Comments (6)
In 'batch_instance.csv', I found 8673603 records with 'real_cpu_max' larger than 1, which takes up ~54% of total records. To this end, I think it is a reasonable number.
Maybe I misunderstand your question?
from clusterdata.
Some tasks in the batch_task table has a plan cpu in decimal.
can you explain what does the plan cpu of 0.5 mean in the batch_task table?
from clusterdata.
It means that we estimate that each instance of the task would use 0.5 cpu. It is only a reference for us to decide where to put an instance of this batch task (and it is not accurate). As you know, the actual cpu used per instance is recorded in batch_instance and does not equal to the plan cpu.
from clusterdata.
Does this mean that the scheduler doesn't allocate the entire cpu for a task?
Do tasks share cpu?
Do batch jobs run on cpu allocated for service jobs?
Can you suggest an way to allocate resources to a task if we simulate the scheduler?
from clusterdata.
Does this mean that the scheduler doesn't allocate the entire cpu for a task?
For batch, no
Do tasks share cpu?
Yes
Do batch jobs run on cpu allocated for service jobs?
Tasks (jobs) have several priorities. For the tasks that have low priority, they can run on cpus that are allocated for service jobs, at a risk of being preempted whenever service jobs need the resource. The mechanism behind cannot be described with a few words but many works are using similar idea.
Can you suggest an way to allocate resources to a task if we simulate the scheduler?
Well, I am not sure if I could help on this one. In fact, you can make any assumption about the priority of tasks as long as it helps with your goal of the design. We could have further discussion on this if you have more specific need. Just send me email and we could also have a call about this.
from clusterdata.
Closing the issue.
from clusterdata.
Related Issues (20)
- P90 GPU utilization for 8-GPU
- How to explain the end_time is less than start_time in 2018 trace's batch_task?
- Will a call graph be stored in multiple files? HOT 2
- cluster-trace-microservices-v2022 release HOT 1
- I can't download 2017 HOT 1
- cluster-trace-microservices-v2021 Time of Data Collection
- Microservice-2021数据集中的um和dm值为”(?)“ HOT 1
- The link of cluster-trace-v2017 is invalid HOT 1
- Why are some call graph in cluster-trace-microservices-v2021 disconnected? HOT 4
- Issue in downloading of cluster-trace-microarchitecture-v2022 HOT 2
- Data type mismatch in container_meta in micro-architecture trace HOT 1
- Changing of curl scripts to wget to download data of micro services 2021 HOT 2
- microService-v2022-clusterData HOT 4
- Problem downloading .tar.gz files HOT 6
- cluster-trace-gpu-v2020中pai_sensor_table 中cpu_usage, gpu_wrk_util, avg_gpu_wrk_mem , max_gpu_wrk_mem 疑问 HOT 1
- cluster-trace-gpu-v2020 论文5.2 Predictable Duration for Recurring Tasks章节中使用cart 模型预测时间是什么时间 HOT 2
- 请问pai_job_duration_estimate_100K.csv 的列中有表示论文5.2中预测的时间吗 HOT 1
- 请问PAI数据中instance 表相同start_time在相同机器执行的instance,是代表多个任务同时在一台机器上运行吗 HOT 2
- No Output generated on NURD straggler prediction
- 请问ICPP'22中的kperf工具会开源吗?
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 clusterdata.