Daniel Haehn PRO
Hi, I am a biomedical imaging and visualization researcher who investigates how computational methods can accelerate biological and medical research.
Make sure you have an SSH client installed!
ENTRYTASK
Due 4/20 11:59pm!
MRI Lecture by Avanith
Wed 4/15
Monday
No class!
Next week: Cluster Assignment!
Sumientra Rampersad's EEG Lab!
Wed 4/22 Fieldtrip
Infrastructure and Algorithms for Data Collections
MACHINE
PSYCHOLOGY
DANIELHAEHN.COM
AI Core
Daniel Haehn
Director
AI Core
We
Computers
GONZO
MONSTER
VAMPIRE
ZOMBIE
HYDRA
2x NVIDIA DGX A100
2x NVIDIA DGX A100
2 TB RAM
30 TB Storage
8x A100 GPUs
High-Performance Switch
NVIDIA DGX H200
1+ TB GPU Memory
Let's try the DGX H200 machine!
Jeff Dusenberry
Director of Research Computing
at UMass Boston
ssh FIRSTNAME.LASTNAME001@chimera.umb.edu
chimera21
salloc -A cs666 --mem=128G -w chimera21 -p pomplun --gres=gpu:1 -t 60
Python Package Manager
curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
chmod +x Mini*
./Mini*conda create -n CS666 python=3.10
conda activate CS666
pip install "tensorflow[and-cuda]"
pip install jupyterlab ipykernel
mkdir -p $CONDA_PREFIX/etc/conda/activate.d
vim $CONDA_PREFIX/etc/conda/activate.d/env_vars.shthis whole block goes into the env_vars.sh file!
export LD_LIBRARY_PATH=$(python - <<'PY'
import site, os, glob
paths=[]
for p in site.getsitepackages():
paths += glob.glob(os.path.join(p,"nvidia","*","lib"))
print(":".join(paths))
PY
):$CONDA_PREFIX/lib:$LD_LIBRARY_PATH
nvidia-smi -L
python -c "import tensorflow as tf; print(tf.config.list_logical_devices('GPU'))"Check if Tensorflow can reach the GPU!
jupyter lab --port 44333 --ip 0.0.0.0HOST=chimera21 PORT=44333 && ssh daniel.haehn@chimera.umb.edu -L $PORT:$HOST:$PORT -N
1. Explain the Story
2. Discuss the Paper
CS666 Lecture 18
By Daniel Haehn
CS666 Lecture 18
Slides for CS666 Biomedical Signal and Image Processing at UMass Boston. See https://cs666.orghttps://cs666.org
