Skip to content
Snippets Groups Projects
Commit 7d728af6 authored by huongdm1896's avatar huongdm1896
Browse files

add GPU

parent 93694e7b
Branches
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
README.md
+ 33
− 6
View file @ 7d728af6
......@@ -8,7 +8,7 @@ This project provides tools to measure the energy consumption of Flower-based fe
- [Installation](#installation)
- [Usage](#usage)
- [FL framework](#fl-framework)
- [Configure instances](#configure-instance)
- [Configure instances for CPU](#configure-instance-for-cpu)
- [Run exp](#run-exp)
- [Quickstart](#quickstart)
- [Step 1. Reserve the Hosts in G5K](#step-1-reserve-the-hosts-in-g5k)
......@@ -51,13 +51,13 @@ cd Run
### FL framework
FL scripts (includes server and client scripts) can be updated, example in dir `Flower_v1`.
### Configure instance
### Configure instance for CPU
Configure instances of experiment in a json format, structure is shown below.
- **instances** includes **"1"**, **"2"** ,... are identifies of each instance.
- **instance**: name of instance.
- **output_dir**: location stores the output files (experiment log and energy monitoring output).
- **dvfs**: choose only one in 3 settings, detects all available CPU frequencies and go through all of them.
- **dvfs_cpu**: choose only one in 3 settings.
- `dummy`: for testing in min and max CPU freq (`false` or `true`).
- `baseline`: for testing in max CPU freq (`false` or `true`).
- `frequencies`: Limits to the provided list of frequencies (`null` or `int list []`).
......@@ -80,7 +80,7 @@ Configure instances of experiment in a json format, structure is shown below.
"1": {
"instance": "",
"output_dir": "",
"dvfs": {
"dvfs_cpu": {
"dummy": true,
"baseline": false,
"frequencies": null
......@@ -112,6 +112,25 @@ Configure instances of experiment in a json format, structure is shown below.
}
```
### Configure instance for GPU
- The configuration is as same CPU, except **dvfs** role. In GPU config, the role is **dvfs_gpu**.
Choose only one in 3 settings (steps - zoomfrom - zoomto use for one setting).
- `dummy`: for testing in min and max GPU freq (`false` or `true`).
- `baseline`: for testing in max GPU freq (`false` or `true`).
- `steps`: steps to jump in range/window of frequencies (int).
- `zoomfrom`: freq start
- `zoomto`: freq stop
```json
"dvfs_gpu": {
"dummy": true,
"baseline": false,
"steps": 2,
"zoomfrom": 0,
"zoomto": 0
},
```
### Run exp
2 options of experiment: run single instance or all instances (a campaign).
......@@ -137,17 +156,25 @@ Configure instances of experiment in a json format, structure is shown below.
### Step 1. Reserve the Hosts in G5K
Reserve the required number of hosts (*See the [document of G5K](https://www.grid5000.fr/w/Getting_Started#Reserving_resources_with_OAR:_the_basics) for more details*)
<u>For example</u>: Reserve 4 hosts (1 server + 3 clients) for 2 hours:
<u>For example</u>:
Reserve 4 hosts (CPU) (1 server + 3 clients) for 2 hours:
```bash
oarsub -I -l host=4,walltime=2
```
Reserve 4 hosts (GPU) (1 server + 3 clients) for 2 hours:
```bash
oarsub -p "gpu_count>0" -l host=4,walltime=2 -t exotic -I #lyon
oarsub -p "gpu_count>0" -l host=4,walltime=2 -I #lille
```
Make sure your are in`eflwr/Run/`:
```bash
cd Run
```
### Step 2. Configure
Create the JSON configuration file (e.g. `config_instances.json`) to specify experiment details includes one or more instances.
Create the JSON configuration file (e.g. `config_instances.json` for CPU and `config_instances_1.json` for GPU) to specify experiment details includes one or more instances.
```bash
vim config_instances.json
......
Run/config_instances.json
+ 2
− 2
View file @ 7d728af6
......@@ -3,7 +3,7 @@
"1": {
"instance": "fedAvg_cifar10",
"output_dir": "/home/mdo/Framework/eflwr/Log",
"dvfs": {
"dvfs_cpu": {
"dummy": true,
"baseline": false,
"frequencies": null
......@@ -57,7 +57,7 @@
"2": {
"instance": "fedAvg2Clients_cifar10",
"output_dir": "/home/mdo/Framework/eflwr/Log",
"dvfs": {
"dvfs_cpu": {
"dummy": true,
"baseline": false,
"frequencies": null
......
