cuda_nsight_compute_project.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ubnGE0v05Eg7",
        "outputId": "412244b9-027b-425a-9d71-fce76c86a857"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Tue May  7 17:20:45 2024       \n",
            "+---------------------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 535.104.05             Driver Version: 535.104.05   CUDA Version: 12.2     |\n",
            "|-----------------------------------------+----------------------+----------------------+\n",
            "| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                                         |                      |               MIG M. |\n",
            "|=========================================+======================+======================|\n",
            "|   0  Tesla T4                       Off | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   52C    P8              10W /  70W |      0MiB / 15360MiB |      0%      Default |\n",
            "|                                         |                      |                  N/A |\n",
            "+-----------------------------------------+----------------------+----------------------+\n",
            "                                                                                         \n",
            "+---------------------------------------------------------------------------------------+\n",
            "| Processes:                                                                            |\n",
            "|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |\n",
            "|        ID   ID                                                             Usage      |\n",
            "|=======================================================================================|\n",
            "|  No running processes found                                                           |\n",
            "+---------------------------------------------------------------------------------------+\n"
          ]
        }
      ],
      "source": [
        "!nvidia-smi\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvcc --version\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5ArpSg_65HC8",
        "outputId": "2d6118e6-5b77-4db9-f50b-83a87ac87841"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "nvcc: NVIDIA (R) Cuda compiler driver\n",
            "Copyright (c) 2005-2023 NVIDIA Corporation\n",
            "Built on Tue_Aug_15_22:02:13_PDT_2023\n",
            "Cuda compilation tools, release 12.2, V12.2.140\n",
            "Build cuda_12.2.r12.2/compiler.33191640_0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "nPV-5PbUFCRB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%%cu\n",
        "#include <iostream>\n",
        "\n",
        "__global__ void helloCUDA()\n",
        "{\n",
        "    printf(\"Hello CUDA from GPU!\\n\");\n",
        "}\n",
        "\n",
        "int main()\n",
        "{\n",
        "    helloCUDA<<<1, 1>>>();\n",
        "    cudaDeviceSynchronize();\n",
        "    return 0;\n",
        "}\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "82BC3d8N5Mip",
        "outputId": "0a8a6e54-0eb9-461c-cb10-97e6a6c5a5b5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "UsageError: Cell magic `%%cu` not found.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%%writefile cuda_code.cu\n",
        "\n",
        "#include <iostream>\n",
        "\n",
        "__global__ void helloCUDA()\n",
        "{\n",
        "    printf(\"Hello CUDA from GPU!\\n\");\n",
        "}\n",
        "\n",
        "int main()\n",
        "{\n",
        "    helloCUDA<<<1, 1>>>();\n",
        "    cudaDeviceSynchronize();\n",
        "    return 0;\n",
        "}\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ncOO6YiC5SSd",
        "outputId": "6c923231-06c0-4c64-f48f-905d686cdb9e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Writing cuda_code.cu\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvcc -o cuda_code_cuda cuda_code.cu\n",
        "!./cuda_code_cuda\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HG_cxWUd5aog",
        "outputId": "4b9e50e6-51a9-4f7d-9f78-64c4b2ac46f8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Hello CUDA from GPU!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%%writefile cuda_code_1.cu\n",
        "\n",
        "#include <iostream>\n",
        "#include <vector>\n",
        "#include <cstdlib>\n",
        "#include <cuda_runtime.h>\n",
        "#include <cublas_v2.h>\n",
        "\n",
        "int main() {\n",
        "    const int batch_size = 128;\n",
        "    const int input_size = 256;\n",
