11:00
2026-09-04
infoq.com
artificial-intelligence
Presentation: From S3 to GPU in One Copy: Rethinking Data Loading for ML Training
Onur Satici of SpiralDB presented a data-loading pipeline that streams columnar data from S3 directly to GPUs at 13 gigabits per second, bypassing NVMe and CPU decompression bottlenecks, using the ope…