Skip to content

CUDA Caching Allocator Fragmentation

CUDA caching allocator fragmentation causes OOM despite enough total free memory, due to non-contiguous blocks.

Quick answer

CUDA caching allocator fragmentation causes OOM despite enough total free memory, due to non-contiguous blocks.

Memory#cuda-allocator#fragmentation#memory#memory-pool#cuda

What this failure is

CUDA Caching Allocator Fragmentation is a Memory failure seen during ML training runs. CUDA caching allocator fragmentation causes OOM despite enough total free memory, due to non-contiguous blocks. Common tags: Cuda Allocator, Fragmentation, Memory, Memory Pool.

Live diagnosis, no signup

Is this what broke your run? Paste your log.

You're reading about CUDA Caching Allocator Fragmentation. Paste your own crash log or traceback below and get the real root cause for YOUR run, not this generic entry. No account, no card. Logs are masked at ingress and never saved to account history.

training_logs.txt
No log to hand? Try one:

3 free diagnoses/day

Want 14 days on the Scale plan?

Request an evaluation code. A verified workplace organization activates up to 50 diagnoses a day, alerts, history, and follow-up questions. No credit card or automatic subscription.

Evaluate one incident

Why it happens (the mechanism)

Variable tensor sizes cause fragmentation. Allocs and frees of different sizes. Long-running training accumulates fragmentation. Memory pool cannot coalesce small blocks. Taken together, these mechanisms explain why the failure is reproducible, why it tends to surface on specific workloads or scales, and why generic mitigation attempts often fall short without addressing the underlying cause.

What you'll observe

  • OOM despite enough free memory
  • Reserved memory is high but allocated is low
  • Memory fragmentation visible in snapshot

Common symptoms and what they mean

SymptomWhy it happens
torch.cuda.memory_reserved() >> memory_allocated()Variable tensor sizes cause fragmentation
Empty cache doesn't helpAllocs and frees of different sizes
Snapshot shows many small free blocksLong-running training accumulates fragmentation

Which systems are affected

  • Variable-shape training
  • Long-running training
  • Models with dynamic shapes

How to confirm this is the problem

Use this checklist to test the hypothesis against a small reproduction. No single line proves the root cause, so preserve the preceding events and compare one variable at a time.

  • Reproduce the failure from a clean checkpoint/seed: the symptom must appear without warm-up state from a previous run.
  • Verified signal present: torch.cuda.memory_reserved() >> memory_allocated()
  • Verified signal present: Empty cache doesn't help
  • Verified signal present: Snapshot shows many small free blocks
  • A targeted fix from the "How to fix it" section eliminates or substantially reduces the symptom within one validation pass.

The fix and the prevention pattern

The root cause is on this page and stays free. A free account adds the exact remediation steps, saved history, and the fix on every entry in the encyclopedia.

Sign up free. Unlock the full analysis

No credit card. Daily allowance follows verified trust tier. Instant access.

Diagnose this failure in VS Code

Select the traceback or open the failed terminal, then run Denpex locally to see the initiating rank, collateral failures, exact fix, and verification command without uploading the log.

Install the free VS Code extension

CUDA errors in context

CUDA reports errors asynchronously, so the traceback usually points at whatever line synchronised next rather than the one at fault. The hub covers every common CUDA error and how to make it report honestly.

Compare every cuda error side by side

Root cause

  • Variable tensor sizes cause fragmentation
  • Allocs and frees of different sizes
  • Long-running training accumulates fragmentation
  • Memory pool cannot coalesce small blocks

The fix and how to prevent it

Evaluate Denpex on your own logs

Request a Scale evaluation code. A verified workplace organization activates 14 days with up to 50 diagnoses a day. Every account keeps its current diagnosis allowance and gets a verification path. No card or automatic subscription.

We send a single-use code tied to that address. Static provider and TLD rules do not reject valid addresses. Account trust determines the benefit after signup.

Don't just read the fix, diagnose your run

The encyclopedia tells you what went wrong. Denpex tells you what went wrong in YOUR training run. With your logs, your config, and your stack.