        "    const int output_size = 10;  // 10 output classes for MNIST\n",
        "\n",
        "    // Allocate host memory for input and weights\n",
        "    std::vector<float> input(batch_size * input_size);\n",
        "    std::vector<float> weights(input_size * output_size);\n",
        "\n",
        "    // Initialize input and weights with random values\n",
        "    for (int i = 0; i < batch_size * input_size; ++i) {\n",
        "        input[i] = static_cast<float>(rand()) / RAND_MAX;\n",
        "    }\n",
        "    for (int i = 0; i < input_size * output_size; ++i) {\n",
        "        weights[i] = static_cast<float>(rand()) / RAND_MAX;\n",
        "    }\n",
        "\n",
        "    // Allocate device memory\n",
        "    float *d_input, *d_weights, *d_output;\n",
        "    cudaMalloc(&d_input, batch_size * input_size * sizeof(float));\n",
        "    cudaMalloc(&d_weights, input_size * output_size * sizeof(float));\n",
        "    cudaMalloc(&d_output, batch_size * output_size * sizeof(float));\n",
        "\n",
        "    // Copy input and weights to device\n",
        "    cudaMemcpy(d_input, input.data(), batch_size * input_size * sizeof(float), cudaMemcpyHostToDevice);\n",
        "    cudaMemcpy(d_weights, weights.data(), input_size * output_size * sizeof(float), cudaMemcpyHostToDevice);\n",
        "\n",
        "    // Create cuBLAS handle\n",
        "    cublasHandle_t handle;\n",
        "    cublasCreate(&handle);\n",
        "\n",
        "    // Perform vector multiplication\n",
        "    float alpha = 1.0f, beta = 0.0f;\n",
        "    cublasSgemv(handle, CUBLAS_OP_N, output_size, input_size, &alpha,\n",
        "                d_weights, output_size, d_input, 1, &beta, d_output, 1);\n",
        "\n",
        "    // Destroy cuBLAS handle\n",
        "    cublasDestroy(handle);\n",
        "\n",
        "    // Copy output back to host\n",
        "    std::vector<float> output(batch_size * output_size);\n",
        "    cudaMemcpy(output.data(), d_output, batch_size * output_size * sizeof(float), cudaMemcpyDeviceToHost);\n",
        "\n",
        "    // Free device memory\n",
        "    cudaFree(d_input);\n",
        "    cudaFree(d_weights);\n",
        "    cudaFree(d_output);\n",
        "\n",
        "    // Print a sample output\n",
        "    std::cout << \"Sample output:\" << std::endl;\n",
        "    for (int i = 0; i < 10; ++i) {\n",
        "        std::cout << output[i] << \" \";\n",
        "    }\n",
        "    std::cout << std::endl;\n",
        "\n",
        "    return 0;\n",
        "}"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R0q5dexbbm3y",
        "outputId": "db9d6190-acf6-4fea-a422-60c2c80f821a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Writing cuda_code_1.cu\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvcc -o cuda_code_1_cuda cuda_code_1.cu\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "E1saqARzcEko",
        "outputId": "65a109d3-d255-4874-bd0e-b7b2b5c61013"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/usr/bin/ld: /tmp/tmpxft_000008d5_00000000-11_cuda_code_1.o: in function `main':\n",
            "tmpxft_000008d5_00000000-6_cuda_code_1.cudafe1.cpp:(.text+0x20a): undefined reference to `cublasCreate_v2'\n",
            "/usr/bin/ld: tmpxft_000008d5_00000000-6_cuda_code_1.cudafe1.cpp:(.text+0x270): undefined reference to `cublasSgemv_v2'\n",
            "/usr/bin/ld: tmpxft_000008d5_00000000-6_cuda_code_1.cudafe1.cpp:(.text+0x280): undefined reference to `cublasDestroy_v2'\n",
            "collect2: error: ld returned 1 exit status\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "os.environ['CUDA_VISIBLE_DEVICES'] = '0'  # Use GPU 0"
      ],
      "metadata": {
        "id": "-VMB6wTVdXPQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!nvidia-smi"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uOLeImOwd1yE",
        "outputId": "98075213-0e73-4969-9242-1992bd6ee44f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mon Apr 29 20:06:35 2024       \n",
            "+---------------------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 535.104.05             Driver Version: 535.104.05   CUDA Version: 12.2     |\n",
            "|-----------------------------------------+----------------------+----------------------+\n",
            "| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                                         |                      |               MIG M. |\n",
            "|=========================================+======================+======================|\n",
