dmitrim

6 ms·

1.
Why Do Nigerian Scammers Say They Are from Nigeria?(microsoft.com)
2 points by dmitrim 4y ago|0 comments
2.
by dmitrim 4y ago
Understand, we'll definitely think about the network part. Just in case it may help, if `nvidia-smi nvlink -gt d` is useful for you in this context then there is a related metric NVLink Throughput Rate to compare runs and monitor. At l
3.
by dmitrim 4y ago
For now it only tries to extract NCCL time percentage from the profile, if available, and show it profile summary. Some hints count be in the step trace timeline as well. We are planning to record some NCCL related counters separately as we
4.
Show HN: Graphsignal – ML profiler to speed up training and inference(github.com)
16 points by dmitrim 4y ago|7 comments
5.
by dmitrim 5y ago
To be more precise, >=2.2 is required for profiler support
6.
by dmitrim 5y ago
Thanks for the feedback, we'll be working on it for sure. At least an explanatory screencast is in the works now, other info material, use cases, etc. are planned.
7.
by dmitrim 5y ago
I'd just use the 2.8, but older 2.x versions should work too. If you encounter are any issues, please let us know via the chat in your account.
8.
by dmitrim 5y ago
Many training and inference workloads run in the cloud or on remote servers and profiling them is not straightforward. Having a SaaS makes things much simpler, and also enables additional features, such as team access and sharing. As far as
9.
Show HN: Graphsignal – Machine learning profiler for training and inference(graphsignal.com)
35 points by dmitrim 5y ago|8 comments
10.
Show HN: Graphsignal – Production Model Monitoring: Anomalies, Drift, Exceptions(github.com)
30 points by dmitrim 5y ago|4 comments
11.
Model Monitoring for Large-Scale Deployments(graphsignal.com)
2 points by dmitrim 5y ago|0 comments
12.
by dmitrim 5y ago
Since ML monitoring is rather a broad term that can be applied to model development, evaluation, retraining and production stages, I'd like to give more context on what Graphsignal is designed for. Our focus is the operational aspects
13.
Show HN: Graphsignal – Operational Machine Learning Model Monitoring(graphsignal.ai)
3 points by dmitrim 5y ago|1 comments
14.
Show HN: StackImpact – Java Production Profiler(github.com)
1 points by dmitrim 8y ago|0 comments
15.
Show HN: TensorFlow Performance Profiler for Development and Production(stackimpact.com)
3 points by dmitrim 8y ago|0 comments
16.
by dmitrim 9y ago
Yep, that meant to be capacity, not length. Corrected. Thanks!
17.
by dmitrim 9y ago
Good point, thanks! The idea behind these benchmarks is to make the results usable in real-world programs, rather than benchmarking real-world programs. I rephrased that sentence to avoid any confusion.
18.
by dmitrim 9y ago
Thanks for pointing it out. Should clearly not depend on the number of iterations. It's fixed now.
19.
Code-Level Go Monitoring(stackimpact.com)
3 points by dmitrim 9y ago|0 comments
20.
Show HN: StackImpact – Performance Profiler as a Service(stackimpact.com)
2 points by dmitrim 9y ago|0 comments
21.
Practical Golang Benchmarks(stackimpact.com)
2 points by dmitrim 9y ago|0 comments
22.
CPU Profiling in Production Node.js Applications(stackimpact.com)
2 points by dmitrim 9y ago|0 comments
23.
Profiling AWS Lambda Go Functions: CPU, Memory and Latency(stackimpact.com)
1 points by dmitrim 9y ago|0 comments
24.
Understanding Application Footprint with Code-Level Detail(stackimpact.com)
2 points by dmitrim 9y ago|0 comments
25.
Show HN: StackImpact – Node.js Production Profiler: CPU, Memory, Async Calls(github.com)
4 points by dmitrim 9y ago|0 comments
26.
Show HN: StackImpact – Go Production Profiler: CPU, Memory, Blocking Calls(github.com)
7 points by dmitrim 9y ago|0 comments
27.
Continuous Performance Profiling(stackimpact.com)
1 points by dmitrim 9y ago|0 comments
28.
by dmitrim 9y ago
We haven't tested it with celery yet. It looks like it should work. gevent is supported by blocking call profiler, and CPU and memory profilers as well as exception and metric reporting are library independent.
29.
by dmitrim 9y ago
We haven't tested the whole agent with asyncio applications yet. I guess only CPU profiler was tested during development. We'll do and include it in the docs. For now, if you see any problems, please just open a ticket. Thanks!
30.
by dmitrim 9y ago
Yes, the apps were under simulated CPU load, memory allocations, etc. The good thing with sampling profilers is that overhead stays relatively stable even under high load.

More ›