Update install instruction for 20.04 LTS

docs/source/installation.md

@@ -16,54 +16,88 @@ docker build -t arch-gym-image .
 docker run -it --name arch-gym-container arch-gym-image
 ```
 
-## 1b) Native Install
-###### Clone ArchGym Repository 
-```shell
-git clone https://github.com/srivatsankrishnan/oss-arch-gym.git
-```
-###### Install Anaconda / Miniconda if not already present
+## 1b) Native Install (Tested on Ubuntu 20.04 LTS)
+
+###### Install Basic Packages & Anaconda / Miniconda if not already present
 ```
+sudo apt-get update -y
+sudo apt-get install -y sudo git wget build-essential vim libgmp-dev 
 wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
 chmod +x Miniconda3-latest-Linux-x86_64.sh
-bash ./Miniconda3-latest-Linux-x86_64.sh -b -f -p /usr/local
+sudo bash ./Miniconda3-latest-Linux-x86_64.sh -b -f -p /usr/local
+conda init
+exit
 ```
 
-###### Create Conda Environment for ArchGym
+###### Clone ArchGym Repository and Create Conda Environment 
+Start a new terminal and run:
 ```
+git clone --recursive https://github.com/srivatsankrishnan/oss-arch-gym.git
 conda env create -f oss-arch-gym/environment.yml
+conda activate arch-gym
 ```
 
-###### Install ACME Framework
-
-Activate the arch-gym conda environment, then in `oss-arch-gym/acme` run: 
+###### Install ACME Framework for Reinforcement Learning
 ```
+cd oss-arch-gym/acme
 pip install .[tf,testing,envs,jax]
 ```
 
 Now run `conda env list`, which will show the path to your arch-gym environment.
 
-Add the following line to your bashrc:
+Add the following line to your `bashrc`:
 ```
 export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:<PATH_TO_ARCH-GYM_CONDA_ENV>/lib/
 ```
+and then source your updated `bashrc` and re-activate the environment:
+```
+source ~/.bashrc
+conda activate arch-gym
+```
 
-###### Install Vizier 
+###### Install Vizier for Collection of Agents
 From the repository root `oss-arch-gym` run:
 ```
 ./install_sim.sh viz
 ```
 
-## 2) Activate ArchGym Environment 
+## 2) Verify Setup
+###### Verify ArchGym Installation 
+Run:
 ```
-conda activate arch-gym
+cd sims/DRAM
+python train_ga_DRAMSys.py
 ```
-
-## 3) Verify Setup
-Navigate to `sims/DRAM` and run:
+You should see the following output (with perhaps different values):
 ```
-python train_ga_DRAMSys.py
+iter: 0
+Agents Action [  3.   2.   2.   8.   1.   0.   3.   6.   2. 103.]
+[  3.   2.   2.   8.   1.   0.   3.   6.   2. 103.]
+Action Dict {'PagePolicy': 'ClosedAdaptive', 'Scheduler': 'FrFcfs', 'SchedulerBuffer': 'Shared', 'RequestBufferSize': 8, 'RespQueue': 'Reorder', 'RefreshPolicy': 'NoRefresh', 'RefreshMaxPostponed': 3, 'RefreshMaxPulledin': 6, 'Arbiter': 'Reorder', 'MaxActiveTransactions': 103}
+[envHelpers][Action] {'PagePolicy': 'ClosedAdaptive', 'Scheduler': 'FrFcfs', 'SchedulerBuffer': 'Shared', 'RequestBufferSize': 8, 'RespQueue': 'Reorder', 'RefreshPolicy': 'NoRefresh', 'RefreshMaxPostponed': 3, 'RefreshMaxPulledin': 6, 'Arbiter': 'Reorder', 'MaxActiveTransactions': 103}
+[Environment] Observation: [2.46681075e-01 2.82248000e+00 8.73850000e-05]
+Power: 2.82248 Latency: 8.7385e-05 Target Power: 1 Target Latency: 0.1
+Episode: 0  Rewards: 0.5487028664237742
+...
 ```
-You should see the following output:
+You can `ctrl-c` to kill the training run when you see the similar output above. You have successfully installed ArchGym!
+
+###### Verify Vizier Installation 
+Run:
 ```
+ python train_randomsearch_vizier.py
+```
+You should see the following output (with perhaps different values):
 
 ```
+...
+['0' '2' '0' '32' '0' '1' '4' '2' '0' '32']
+Action Dict {'PagePolicy': 'Open', 'Scheduler': 'FrFcfs', 'SchedulerBuffer': 'Bankwise', 'RequestBufferSize': 32, 'RespQueue': 'Fifo', 'RefreshPolicy': 'AllBank', 'RefreshMaxPostponed': 4, 'RefreshMaxPulledin': 2, 'Arbiter': 'Simple', 'MaxActiveTransactions': 32}
+[envHelpers][Action] {'PagePolicy': 'Open', 'Scheduler': 'FrFcfs', 'SchedulerBuffer': 'Bankwise', 'RequestBufferSize': 32, 'RespQueue': 'Fifo', 'RefreshPolicy': 'AllBank', 'RefreshMaxPostponed': 4, 'RefreshMaxPulledin': 2, 'Arbiter': 'Simple', 'MaxActiveTransactions': 32}
+E1018 11:11:07.816930157   15103 fork_posix.cc:76]           Other threads are currently calling into gRPC, skipping fork() handlers
+[Environment] Observation: [2.41416e-01 3.17027e+00 7.61500e-05]
+Power: 3.17027 Latency: 7.615e-05 Target Power: 1 Target Latency: 0.1
+Episode: 0  Rewards: 0.4607721619890613
+```
+
+You can `ctrl-c` to kill the training run when you see the similar output above. You have successfully installed Vizier's Agents!