# How Object Storage Performance Affects GPU Training Workloads

DevFeed: [How Object Storage Performance Affects GPU Training Workloads](<https://devfeed.tech/articles/keeping-gpus-fed-34013.md>)

Original publisher: [Read original article](<https://sridharrajarao.com/blog/gpus-need-fast-object-storage/>)

Author: Sridhar Rajarao

Published: 2026-08-30T00:00:00Z

Content type: article

Language: en

Sources: [Sridhar Rajarao](<https://devfeed.tech/sources/sridhar-rajarao.md>)

Topics: [GPU](<https://devfeed.tech/topics/gpu.md>), [cloud-infrastructure](<https://devfeed.tech/topics/cloud-infrastructure.md>), [Caching](<https://devfeed.tech/topics/caching.md>), [Disk image](<https://devfeed.tech/topics/disk-image.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [aws](<https://devfeed.tech/tags/aws.md>), [cache](<https://devfeed.tech/tags/cache.md>), [compute](<https://devfeed.tech/tags/compute.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [nvme](<https://devfeed.tech/tags/nvme.md>), [object-storage](<https://devfeed.tech/tags/object-storage.md>), [performance](<https://devfeed.tech/tags/performance.md>), [storage](<https://devfeed.tech/tags/storage.md>), [training-data](<https://devfeed.tech/tags/training-data.md>)

## AI overview

This article explains how object-storage latency, network transfers, many small files, and concurrent requests can leave GPUs idle during AI training. It recommends sharding small samples into larger files, reusing connections, and caching data near the GPU; it also notes AWS's suggested shard range of roughly 100 MB to 1 GB for many training workloads.

## Source excerpt

The GPU does not care that object storage is durable and scalable. It cares whether the next batch of data arrives before it goes idle.