, ordered, singly linked list as proposed in [5, 8] is a textbook example of a concurrent data structure [6, 10]. The data structure supports lock-free insertion and deletion, and wait-free contains operations on items identified by a unique key. The lock-free implementation is actually quite subtle. The ordering condition and a relaxed invariant makes it possible to do with a single-word Compare-And-Swap operation (CAS), and all operations can be shown to be linearizable even though linearization does not always happen at fixed points in the code. The lock-free data structure has many direct and indirect applications, notably in the implementation of concurrent skiplists and hash tables [8, 9, 11, 12].
Jesper Larsson Träff, Manuel Pöter
PPoPP2
2018 Stamp-it, amortized constant-time memory reclamation in comparison to five other schemes
abstract
The memory reclamation problem is to determine, for any given allocated memory node, when there are no more references to the node, allowing it to be safely returned to the memory management system. In a concurrent context, the memory reclamation problem is highly non-trivial, since there may be more than one thread referencing an allocated node unbeknownst to the other threads.
Manuel Pöter, Jesper Larsson Träff
PPoPP1
2018 Brief Announcement: Stamp-it, a more Thread-efficient, Concurrent Memory Reclamation Scheme in the C++ Memory Model
abstract
We present Stamp-it, a new, general, portable, lock-less concurrent memory reclamation scheme with amortized, constant-time (thread-count independent) reclamation overhead. Stamp-it has been implemented and proved correct in the C++ memory model using as weak memory-consistency assumptions as possible. We have (re)implemented six other comparable reclamation schemes. By a detailed performance comparison, we show that Stamp-it performs favorably, sometimes better, but at least as good as these other schemes while being able to reclaim free memory nodes earlier.
Manuel Pöter, Jesper Larsson Träff
SPAA1