Tootfinder

MLP-Offload: Multi-Level, Multi-Path Offloading for LLM Pre-training to Break the GPU Memory Wall
Avinash Maurya, M. Mustafa Rafique, Franck Cappello, Bogdan Nicolae
https://arxiv.org/abs/2509.02480

MLP-Offload: Multi-Level, Multi-Path Offloading for LLM Pre-training to Break the GPU Memory Wall
Training LLMs larger than the aggregated memory of multiple GPUs is increasingly necessary due to the faster growth of LLM sizes compared to GPU memory. To this end, multi-tier host memory or disk offloading techniques are proposed by state of art. Despite advanced asynchronous multi-tier read/write strategies, such offloading strategies result in significant I/O overheads in the critical path of training, resulting in slower iterations. To this end, we propose MLP-Offload, a novel multi-level,…

Autonomous Task Offloading of Vehicular Edge Computing with Parallel Computation Queues
Sungho Cho, Sung Il Choi, Seung Hyun Oh, Ian P. Roberts, Sang Hyun Lee
https://arxiv.org/abs/2509.03935

Autonomous Task Offloading of Vehicular Edge Computing with Parallel Computation Queues
This work considers a parallel task execution strategy in vehicular edge computing (VEC) networks, where edge servers are deployed along the roadside to process offloaded computational tasks of vehicular users. To minimize the overall waiting delay among vehicular users, a novel task offloading solution is implemented based on the network cooperation balancing resource under-utilization and load congestion. Dual evaluation through theoretical and numerical ways shows that the developed solution…

🇺🇦 Auf radioeins läuft...
Electric Light Orchestra (ELO):
🎵 Evil Woman
#NowPlaying #ElectricLightOrchestraELO
https://opaledit.bandcamp.com/track/electric-light-orchestra-evil-woman-opal-gta-edit
https://open.spotify.com/track/4N6HYihypXOVL24bnaR4fp

#PennedPossibilities 690 — SC POV: What do you fear losing the most?
This question is for the MC's brother, I believe.
SC: My brother.
It terrifies me to think that my brother could become completely corrupted and lose his true Ethereal essence to Ianthe's manipulation. You see, The Unmaking is not just simple indoctrination, it is the obliteration of free will,…

Gravitational Redshift for Rapidly Rotating Neutron Stars
Zsombor Jakab, Sharon Morsink
https://arxiv.org/abs/2507.02234 https://arxi…

Gravitational Redshift for Rapidly Rotating Neutron Stars
The Oblate Schwarzschild (OS) approximation is a method often used to compute the flux of X-rays emitted from a rapidly rotating neutron star. In this approximation, the oblate shape of the rotating star is embedded in the Schwarzschild metric, which is used to compute the redshift of photon energies as they propagate from the star to the telescope. In this paper, we demonstrate that there are small errors introduced by the standard treatment of photon redshift in the OS approximation and provi…

On the impact of unlimited computational power in OBLOT: consequences for synchronous robots on graphs
Serafino Cicerone, Alessia Di Fonso, Gabriele Di Stefano, Alfredo Navarra
https://arxiv.org/abs/2509.04383

On the impact of unlimited computational power in OBLOT: consequences for synchronous robots on graphs
The OBLOT model has been extensively studied in theoretical swarm robotics. It assumes weak capabilities for the involved mobile robots, such as they are anonymous, disoriented, no memory of past events (oblivious), and silent. Their only means of (implicit) communication is transferred to their positioning, i.e., stigmergic information. These limited capabilities make the design of distributed algorithms a challenging task. Over the last two decades, numerous research papers have addressed the…

Crosslisted article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[15/15]:
- MLP-Offload: Multi-Level, Multi-Path Offloading for LLM Pre-training to Break the GPU Memory Wall
Avinash Maurya, M. Mustafa Rafique, Franck Cappello, Bogdan Nicolae

Extremely large oblate deformation of the first excited state in $^{12}$C: a new challenge to modern nuclear theory
C. Ngwetsheni, J. N. Orce, P. Navr\'atil, P. E. Garrett, T. Faestermann, A. Bergmaier, M. Frosini, V. Bildstein, B. A. Brown, C. Burbadge, T. Duguet, K. Hady\'nska-Kl\c{e}k, M. Mahgoub, C. V. Mehl, A. Pastore, A. Radich, S. Triambak

Extremely large oblate deformation of the first excited state in $^{12}$C: a new challenge to modern nuclear theory
A Coulomb-excitation study of the high-lying first excited state at 4.439 MeV in the nucleus $^{12}$C has been carried out using the $^{208}$Pb($^{12}$C,$^{12}$C$^*$)$^{208}$Pb$^*$ reaction at 56 MeV and the {\sc Q3D} magnetic spectrograph at the Maier-Leibnitz Laboratorium in Munich. High-statistics achieved with an average beam intensity of approximately 10$^{11}$ ions/s together with state-of-the-art {\it ab initio} calculations of the nuclear dipole polarizability permitted the accurate det…

🇺🇦 Auf radioeins läuft...
Bruce Springsteen:
🎵 Streets Of Philadelphia
#NowPlaying #BruceSpringsteen
https://opaledit.bandcamp.com/track/bruce-springsteen-streets-of-philadelphia-opal-fx-remix
https://open.spotify.com/track/6WgBIPHwkQhCqMNnR5hQCN

Parallel CPU-GPU Execution for LLM Inference on Constrained GPUs
Jiakun Fan, Yanglin Zhang, Xiangchen Li, Dimitrios S. Nikolopoulos
https://arxiv.org/abs/2506.03296

Parallel CPU-GPU Execution for LLM Inference on Constrained GPUs
Deploying large language models (LLMs) for online inference is often constrained by limited GPU memory, particularly due to the growing KV cache during auto-regressive decoding. Hybrid GPU-CPU execution has emerged as a promising solution by offloading KV cache management and parts of attention computation to the CPU. However, a key bottleneck remains: existing schedulers fail to effectively overlap CPU-offloaded tasks with GPU execution during the latency-critical, bandwidth-bound decode phase…