Explore the Newsverum news archive, editorial policies and categories in Turkish and English, with Arabic, French, Spanish, German and Portuguese editions. (Page 310 of 436)
Researchers at the Technical University of Munich developed Ember, a memory manager that tracks which data should remain on a GPU when its memory fills. In the paper’s data-intensive workloads, it reported higher speeds than standard Unified Virtual Memory on three NVIDIA GPUs.
Researchers at KTH Royal Institute of Technology introduced GraphTok, a system that tokenizes large language model requests on a GPU. In a Dynamo frontend experiment, a real H100 NVL performed tokenization while backend GPU workers were simulated.
Researchers grew stoichiometric hexagonal TiAs single crystals by chemical vapour transport and combined spectroscopy, ARPES, transport measurements and calculations. Four Raman-active vibrational modes were observed, while calculations indicated symmetry-protected topological band crossings.
An experimental and theoretical study found Ni/Ru and Mn/Ga site mixing in the quaternary Heusler alloy NiRuMnGa and magnetic irreversibility near 48 K. The authors report ac susceptibility, magnetic relaxation and memory measurements as evidence supporting a cluster-glass state.
Researchers affiliated with Hong Kong Polytechnic University and other institutions in China modeled optical circuit switching alongside network scheduling to shift AI data-center training toward renewable-energy windows. In a controlled five-center scenario, modeled grid-purchase cost was 12.50% lower and local renewable-energy use rose by 2.724 percentage points.
Researchers compared monolayer, bilayer and trilayer regions of WS₂ in one specimen using electron ptychography. The measured phase of tungsten columns rose about 13.2% from one to two layers, but only 4.7% from two to three.
Researchers at UC Merced and the University of Texas at Austin introduced SlackSwap, a planner for deep-learning training. It recomputes some activations when the transfer budget between a GPU and host memory is exceeded; reported results come from simulation, not hardware measurements.