            "|   0  Tesla T4                       Off | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   34C    P8               8W /  70W |      0MiB / 15360MiB |      0%      Default |\n",
            "|                                         |                      |                  N/A |\n",
            "+-----------------------------------------+----------------------+----------------------+\n",
            "                                                                                         \n",
            "+---------------------------------------------------------------------------------------+\n",
            "| Processes:                                                                            |\n",
            "|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |\n",
            "|        ID   ID                                                             Usage      |\n",
            "|=======================================================================================|\n",
            "|  No running processes found                                                           |\n",
            "+---------------------------------------------------------------------------------------+\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ldconfig /usr/local/cuda-11.8/targets/x86_64-linux/lib/\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Mlutg6fodXSn",
        "outputId": "38fa4d02-bc85-410d-a7f5-7105b7d02775"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind_2_0.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind_2_5.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbb.so.12 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbmalloc.so.2 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbmalloc_proxy.so.2 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind.so.3 is not a symbolic link\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ldconfig /usr/local/cuda/targets/x86_64-linux/lib/\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L7f9xt86d76G",
        "outputId": "75313d0a-74fe-4ddc-d193-089a81ee2e60"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind_2_0.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind_2_5.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbb.so.12 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbmalloc.so.2 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbmalloc_proxy.so.2 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind.so.3 is not a symbolic link\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%%cuda --compiler-options '-lcublas -lcudart -lcudnn -lcurand' --include-paths '/usr/local/cuda/targets/x86_64-linux/include/' --libraries '/usr/local/cuda/targets/x86_64-linux/lib/'"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-8Od8P83d79S",
        "outputId": "1af6f051-00ee-4f07-9858-28d0162e3505"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "UsageError: Cell magic `%%cuda` not found.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%%writefile cuda_code_1.cu\n",
        "\n",
        "#include <iostream>\n",
        "#include <vector>\n",
        "#include <cstdlib>\n",
        "#include <cuda_runtime.h>\n",
        "#include <cublas_v2.h>\n",
        "\n",
        "int main() {\n",
        "    const int batch_size = 128;\n",
        "    const int input_size = 256;\n",
        "    const int output_size = 10;  // 10 output classes for MNIST\n",
        "\n",
        "    // Allocate host memory for input and weights\n",
        "    std::vector<float> input(batch_size * input_size);\n",
        "    std::vector<float> weights(input_size * output_size);\n",
        "\n",
        "    // Initialize input and weights with random values\n",
        "    for (int i = 0; i < batch_size * input_size; ++i) {\n",
        "        input[i] = static_cast<float>(rand()) / RAND_MAX;\n",
        "    }\n",
        "    for (int i = 0; i < input_size * output_size; ++i) {\n",
        "        weights[i] = static_cast<float>(rand()) / RAND_MAX;\n",
        "    }\n",
        "\n",
        "    // Allocate device memory\n",
        "    float *d_input, *d_weights, *d_output;\n",
        "    cudaMalloc(&d_input, batch_size * input_size * sizeof(float));\n",
        "    cudaMalloc(&d_weights, input_size * output_size * sizeof(float));\n",
        "    cudaMalloc(&d_output, batch_size * output_size * sizeof(float));\n",
        "\n",
        "    // Copy input and weights to device\n",
        "    cudaMemcpy(d_input, input.data(), batch_size * input_size * sizeof(float), cudaMemcpyHostToDevice);\n",
        "    cudaMemcpy(d_weights, weights.data(), input_size * output_size * sizeof(float), cudaMemcpyHostToDevice);\n",
