Posts

How to use TRUBA for deep learning with PyTorch - 2

Image
Today, we’ll build on the basics by diving deeper into TRUBA's functionalities. I will guide you through more advanced commands and configurations in your home directory. Following our baby steps metaphor, we’ll continue nurturing your project from its early stages to more complex operations. This tutorial is designed for macOS systems with Apple Silicon computers. Breathe When you want to execute computation, you need to prepare a .slurm file How to prepare .slurm file (e.g., job.slurm) ( 1 , 2 , 3 )  Go to personal directory: type cd /truba_scratch/username Start a .slurm file with touch jobName.slurm Delete a .slurm file with rm jobName.slurm Edit .slurm with command vi  To enter insert mode: Press i To save changes: Press Esc , then type :w and press Enter To exit without save: Press Esc , then type :q and press Enter To exit with save: Press Esc , then type :wq and press Enter Example .slurm file #!/bin/bash #SBATCH -J "trial" #SBATCH -A username #SBATCH

How to use TRUBA for deep learning with PyTorch - 1

Image
In this tutorial, we will guide you through connecting to TRUBA and executing basic commands in your home directory. We’ll use a baby steps metaphor : starting with project creation and progressing through initial operations. This guide is tailored for macOS systems with Apple Silicon computers. Birth How to connect to TRUBA:  If you are connecting from the university internet, you can directly use the terminal.  If you want to connect from outside of the university we will use OpenVPN To download OpenVPN To OpenVPN for Macos How to reach interface:  Interface:  levrek1.ulakbim.gov.tr  Then we will open our terminal from our local computer  Barbun1 connection to queue (using terminal):  Please type: ssh  username@172.16.7.1 Please type: password   The username and the password is given by the managers of the TRUBA  In here you can learn how to get it When you reach to the interface (using your termina):  You will see a line like that: username@barbun1:~$  Use pwd to learn where your

Urban Heat Island Measurement in Ankara, 2021

Image
  Ankara, Bahçelievler Neighborhood

Young CAADRIA Award - 2022

Image
 

An Exploratory Multi-objective Retrofit Decision-making Process

Image
The retrofit processes for buildings necessitates long-term planning and costly operations and requires a collaborative approach where a high number of alternative solutions should be explored by stakeholders. However, the evaluation of a range of retrofit solutions is a complex process wherein various design parameters and objectives are involved. The identification of the most effective solutions requires a collaborative evaluation in order to satisfy all stakeholders' expectations; however, during the decision-making process, stakeholders may generally have conflicting objectives. This paper discusses different user preference-based decision-making approaches for building retrofit that involves the collaborative evaluation of multiple design parameters and objectives simultaneously. For this purpose, we demonstrate a simulation-based approach for performative exploration for building retrofits, which may allow a broader consideration of alternative retrofit solutions to increase

The Comparative Study on the Influence of Early Architectural Design Decisions on Energy Demand: A Case Study in Turkey

Image
The early design process has the most salient design decisions for architects. It is crucial to observe the impact of these design decisions in terms of performance-based design. However, because of the large amount of variance of the performance criteria in the early design parameters, the decision-making is highly arduous. The current study proposes a method to quantify output uncertainty and presents the relationship between independent and dependent variables for providing insight into the decision-making process. The energy simulations for hypothetical office building based on TS-825 requirements are executed with cooling and heating demand (kWh/m 2-year) outputs for two different regions, i.e., Erzurum as a cold climate and Izmir as a hot-humid climate. Researchers compute the input parameters' impact on building performance with quasi-random statistical sampling and filtering techniques. Respectively, ineffective parameters are eliminated with factor fixing, and factor prior

Simaud2020 - Best Student Paper Award

Image