Comments (4)
It's on the list but pretty low priority since we seldom run into it. If you can figure out why the lock mechanism in torch.utils.cpp_extension.load
doesn't always work as I think it should I'm all ears. :) Might be we are expecting more from it than it is meant to handle, might be we are using it wrong, might be a bug in torch, etc.
from megatron-lm.
We've had reports of this happening internally as well. The problem with only compiling on the master rank is we need to support the use case that the code lives locally on several nodes (i.e. a multinode environment and the code isn't on a shared disk). So the kernels need to be compiled on several ranks.
torch.utils.cpp_extension.load
is supposed to lock the directory so that only one rank runs the build at a time (and only the first run actually does anything, subsequent runs just see that it is already built and returns). Apparently that lock isn't properly deleted/released in some instances.
from megatron-lm.
Thanks for your response! I'm glad that your group already had reports of this happening.
Is your group internally planning to resolve this issue?
from megatron-lm.
I will close this issue, and if there is an additional issue then I will re-open it. Thank you :)
from megatron-lm.
Related Issues (20)
- [QUESTION] Training Llama3 70B on 16 x A100 only achieves low throughput of 20 TFLOPS HOT 1
- [QUESTION]Why is the operator of computation slower when computation overlaps with communication
- [QUESTION] Will the data get re-shuffled if the sequence length is modified during training?
- [BUG] bugs when using pytorch2.4.0 to run run_simple_mcore_train_loop.py HOT 1
- [BUG] T5 extended attention mask shape mismatch with transformer engine HOT 5
- [QUESTION]
- [QUESTION] Checkpoint storage format
- [QUESTION] Megatron-LM `DistributedOptimizer` or NeMo `MegatronDistributedFusedAdam` Optimizer?
- [QUESTION]Splitting large document and bucketing
- [QUESTION] glu activation with tensor parallel in GroupedMLP HOT 2
- [BUG] arguments of get_cpu_offload_context() in transformer_engine.py for different version of te HOT 4
- [QUESTION] Does it support Knowledge Distillation? HOT 1
- No pre-norm for non-moe gpt style model when using TE-transformer layer spec??? HOT 2
- [BUG] clip key mismatch HOT 1
- [BUG] error raised while converting llm to megatron
- FASE 6 LILITI STK 3.6.9 INTELIGÊNCIA ARTIFICIAL ANTI CARBONO.
- [QUESTION] How to freeze specific modules while training? HOT 1
- [QUESTION] Why and When dose matmul call different kernels? HOT 1
- [BUG] llava pipeline parallel initialization problem HOT 1
- Why is gather_output not supported in ColumnParallelLinear when using sequence parallelism?
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from megatron-lm.