        "\n",
        "    // Create cuBLAS handle\n",
        "    cublasHandle_t handle;\n",
        "    cublasCreate(&handle);\n",
        "\n",
        "    // Perform vector multiplication\n",
        "    float alpha = 1.0f, beta = 0.0f;\n",
        "    cublasSgemv(handle, CUBLAS_OP_N, output_size, input_size, &alpha,\n",
        "                d_weights, output_size, d_input, 1, &beta, d_output, 1);\n",
        "\n",
        "    // Destroy cuBLAS handle\n",
        "    cublasDestroy(handle);\n",
        "\n",
        "    // Copy output back to host\n",
        "    std::vector<float> output(batch_size * output_size);\n",
        "    cudaMemcpy(output.data(), d_output, batch_size * output_size * sizeof(float), cudaMemcpyDeviceToHost);\n",
        "\n",
        "    // Free device memory\n",
        "    cudaFree(d_input);\n",
        "    cudaFree(d_weights);\n",
        "    cudaFree(d_output);\n",
        "\n",
        "    // Print a sample output\n",
        "    std::cout << \"Sample output:\" << std::endl;\n",
        "    for (int i = 0; i < 10; ++i) {\n",
        "        std::cout << output[i] << \" \";\n",
        "    }\n",
        "    std::cout << std::endl;\n",
        "\n",
        "    return 0;\n",
        "}\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uhPEi-bHeUd9",
        "outputId": "6c0dccd5-5db3-4af8-8610-22275c53b4c9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overwriting cuda_code_1.cu\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvcc -o cuda_executable cuda_code_1.cu -lcublas\n"
      ],
      "metadata": {
        "id": "1NgSe85Zd7_I"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!./cuda_executable\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KAk0Yuzfd8Cf",
        "outputId": "befde94b-de70-42b8-85f1-fb2508450d13"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Sample output:\n",
            "66.7555 68.8272 68.0738 61.9798 67.4468 64.2662 68.7616 64.612 60.7305 67.7168 \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import time\n",
        "import subprocess\n",
        "\n",
        "# Compile CUDA code\n",
        "os.system('nvcc -o cuda_executable cuda_code_1.cu')\n",
        "\n",
        "# Record start time\n",
        "start_time = time.time()\n",
        "\n",
        "# Execute compiled CUDA code\n",
        "os.system('./cuda_executable')\n",
        "\n",
        "# Record end time\n",
        "end_time = time.time()\n",
        "\n",
        "# Calculate execution time\n",
        "execution_time = end_time - start_time\n",
        "print(\"Execution time:\", execution_time, \"seconds\")\n",
        "\n",
        "# Calculate memory usage\n",
        "process = subprocess.Popen(['nvidia-smi', '--query-gpu=memory.used', '--format=csv,nounits,noheader'],\n",
        "                           stdout=subprocess.PIPE, stderr=subprocess.PIPE)\n",
        "stdout, _ = process.communicate()\n",
        "memory_used = sum(int(x) for x in stdout.strip().split(b'\\n'))\n",
        "print(\"Memory used:\", memory_used, \"MiB\")\n"
      ],
      "metadata": {
        "id": "u_VLBGhkn4s3",
        "outputId": "6a7adc70-2ba0-4180-ded4-679495e11f70",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Execution time: 0.001447916030883789 seconds\n",
            "Memory used: 0 MiB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://raw.githubusercontent.com/cteqeu/HAC/main/GPU/cuda_profile.cu"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mmD5-X6kFDjx",
        "outputId": "55126e57-cd2b-488d-b33b-58fd35d7e233"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-05-07 17:25:22--  https://raw.githubusercontent.com/cteqeu/HAC/main/GPU/cuda_profile.cu\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.110.133, 185.199.108.133, 185.199.109.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.110.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 2331 (2.3K) [text/plain]\n",
            "Saving to: ‘cuda_profile.cu’\n",
            "\n",
            "\rcuda_profile.cu       0%[                    ]       0  --.-KB/s               \rcuda_profile.cu     100%[===================>]   2.28K  --.-KB/s    in 0s      \n",
            "\n",
            "2024-05-07 17:25:22 (54.5 MB/s) - ‘cuda_profile.cu’ saved [2331/2331]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvidia-smi"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4tuvmdxEF2Fr",
        "outputId": "b48e4570-ced2-4313-844e-0078533c1791"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Tue May  7 17:55:02 2024       \n",