Run/config_instances_1.json 0 → 100644
+ 116
− 0
View file @ 7d728af6
{
"instances": {
"1": {
"instance": "fedAvg_cifar10",
"output_dir": "/home/mdo/Framework/eflwr/Log",
"dvfs_gpu": {
"dummy": true,
"baseline": false,
"steps": 2,
"zoomfrom": 0,
"zoomto": 0
},
"server": {
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/server_1.py",
"-r 1",
"-s fedAvg"
],
"ip": "172.16.66.76",
"port": 8080
},
"clients": [
{
"name": "client1",
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/client_1.py",
"cifar10",
"1",
"3"
],
"ip": "172.16.66.77"
},
{
"name": "client2",
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/client_1.py",
"cifar10",
"2",
"3"
],
"ip": "172.16.66.78"
},
{
"name": "client3",
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/client_1.py",
"cifar10",
"3",
"3"
],
"ip": "172.16.66.79"
}
]
},
"2": {
"instance": "fedAvg2Clients_cifar10",
"output_dir": "/home/mdo/Framework/eflwr/Log",
"dvfs_gpu": {
"dummy": true,
"baseline": false,
"steps": 2,
"zoomfrom": 0,
"zoomto": 0
},
"server": {
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/server_1.py",
"-r 1",
"-s fedAvg2Clients"
],
"ip": "172.16.66.76",
"port": 8080
},
"clients": [
{
"name": "client1",
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/client_1.py",
"cifar10",
"1",
"3"
],
"ip": "172.16.66.77"
},
{
"name": "client2",
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/client_1.py",
"cifar10",
"2",
"3"
],
"ip": "172.16.66.78"
},
{
"name": "client3",
"command": "python3",
"args": [
"/home/mdo/Framework/eflwr/Flower_v1/client_1.py",
"cifar10",
"3",
"3"
],
"ip": "172.16.66.79"
}
]
}
}
}
Run/measure_instance.py
+ 23
− 9
View file @ 7d728af6
......@@ -6,7 +6,7 @@ import json
import time
import expetator.experiment as experiment
from expetator.monitors import Mojitos
from expetator.leverages import Dvfs
from expetator.leverages import Dvfs, GpuClock
#import run_flwr as run_flwr
# Set up argument parser
......@@ -37,13 +37,27 @@ if instance_key not in config_data["instances"]:
# Load config instance
config = config_data["instances"][instance_key]
# Extract DVFS from config
dvfs_config = config.get("dvfs", {})
dvfs = Dvfs(
dummy=dvfs_config.get("dummy", False),
baseline=dvfs_config.get("baseline", False),
frequencies=dvfs_config.get("frequencies", None)
)
# extract DVFS infor (cpu and gpu)
if "dvfs_cpu" in config:
dvfs_config = config["dvfs_cpu"]
dvfs = Dvfs(
dummy=dvfs_config.get("dummy", False),
baseline=dvfs_config.get("baseline", False),
frequencies=dvfs_config.get("frequencies", None)
)
LEVERAGES = [dvfs]
elif "dvfs_gpu" in config:
dvfs_config = config["dvfs_gpu"]
dvfs = GpuClock(
dummy=dvfs_config.get("dummy", False),
baseline=dvfs_config.get("baseline", False),
steps=dvfs_config.get("steps", 2),
zoomfrom=dvfs_config.get("zoomfrom", 0),
zoomto=dvfs_config.get("zoomto", 0)
)
LEVERAGES = [dvfs]
else:
LEVERAGES = []
# Log directory
log_dir = config["output_dir"]
......@@ -78,7 +92,7 @@ if __name__ == "__main__":
experiment.run_experiment(
flower_log_dir,
[DemoBench()],
leverages=[dvfs],
leverages= LEVERAGES,
monitors=[Mojitos(sensor_set={'user', 'rxp', 'dram0'})],
times=args.repeat
)
\ No newline at end of file
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment