Julie Dumas

Dr. Julie Dumas

Univ. Grenoble Alpes, CNRS, Grenoble INP*, TIMA, 38000 Grenoble, France * Institute of Engineering Univ. Grenoble Alpes
orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
KMKévin Mambu
Kévin Mambu
HCHenri-Pierre Charles
Henri-Pierre Charles
MKMaha Kooli
Maha Kooli
JDJulie Dumas
Julie Dumas
In-memory computing (IMC) aims to solve the performance gap between CPU and memories introduced by the memory wall. However, it does not address the energy wall problem caused by data transfer over memory hierarchies. This paper proposes the data-locality management unit (DMU) to efficiently transfer data from a DRAM memory to a computational SRAM (C-SRAM) memory allowing IMC operations. The DMU is tightly coupled within the C-SRAM and allows one to align the data structure in order to perform effective in-memory computation. We propose a dedicated instruction set within the DMU to issue data transfers. The performance evaluation of a system integrating C-SRAM within the DMU compared to a reference scalar system architecture shows an increase from ×5.73 to ×11.01 in speed-up and from ×29.49 to ×46.67 in energy reduction, versus a system integrating C-SRAM without any transfer mechanism compared to a reference scalar system architecture.

1 July 2018
Verified
Conference Paper
Closed access
JDJulie Dumas
Julie Dumas
EGEric Guthmuller
Eric Guthmuller
FPFrederic Petrot
Frederic Petrot
The most widely used programming models expect hardware to guarantee coherent shared memory accesses. However, with the increasing number of integrated cores on chip, resource and performance efficient scalable cache coherence protocols are needed. To address the scalability issues due to the size of the sharing set, we propose to encode, on a fixed size bit-vector, a rectangular cluster whose goal is to cover most of the sharers. The cluster size is fixed but its height, width and position are determined for each cache block and can change during execution. We use a fixed size linked list for the first few outliers, and resort to broadcast when the list overflows. We compare our solution to snoop, directory-based full bit-vector, and Ackwise. It leads to similar mean latency and 10% less traffic than Ackwise, and only a few percent more than the complete sharing set on these metrics. More importantly, it generates ten times less broadcasts than Ackwise while using similar hardware resources for a 64 cores architecture.

Verified
Journal Article
Open access
PDF file
JDJulie Dumas
Julie Dumas
HCHenri-Pierre Charles
Henri-Pierre Charles
KMKévin Mambu
Kévin Mambu
MKMaha Kooli
Maha Kooli
This article describes a software environment called HybroGen, which helps to experiment binary code generation at run time. As computing architectures are getting more complex, the application performance is becoming data-dependent. The proposed experimental platform is helpful in programming applications that can be reconfigured at run time in order to be adapted for a new data environment. The HybroGen platform is adapted to heterogeneous architectures and can generate instructions for different targets. This platform allows to go farther than classical JIT compilation in many directions: the code generator is smaller by three orders of magnitude and faster by three orders of magnitude, compared to JIT (Just-In-Time) platforms, and allows making code transformation that is impossible in traditional compilation schemes, such as code generation for non von Neumann accelerators or dynamic code transformations for transprecision. The latter is illustrated in a code example: the square root with Newton’s algorithm. We also illustrate the proposed HybroGen platform with two other examples: a multiplication with a specialization on a value determined at run time, and a conversion of degrees Celsius to degrees Fahrenheit. This article presents a proof of concept of the proposed HybroGen platform in terms of its functionalities, and demonstrates the working status.

17 January 2023
Verified
Conference Paper
Open access
MBMarie Badaroux
Marie Badaroux
JDJulie Dumas
Julie Dumas
FPFrederic Petrot
Frederic Petrot
Given the performances it achieves, dynamic binary translation is the most compelling simulation approach for cross-emulation of software centric systems. This speed comes at a cost: simulation is purely functional. Modeling instruction caches by instrumenting each target instruction is feasible, but severely degrades performances. As the translation occurs per target instruction block, we propose to model instruction caches at that granularity. This raises a few issues that we detail and mitigate. We implement this solution in the QEMU dynamic binary translation engine, which brings up an interesting problem inherent to this simulation strategy. Using as test vehicle a multicore RISC-V based platform, we show that a proper model can be nearly as accurate as an instruction accurate model. On the PolyBench/C and PARSEC benchmarks, our model slows down simulation by a factor of 2 to 10 compared to vanilla QEMU. Although not negligible, this is to be balanced with the factor of 20 to 60 for the instruction accurate approach.

1 June 2017
Verified
Conference Paper
Closed access
JDJulie Dumas
Julie Dumas
EGEric Guthmuller
Eric Guthmuller
CTCesar Fuguet Tortolero
Cesar Fuguet Tortolero
FPFrederic Petrot
Frederic Petrot
Doing early design space exploration for manycore architectures is a challenge, all the more when the focus is on complex coherence protocols. Implementing such protocols in realistic simulation models is costly both in modelling effort and execution time. We propose a trace-driven method to accurately compare cache coherence protocols while keeping cache modelling at a high level of abstraction. We show what kind of design space exploration can be performed on an existing sharing set implementation proposal, the linked list sharing set management. By doing so, we demonstrate that our approach, while being still fairly accurate, is much easier to develop and much faster to execute than state of the art low level simulators.

4 March 2017
Verified
Book Chapter
Closed access
PDF file
JDJulie Dumas
Julie Dumas
EGEric Guthmuller
Eric Guthmuller
CTCésar Fuguet Tortolero
César Fuguet Tortolero
FPFrédéric Pétrot
Frédéric Pétrot