            "+---------------------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 535.104.05             Driver Version: 535.104.05   CUDA Version: 12.2     |\n",
            "|-----------------------------------------+----------------------+----------------------+\n",
            "| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                                         |                      |               MIG M. |\n",
            "|=========================================+======================+======================|\n",
            "|   0  Tesla T4                       Off | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   38C    P8               9W /  70W |      0MiB / 15360MiB |      0%      Default |\n",
            "|                                         |                      |                  N/A |\n",
            "+-----------------------------------------+----------------------+----------------------+\n",
            "                                                                                         \n",
            "+---------------------------------------------------------------------------------------+\n",
            "| Processes:                                                                            |\n",
            "|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |\n",
            "|        ID   ID                                                             Usage      |\n",
            "|=======================================================================================|\n",
            "|  No running processes found                                                           |\n",
            "+---------------------------------------------------------------------------------------+\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvcc --version"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VwWLhaNIF2JM",
        "outputId": "5a068647-b2d4-4fed-cdc0-45b21d0d19ff"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "nvcc: NVIDIA (R) Cuda compiler driver\n",
            "Copyright (c) 2005-2023 NVIDIA Corporation\n",
            "Built on Tue_Aug_15_22:02:13_PDT_2023\n",
            "Cuda compilation tools, release 12.2, V12.2.140\n",
            "Build cuda_12.2.r12.2/compiler.33191640_0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvcc cuda_profile.cu -o cuda_profile --generate-line-info"
      ],
      "metadata": {
        "id": "gdYzcvinF2RY"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!compute-sanitizer ./cuda_profile"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k9P8_krwF2Tx",
        "outputId": "3313ab72-8231-4e9b-8462-407a5a5cfe8a"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "========= COMPUTE-SANITIZER\n",
            "cpu control: \t\t18108465020928.000000\n",
            "gpu control: \t\t18108465020928.000000\n",
            "gpu um control: \t18108465020928.000000\n",
            "========= ERROR SUMMARY: 0 errors\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvprof --print-gpu-trace ./cuda_profile"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VDWVyKdlF2XK",
        "outputId": "171355a1-1c7c-423c-9c99-ee937174b0e4"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "==11364== NVPROF is profiling process 11364, command: ./cuda_profile\n",
            "cpu control: \t\t18108465020928.000000\n",
            "gpu control: \t\t18108465020928.000000\n",
            "gpu um control: \t18108465020928.000000\n",
            "==11364== Profiling application: ./cuda_profile\n",
            "==11364== Profiling result:\n",
            "   Start  Duration            Grid Size      Block Size     Regs*    SSMem*    DSMem*      Size  Throughput  SrcMemType  DstMemType           Device   Context    Stream        Unified Memory  Virtual Address  Name\n",
            "293.49ms  35.359us                    -               -         -         -         -  390.63KB  10.536GB/s    Pageable      Device     Tesla T4 (0)         1         7                     -                -  [CUDA memcpy HtoD]\n",
            "293.62ms  34.720us                    -               -         -         -         -  390.63KB  10.730GB/s    Pageable      Device     Tesla T4 (0)         1         7                     -                -  [CUDA memcpy HtoD]\n",
            "293.98ms  7.4880us             (98 1 1)      (1024 1 1)        16        0B        0B         -           -           -           -     Tesla T4 (0)         1         7                     -                -  gpu_testfunction(float*, float*, float*, int) [133]\n",
            "294.00ms  32.383us                    -               -         -         -         -  390.63KB  11.504GB/s      Device    Pageable     Tesla T4 (0)         1         7                     -                -  [CUDA memcpy DtoH]\n",
