Agent skill

opencl-runtime

Cross-vendor OpenCL runtime management and kernel development. Query platforms/devices, generate portable OpenCL C kernel code, handle vendor-specific extensions, manage contexts and command queues, compile and cache programs.

View SKILL.md on GitHub Repository
Stars 514
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/a5c-ai/babysitter/tree/main/library/specializations/gpu-programming/skills/opencl-runtime

Metadata

Additional technical details for this skill

author
babysitter-sdk
version
1.0.0
category
opencl-development
backlog id
SK-003

SKILL.md

opencl-runtime

You are opencl-runtime - a specialized skill for cross-vendor OpenCL runtime management and kernel development. This skill provides expert capabilities for portable GPU programming across NVIDIA, AMD, and Intel platforms.

Overview

This skill enables AI-powered OpenCL development operations including:

  • Query and enumerate OpenCL platforms/devices
  • Generate portable OpenCL C kernel code
  • Handle vendor-specific extensions and workarounds
  • Manage OpenCL contexts and command queues
  • Compile and cache OpenCL programs/binaries
  • Configure NDRange and work-group dimensions
  • Validate OpenCL memory object usage
  • Support OpenCL 1.2, 2.0, and 3.0 specifications

Prerequisites

  • OpenCL SDK (NVIDIA, AMD, or Intel)
  • OpenCL ICD Loader
  • OpenCL-capable GPU or CPU
  • clinfo utility (for device enumeration)

Capabilities

1. Platform and Device Enumeration

Query available OpenCL resources:

c
// Query platforms
cl_uint numPlatforms;
clGetPlatformIDs(0, NULL, &numPlatforms);
cl_platform_id* platforms = malloc(numPlatforms * sizeof(cl_platform_id));
clGetPlatformIDs(numPlatforms, platforms, NULL);

// Get platform info
char platformName[128];
clGetPlatformInfo(platforms[0], CL_PLATFORM_NAME, 128, platformName, NULL);

// Query devices
cl_uint numDevices;
clGetDeviceIDs(platforms[0], CL_DEVICE_TYPE_GPU, 0, NULL, &numDevices);
cl_device_id* devices = malloc(numDevices * sizeof(cl_device_id));
clGetDeviceIDs(platforms[0], CL_DEVICE_TYPE_GPU, numDevices, devices, NULL);
bash
# Using clinfo utility
clinfo --list

# Detailed device info
clinfo -a

2. OpenCL Kernel Code Generation

Generate portable kernels:

opencl
// Basic kernel pattern
__kernel void vectorAdd(
    __global const float* a,
    __global const float* b,
    __global float* c,
    const int n)
{
    int gid = get_global_id(0);
    if (gid < n) {
        c[gid] = a[gid] + b[gid];
    }
}

// 2D kernel pattern
__kernel void matrixMultiply(
    __global const float* A,
    __global const float* B,
    __global float* C,
    const int M, const int N, const int K)
{
    int row = get_global_id(0);
    int col = get_global_id(1);

    if (row < M && col < N) {
        float sum = 0.0f;
        for (int k = 0; k < K; k++) {
            sum += A[row * K + k] * B[k * N + col];
        }
        C[row * N + col] = sum;
    }
}

// Shared memory (local memory) kernel
__kernel void reductionSum(
    __global const float* input,
    __global float* output,
    __local float* localData,
    const int n)
{
    int gid = get_global_id(0);
    int lid = get_local_id(0);
    int groupSize = get_local_size(0);

    localData[lid] = (gid < n) ? input[gid] : 0.0f;
    barrier(CLK_LOCAL_MEM_FENCE);

    for (int stride = groupSize / 2; stride > 0; stride >>= 1) {
        if (lid < stride) {
            localData[lid] += localData[lid + stride];
        }
        barrier(CLK_LOCAL_MEM_FENCE);
    }

    if (lid == 0) {
        output[get_group_id(0)] = localData[0];
    }
}

3. Context and Command Queue Management

Create and manage OpenCL contexts:

c
// Create context
cl_context context = clCreateContext(NULL, 1, &device, NULL, NULL, &err);

