A platform for GPU programming challenges. Write efficient GPU kernels and compare your solutions with other developers!

• Node.js 18+ and npm/yarn/pnpm
• PostgreSQL (for local development) or a Supabase account
• Git

1. Clone the repository:

git clone https://github.com/tensara/tensara
cd tensara

2. Install dependencies:

npm install
# or
yarn
# or
pnpm install

3. Set up your environment variables by copying the example file:

cp .env.example .env

You have a few options for setting up the database:

1. Install PostgreSQL on your machine if you haven't already:

   • macOS (using Homebrew): brew install postgresql
   • Linux: sudo apt-get install postgresql
   • Windows: Download from PostgreSQL website

2. Start PostgreSQL service:

   • macOS: brew services start postgresql
   • Linux: sudo service postgresql start
   • Windows: It should start automatically as a service

3. Create a new database:

createdb tensara_db

4. Update your .env file with local PostgreSQL connection string:

DATABASE_URL="postgresql://your_username@localhost:5432/tensara_db"

1. Create a new project on Supabase

2. Once your project is created, go to Settings > Database to find your connection string

3. Update your .env file with the Supabase connection string:

DATABASE_URL="postgresql://postgres:[YOUR-PASSWORD]@db.[YOUR-PROJECT-REF].supabase.co:5432/postgres"

1. Push the database schema:

npx prisma db push

2. Generate Prisma Client:

npx prisma generate

3. Push the problems to the database. Follow the steps in the problems repository.
4. Start the development server:

npm run dev
# or
yarn dev
# or
pnpm dev

Your app should now be running at http://localhost:3000!

Create a .env file in the root directory with the following variables:

# Database connection string
DATABASE_URL="your_connection_string_here"

# NextAuth configuration
NEXTAUTH_SECRET="your_nextauth_secret"
AUTH_GITHUB_ID=""
AUTH_GITHUB_SECRET=""
NEXTAUTH_URL="http://localhost:3000"

# Google Analytics ID (can be ignored until production)
NEXT_PUBLIC_GA_ID=""

MODAL_CHECKER_SLUG=""
MODAL_BENCHMARK_SLUG=""

MODAL_ENDPOINT=""

Thank you to our sponsors who help make Tensara possible:

• Modal - Modal lets you run jobs in the cloud, by just writing a few lines of Python. Customers use Modal to deploy Gen AI models at large scale, fine-tune large language models, run protein folding simulations, and much more.

We use Modal to securely run accurate benchmarks on various GPUs.

Interested in sponsoring? Contact us at sponsor@tensara.org