cuda

cuda is a cursor rule for coding agents from sanjeed5/awesome-cursor-rules-mdc. It costs 2,336 tokens per session, scanned A, original, CC0-1.0.

A set of rules for writing CUDA C++ programs that run work on NVIDIA graphics processors. It covers checking errors, handling memory, and improving GPU program performance.

In plain words
What is it for?
Use it when writing CUDA kernels and C++ applications that allocate GPU memory, move data, launch GPU work, and check CUDA API calls.
Why use it?
It helps expose GPU errors immediately instead of allowing them to cause failures later. It also provides guidance for code that is easier to maintain and debug.

Cursor rule

About the project

awesome-cursor-rules-mdc is a generator that creates Cursor MDC rule files from structured library information, using semantic search and language models to gather and organize guidance. Developers use it to produce reusable rules for libraries in Cursor, and the catalogue includes 200 of those rules.

sanjeed5/awesome-cursor-rules-mdc · 3,571 stars · on GitHub

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add rules/sanjeed5/awesome-cursor-rules-mdc/cuda
Clone the repo
git clone --depth 1 https://github.com/sanjeed5/awesome-cursor-rules-mdc

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for cuda

README.md
[![agentmods](https://agentmods.dev/badge/rules/sanjeed5/awesome-cursor-rules-mdc/cuda.svg)](https://agentmods.dev/rules/sanjeed5/awesome-cursor-rules-mdc/cuda)
Your own site
<a href="https://agentmods.dev/rules/sanjeed5/awesome-cursor-rules-mdc/cuda"><img src="https://agentmods.dev/badge/rules/sanjeed5/awesome-cursor-rules-mdc/cuda.svg" alt="Measured on agentmods" height="20"></a>
Per session 2,336 This file is loaded in full into every session.
When invoked 2,336 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.02336 $0.02336
Opus 5 $0.01168 $0.01168
Sonnet 5 $0.00467 $0.00467
Haiku 4.5 $0.00234 $0.00234

Measured 5d ago against content hash 7650ffc664ca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cuda scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

rules-mdc/cuda.mdc · 286 lines

How it starts

The opening of the file, as written. The whole thing — 286 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CUDA Best Practices

This guide outlines the essential practices for developing efficient and maintainable CUDA C++ applications. Adhere to these rules to maximize GPU throughput, reduce debugging time, and ensure code quality across projects.

1. Code Organization and Structure

1.1 Robust Error Handling

Always wrap CUDA API calls in an error-checking macro. This prevents silent failures and provides immediate, actionable debugging information.

❌ BAD:

cudaMalloc(&d_data, size);
// ... potentially crash later without knowing why

✅ GOOD:

#define CUDA_CHECK(call)                                                          \
    do {                                                                          \
        cudaError_t err = call;                                                   \
        if (err != cudaSuccess) {                                                 \
            fprintf(stderr, "CUDA Error: %s:%d: %s\n", __FILE__, __LINE__, cudaGetErrorString(err)); \
            exit(EXIT_FAILURE);                                                   \
        }                                                                         \
    } while (0)

// Usage:
CUDA_CHECK(cudaMalloc(&d_data, size));
kernel<<<grid, block>>>(d_data);
CUDA_CHECK(cudaGetLastError()); // Check for kernel launch errors

1.2 Host-Device Separation

Clearly separate host (CPU) orchestration logic from device (GPU) computation.

❌ BAD:

// host_code.cu
void processData() {
    cudaMalloc(&d_data, size); // Mixed host/device
    kernel<<<...>>>(d_data);
}

✅ GOOD:

// host_manager.cpp
void allocateAndLaunch(float* h_in, int N) {
    float *d_in;
    CUDA_CHECK(cudaMalloc((void**)&d_in, N * sizeof(float)));
    CUDA_CHECK(cudaMemcpy(d_in, h_in, N * sizeof(float), cudaMemcpyHostToDevice));
    myKernel_kernel<<<N/256 + 1, 256>>>(d_in, N);
    CUDA_CHECK(cudaGetLastError());
    CUDA_CHECK(cudaFree(d_in));
}

// device_kernels.cu
__global__ void myKernel_kernel(float* d_in, int N) {
    int idx = blockIdx.x * blockDim.x + threadIdx.x;
    if (idx < N) { d_in[idx] *= 2.0f; }
}

Read the full file on GitHub · 286 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 286 lines · 0 tokens per session scan A 7650ffc664ca

Subscribe to this mod's changes

cuda is a cursor rule published in the GitHub repository sanjeed5/awesome-cursor-rules-mdc (3,571 stars, last pushed 3mo ago), licensed CC0-1.0. It adds 2,336 tokens to every session, about $0.0117 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.