// Create command queue (OpenCL 1.x)
cl_command_queue queue = clCreateCommandQueue(context, device,
    CL_QUEUE_PROFILING_ENABLE, &err);

// Create command queue (OpenCL 2.0+)
cl_queue_properties props[] = {
    CL_QUEUE_PROPERTIES, CL_QUEUE_PROFILING_ENABLE | CL_QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE,
    0
};
cl_command_queue queue = clCreateCommandQueueWithProperties(context, device, props, &err);

4. Program Compilation and Caching

Compile and cache OpenCL programs:

c
// Create program from source
const char* source = loadKernelSource("kernel.cl");
cl_program program = clCreateProgramWithSource(context, 1, &source, NULL, &err);

// Build with options
const char* options = "-cl-fast-relaxed-math -cl-mad-enable";
err = clBuildProgram(program, 1, &device, options, NULL, NULL);

// Get build log on error
if (err != CL_SUCCESS) {
    size_t logSize;
    clGetProgramBuildInfo(program, device, CL_PROGRAM_BUILD_LOG, 0, NULL, &logSize);
    char* log = malloc(logSize);
    clGetProgramBuildInfo(program, device, CL_PROGRAM_BUILD_LOG, logSize, log, NULL);
    printf("Build error:\n%s\n", log);
    free(log);
}

// Get compiled binary for caching
size_t binarySize;
clGetProgramInfo(program, CL_PROGRAM_BINARY_SIZES, sizeof(size_t), &binarySize, NULL);
unsigned char* binary = malloc(binarySize);
clGetProgramInfo(program, CL_PROGRAM_BINARIES, sizeof(unsigned char*), &binary, NULL);
saveBinaryToFile("kernel.bin", binary, binarySize);

// Load cached binary
cl_program programFromBinary = clCreateProgramWithBinary(
    context, 1, &device, &binarySize, (const unsigned char**)&binary, NULL, &err);

5. NDRange Configuration

Configure work dimensions:

c
// 1D NDRange
size_t globalSize = ((n + 255) / 256) * 256;  // Round up to multiple of work-group size
size_t localSize = 256;
clEnqueueNDRangeKernel(queue, kernel, 1, NULL, &globalSize, &localSize, 0, NULL, NULL);

// 2D NDRange
size_t globalSize2D[2] = {width, height};
size_t localSize2D[2] = {16, 16};
clEnqueueNDRangeKernel(queue, kernel, 2, NULL, globalSize2D, localSize2D, 0, NULL, NULL);

// Query max work-group size
size_t maxWorkGroupSize;
clGetDeviceInfo(device, CL_DEVICE_MAX_WORK_GROUP_SIZE, sizeof(size_t), &maxWorkGroupSize, NULL);

6. Memory Object Management

Create and manage buffers:

c
// Create buffers
cl_mem bufferA = clCreateBuffer(context, CL_MEM_READ_ONLY, size, NULL, &err);
cl_mem bufferB = clCreateBuffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
    size, hostDataB, &err);
cl_mem bufferC = clCreateBuffer(context, CL_MEM_WRITE_ONLY, size, NULL, &err);

// Write to buffer
clEnqueueWriteBuffer(queue, bufferA, CL_TRUE, 0, size, hostDataA, 0, NULL, NULL);

// Read from buffer
clEnqueueReadBuffer(queue, bufferC, CL_TRUE, 0, size, hostResult, 0, NULL, NULL);

// Map buffer for direct access
float* mappedPtr = clEnqueueMapBuffer(queue, bufferA, CL_TRUE, CL_MAP_WRITE,
    0, size, 0, NULL, NULL, &err);
// ... modify data ...
clEnqueueUnmapMemObject(queue, bufferA, mappedPtr, 0, NULL, NULL);

7. Vendor Extensions

Handle vendor-specific features:

c
// Check for extension
char extensions[4096];
clGetDeviceInfo(device, CL_DEVICE_EXTENSIONS, sizeof(extensions), extensions, NULL);

if (strstr(extensions, "cl_khr_fp16")) {
    // Half precision available
}

if (strstr(extensions, "cl_nv_device_attribute_query")) {
    // NVIDIA-specific queries available
    cl_uint smCount;
    clGetDeviceInfo(device, CL_DEVICE_COMPUTE_CAPABILITY_MAJOR_NV,
        sizeof(cl_uint), &smCount, NULL);
}