            "294.46ms  615.71us             (98 1 1)      (1024 1 1)        16        0B        0B         -           -           -           -     Tesla T4 (0)         1         7                     -                -  gpu_testfunction_uni_mem(float*, float*, float*, int) [135]\n",
            "294.47ms  447.74us                    -               -         -         -         -         -           -           -           -     Tesla T4 (0)         -         -                     4   0x7f0ef2000000  [Unified Memory GPU page faults]\n",
            "294.91ms  77.279us                    -               -         -         -         -         -           -           -           -     Tesla T4 (0)         -         -                     1   0x7f0ef20c4000  [Unified Memory GPU page faults]\n",
            "295.00ms  77.567us                    -               -         -         -         -         -           -           -           -     Tesla T4 (0)         -         -                     8   0x7f0ef2100000  [Unified Memory GPU page faults]\n",
            "\n",
            "Regs: Number of registers used per CUDA thread. This number includes registers used internally by the CUDA driver and/or tools and can be more than what the compiler shows.\n",
            "SSMem: Static shared memory allocated per CUDA block.\n",
            "DSMem: Dynamic shared memory allocated per CUDA block.\n",
            "SrcMemType: The type of source memory accessed by memory operation/copy\n",
            "DstMemType: The type of destination memory accessed by memory operation/copy\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nsys profile cuda_profile"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dIVEMTFrOKit",
        "outputId": "1c48f1a1-2176-42bc-dd5f-1ec922e08b0e"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/bin/bash: line 1: nsys: command not found\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "0WrAwQjUO4oE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/nsight-systems-2023.2.3_2023.2.3.1001-1_amd64.deb\n",
        "!apt update\n",
        "!apt install ./nsight-systems-2023.2.3_2023.2.3.1001-1_amd64.deb\n",
        "!apt --fix-broken install"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qew0cKciOs9O",
        "outputId": "df215308-259f-4112-fd8e-cb489f18374c"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-05-07 18:04:53--  https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/nsight-systems-2023.2.3_2023.2.3.1001-1_amd64.deb\n",
            "Resolving developer.download.nvidia.com (developer.download.nvidia.com)... 152.195.19.142\n",
            "Connecting to developer.download.nvidia.com (developer.download.nvidia.com)|152.195.19.142|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 317705436 (303M) [application/x-deb]\n",
            "Saving to: ‘nsight-systems-2023.2.3_2023.2.3.1001-1_amd64.deb’\n",
            "\n",
            "nsight-systems-2023 100%[===================>] 302.99M  18.9MB/s    in 5.0s    \n",
            "\n",
            "2024-05-07 18:04:58 (60.1 MB/s) - ‘nsight-systems-2023.2.3_2023.2.3.1001-1_amd64.deb’ saved [317705436/317705436]\n",
            "\n",
            "Get:1 https://cloud.r-project.org/bin/linux/ubuntu jammy-cran40/ InRelease [3,626 B]\n",
            "Hit:2 https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64  InRelease\n",
            "Get:3 http://security.ubuntu.com/ubuntu jammy-security InRelease [110 kB]\n",
            "Hit:4 http://archive.ubuntu.com/ubuntu jammy InRelease\n",
            "Get:5 http://archive.ubuntu.com/ubuntu jammy-updates InRelease [119 kB]\n",
            "Hit:6 https://ppa.launchpadcontent.net/c2d4u.team/c2d4u4.0+/ubuntu jammy InRelease\n",
            "Hit:7 https://ppa.launchpadcontent.net/deadsnakes/ppa/ubuntu jammy InRelease\n",
            "Get:8 http://archive.ubuntu.com/ubuntu jammy-backports InRelease [109 kB]\n",
            "Hit:9 https://ppa.launchpadcontent.net/graphics-drivers/ppa/ubuntu jammy InRelease\n",
            "Get:10 http://security.ubuntu.com/ubuntu jammy-security/main amd64 Packages [1,789 kB]\n",
            "Hit:11 https://ppa.launchpadcontent.net/ubuntugis/ppa/ubuntu jammy InRelease\n",
            "Get:12 http://archive.ubuntu.com/ubuntu jammy-updates/restricted amd64 Packages [2,382 kB]\n",
            "Get:13 http://security.ubuntu.com/ubuntu jammy-security/restricted amd64 Packages [2,308 kB]\n",
            "Get:14 http://archive.ubuntu.com/ubuntu jammy-updates/main amd64 Packages [2,069 kB]\n",
