4 papers
KernelEvolve: Scaling Agentic Kernel Coding for Heterogeneous AI Accelerators at Meta
Gang Liao, Hongsen Qin, Ying Wang +36
Making deep learning recommendation model (DLRM) training and inference fast and efficient is important. However, this presents three key system challenges - model architecture div…
Agentic Operator Generation for ML ASICs
Alec M. Hammond, Aram Markosyan, Aman Dontula +15
We present TritorX, an agentic AI system designed to generate functionally correct Triton PyTorch ATen kernels at scale for emerging accelerator platforms. TritorX integrates open-…
Unifying and accelerating level-set and density-based topology optimization by subpixel-smoothed projection
Alec M. Hammond, Ardavan Oskooi, Ian M. Hammond +3
We introduce a new "subpixel-smoothed projection" (SSP) formulation for differentiable binarization in topology optimization (TopOpt) as a drop-in replacement for previous projecti…
Inverse design of 3D-printable metalenses with complementary dispersion for terahertz imaging
Mo Chen, Ka Fai Chan, Alec M. Hammond +2
This study formulates a volumetric inverse-design methodology to generate a pair of complementary focusing metalenses for terahertz imaging: the two lenses exhibit equal and opposi…