// AMD-specific
if (strstr(extensions, "cl_amd_device_attribute_query")) {
    cl_uint simdPerCU;
    clGetDeviceInfo(device, CL_DEVICE_SIMD_PER_COMPUTE_UNIT_AMD,
        sizeof(cl_uint), &simdPerCU, NULL);
}

8. OpenCL Version Support

Support multiple OpenCL versions:

c
// Query OpenCL version
char version[128];
clGetDeviceInfo(device, CL_DEVICE_VERSION, sizeof(version), version, NULL);

// OpenCL 2.0+ features
#ifdef CL_VERSION_2_0
// Shared Virtual Memory
cl_device_svm_capabilities svmCaps;
clGetDeviceInfo(device, CL_DEVICE_SVM_CAPABILITIES, sizeof(svmCaps), &svmCaps, NULL);

if (svmCaps & CL_DEVICE_SVM_COARSE_GRAIN_BUFFER) {
    void* svmPtr = clSVMAlloc(context, CL_MEM_READ_WRITE, size, 0);
    clEnqueueSVMMap(queue, CL_TRUE, CL_MAP_WRITE, svmPtr, size, 0, NULL, NULL);
}
#endif

// OpenCL 3.0 optional features
#ifdef CL_VERSION_3_0
cl_device_atomic_capabilities atomicCaps;
clGetDeviceInfo(device, CL_DEVICE_ATOMIC_MEMORY_CAPABILITIES,
    sizeof(atomicCaps), &atomicCaps, NULL);
#endif

Process Integration

This skill integrates with the following processes:

  • opencl-application-development.js - OpenCL application development
  • hip-porting-cross-platform.js - Cross-platform porting

Output Format

json
{
  "operation": "enumerate-devices",
  "status": "success",
  "platforms": [
    {
      "name": "NVIDIA CUDA",
      "version": "OpenCL 3.0 CUDA",
      "devices": [
        {
          "name": "NVIDIA GeForce RTX 4090",
          "type": "GPU",
          "computeUnits": 128,
          "maxWorkGroupSize": 1024,
          "globalMemory": "24 GB",
          "extensions": ["cl_khr_fp16", "cl_khr_fp64"]
        }
      ]
    }
  ]
}

Dependencies

  • OpenCL SDK (NVIDIA, AMD, or Intel)
  • OpenCL ICD Loader
  • clinfo utility

Constraints

  • OpenCL 2.0+ features not available on all platforms
  • Vendor extensions are not portable
  • Binary caching requires same device/driver
  • SVM requires OpenCL 2.0+ and device support

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

a5c-ai/babysitter

gsd-tools

Central utility skill for GSD operations. Provides config parsing, slug generation, timestamps, path operations, and orchestrates calls to other specialized skills. Acts as the unified entry point that the original gsd-tools.cjs provided via its lib/ modules (commands, config, core, init).

514 31
Explore
a5c-ai/babysitter

model-profile-resolution

Resolve model profile (quality/balanced/budget) at orchestration start and map agents to specific models. Enables cost/quality tradeoffs by selecting appropriate AI models for each agent role.

514 31
Explore
a5c-ai/babysitter

verification-suite

Plan structure validation, phase completeness checks, reference integrity verification, and artifact existence confirmation. Provides the structured verification layer ensuring GSD artifacts are well-formed and complete.

514 31
Explore
a5c-ai/babysitter

state-management

STATE.md reading, writing, and field-level updates. Provides cross-session state persistence via .planning/STATE.md with structured fields for current task, completed phases, blockers, decisions, and quick tasks.

514 31
Explore
a5c-ai/babysitter

git-integration

Git commit patterns, formats, and conventions for GSD methodology. Provides atomic commits per task, structured commit messages, planning file commits, branch management, and milestone tag operations.

514 31
Explore
a5c-ai/babysitter

frontmatter-parsing

YAML frontmatter parsing and manipulation for .planning/ documents. Provides read, write, update, query, and validation operations on frontmatter blocks in GSD markdown artifacts.

514 31
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results