            "Get:15 http://archive.ubuntu.com/ubuntu jammy-updates/universe amd64 Packages [1,371 kB]\n",
            "Get:16 http://archive.ubuntu.com/ubuntu jammy-backports/universe amd64 Packages [33.3 kB]\n",
            "Fetched 10.3 MB in 2s (5,349 kB/s)\n",
            "Reading package lists... Done\n",
            "Building dependency tree... Done\n",
            "Reading state information... Done\n",
            "46 packages can be upgraded. Run 'apt list --upgradable' to see them.\n",
            "Reading package lists... Done\n",
            "Building dependency tree... Done\n",
            "Reading state information... Done\n",
            "Note, selecting 'nsight-systems-2023.2.3' instead of './nsight-systems-2023.2.3_2023.2.3.1001-1_amd64.deb'\n",
            "The following additional packages will be installed:\n",
            "  libtinfo5 libxcb-icccm4 libxcb-image0 libxcb-keysyms1 libxcb-render-util0 libxcb-util1\n",
            "  libxcb-xinerama0 libxcb-xinput0 libxcb-xkb1 libxkbcommon-x11-0 libxtst6\n",
            "The following NEW packages will be installed:\n",
            "  libtinfo5 libxcb-icccm4 libxcb-image0 libxcb-keysyms1 libxcb-render-util0 libxcb-util1\n",
            "  libxcb-xinerama0 libxcb-xinput0 libxcb-xkb1 libxkbcommon-x11-0 libxtst6 nsight-systems-2023.2.3\n",
            "0 upgraded, 12 newly installed, 0 to remove and 46 not upgraded.\n",
            "Need to get 318 MB of archives.\n",
            "After this operation, 1,269 kB of additional disk space will be used.\n",
            "Get:1 https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64  nsight-systems-2023.2.3 2023.2.3.1001-32894139v0 [318 MB]\n",
            "Get:2 http://archive.ubuntu.com/ubuntu jammy-updates/universe amd64 libtinfo5 amd64 6.3-2ubuntu0.1 [100 kB]\n",
            "Get:3 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-xinerama0 amd64 1.14-3ubuntu3 [5,414 B]\n",
            "Get:4 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-xinput0 amd64 1.14-3ubuntu3 [34.3 kB]\n",
            "Get:5 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-xkb1 amd64 1.14-3ubuntu3 [32.8 kB]\n",
            "Get:6 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxkbcommon-x11-0 amd64 1.4.0-1 [14.4 kB]\n",
            "Get:7 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxtst6 amd64 2:1.2.3-1build4 [13.4 kB]\n",
            "Get:8 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-icccm4 amd64 0.4.1-1.1build2 [11.5 kB]\n",
            "Get:9 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-util1 amd64 0.4.0-1build2 [11.4 kB]\n",
            "Get:10 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-image0 amd64 0.4.0-2 [11.5 kB]\n",
            "Get:11 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-keysyms1 amd64 0.4.0-1build3 [8,746 B]\n",
            "Get:12 http://archive.ubuntu.com/ubuntu jammy/main amd64 libxcb-render-util0 amd64 0.3.9-1build3 [10.3 kB]\n",
            "Fetched 318 MB in 4s (79.7 MB/s)\n",
            "Selecting previously unselected package libtinfo5:amd64.\n",
            "(Reading database ... 121920 files and directories currently installed.)\n",
            "Preparing to unpack .../00-libtinfo5_6.3-2ubuntu0.1_amd64.deb ...\n",
            "Unpacking libtinfo5:amd64 (6.3-2ubuntu0.1) ...\n",
            "Selecting previously unselected package libxcb-xinerama0:amd64.\n",
            "Preparing to unpack .../01-libxcb-xinerama0_1.14-3ubuntu3_amd64.deb ...\n",
            "Unpacking libxcb-xinerama0:amd64 (1.14-3ubuntu3) ...\n",
            "Selecting previously unselected package libxcb-xinput0:amd64.\n",
            "Preparing to unpack .../02-libxcb-xinput0_1.14-3ubuntu3_amd64.deb ...\n",
            "Unpacking libxcb-xinput0:amd64 (1.14-3ubuntu3) ...\n",
            "Selecting previously unselected package libxcb-xkb1:amd64.\n",
            "Preparing to unpack .../03-libxcb-xkb1_1.14-3ubuntu3_amd64.deb ...\n",
            "Unpacking libxcb-xkb1:amd64 (1.14-3ubuntu3) ...\n",
            "Selecting previously unselected package libxkbcommon-x11-0:amd64.\n",
            "Preparing to unpack .../04-libxkbcommon-x11-0_1.4.0-1_amd64.deb ...\n",
            "Unpacking libxkbcommon-x11-0:amd64 (1.4.0-1) ...\n",
            "Selecting previously unselected package libxtst6:amd64.\n",
            "Preparing to unpack .../05-libxtst6_2%3a1.2.3-1build4_amd64.deb ...\n",
            "Unpacking libxtst6:amd64 (2:1.2.3-1build4) ...\n",
            "Selecting previously unselected package libxcb-icccm4:amd64.\n",
            "Preparing to unpack .../06-libxcb-icccm4_0.4.1-1.1build2_amd64.deb ...\n",
            "Unpacking libxcb-icccm4:amd64 (0.4.1-1.1build2) ...\n",
            "Selecting previously unselected package libxcb-util1:amd64.\n",
            "Preparing to unpack .../07-libxcb-util1_0.4.0-1build2_amd64.deb ...\n",
            "Unpacking libxcb-util1:amd64 (0.4.0-1build2) ...\n",
            "Selecting previously unselected package libxcb-image0:amd64.\n",
            "Preparing to unpack .../08-libxcb-image0_0.4.0-2_amd64.deb ...\n",
            "Unpacking libxcb-image0:amd64 (0.4.0-2) ...\n",
            "Selecting previously unselected package libxcb-keysyms1:amd64.\n",
            "Preparing to unpack .../09-libxcb-keysyms1_0.4.0-1build3_amd64.deb ...\n",
            "Unpacking libxcb-keysyms1:amd64 (0.4.0-1build3) ...\n",
            "Selecting previously unselected package libxcb-render-util0:amd64.\n",
            "Preparing to unpack .../10-libxcb-render-util0_0.3.9-1build3_amd64.deb ...\n",
            "Unpacking libxcb-render-util0:amd64 (0.3.9-1build3) ...\n",
            "Selecting previously unselected package nsight-systems-2023.2.3.\n",
            "Preparing to unpack .../11-nsight-systems-2023.2.3_2023.2.3.1001-32894139v0_amd64.deb ...\n",
            "Unpacking nsight-systems-2023.2.3 (2023.2.3.1001-32894139v0) ...\n",
            "Setting up libxcb-xinput0:amd64 (1.14-3ubuntu3) ...\n",
            "Setting up libxtst6:amd64 (2:1.2.3-1build4) ...\n",
            "Setting up libxcb-keysyms1:amd64 (0.4.0-1build3) ...\n",
            "Setting up libxcb-render-util0:amd64 (0.3.9-1build3) ...\n",
            "Setting up libxcb-icccm4:amd64 (0.4.1-1.1build2) ...\n",
            "Setting up libxcb-util1:amd64 (0.4.0-1build2) ...\n",
            "Setting up libxcb-xkb1:amd64 (1.14-3ubuntu3) ...\n",
            "Setting up libxcb-image0:amd64 (0.4.0-2) ...\n",
            "Setting up libxcb-xinerama0:amd64 (1.14-3ubuntu3) ...\n",
            "Setting up libxkbcommon-x11-0:amd64 (1.4.0-1) ...\n",
            "Setting up libtinfo5:amd64 (6.3-2ubuntu0.1) ...\n",
            "Setting up nsight-systems-2023.2.3 (2023.2.3.1001-32894139v0) ...\n",
            "update-alternatives: using /opt/nvidia/nsight-systems/2023.2.3/target-linux-x64/nsys to provide /usr/local/bin/nsys (nsys) in auto mode\n",
            "update-alternatives: using /opt/nvidia/nsight-systems/2023.2.3/host-linux-x64/nsys-ui to provide /usr/local/bin/nsys-ui (nsys-ui) in auto mode\n",
            "Processing triggers for libc-bin (2.35-0ubuntu3.4) ...\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbb.so.12 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind_2_0.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbmalloc.so.2 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbbind_2_5.so.3 is not a symbolic link\n",
            "\n",
            "/sbin/ldconfig.real: /usr/local/lib/libtbbmalloc_proxy.so.2 is not a symbolic link\n",
            "\n",
            "Reading package lists... Done\n",
            "Building dependency tree... Done\n",
            "Reading state information... Done\n",
            "0 upgraded, 0 newly installed, 0 to remove and 47 not upgraded.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nsys profile cuda_profile"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "srFmIfT7O5bZ",
        "outputId": "feb50772-f190-4ed4-dd39-9a62e7300b69"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "cpu control: \t\t18108465020928.000000\n",
            "gpu control: \t\t18108465020928.000000\n",
            "gpu um control: \t18108465020928.000000\n",
            "Generating '/tmp/nsys-report-bb3f.qdstrm'\n",
            "[1/1] [========================100%] report1.nsys-rep\n",
            "Generated:\n",
            "    /content/report1.nsys-rep\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ncu --set full -o cuda_profile ./cuda_profile"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KHdy1mAQUzrd",
        "outputId": "94ffc424-d5cb-44c7-dc89-a7592b51af25"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "==PROF== Connected to process 19587 (/content/cuda_profile)\n",
            "cpu control: \t\t18108465020928.000000\n",
            "==PROF== Profiling \"gpu_testfunction\" - 0: 0%....50%....100% - 31 passes\n",
            "gpu control: \t\t18108465020928.000000\n",
            "==PROF== Profiling \"gpu_testfunction_uni_mem\" - 1: 0%....50%....100% - 31 passes\n",
            "gpu um control: \t18108465020928.000000\n",
            "==PROF== Disconnected from process 19587\n",
            "==PROF== Report: /content/cuda_profile.ncu-rep\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "_l_h7wohU9lu"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}