• Stars
    star
    510
  • Rank 83,210 (Top 2 %)
  • Language
    C++
  • License
    Other
  • Created almost 4 years ago
  • Updated 7 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Repository for nvCOMP docs and examples. nvCOMP is a library for fast lossless compression/decompression on the GPU that can be downloaded from https://developer.nvidia.com/nvcomp.

What is nvCOMP?

nvCOMP is a CUDA library that features generic compression interfaces to enable developers to use high-performance GPU compressors and decompressors in their applications.

Example benchmarking results and a brief description of each algorithm are available on the nvCOMP Developer Page.

From version 2.3 onwards, the compression / decompression source code will not be released.

Known issues

  • Cascaded, GDeflate, zStandard, Deflate, Gzip and Bitcomp decompressors can only operate on valid input data (data that was compressed using the same compressor). Other decompressors can sometimes detect errors in the compressed stream.
  • Cascaded, zStandard and Bitcomp batched decompression C APIs cannot currently accept nullptr for actual_decompressed_bytes or device_statuses values. Deflate and Gzip cannot accept nullptr for device_statuses values.
  • The Bitcomp low-level batched decompression function is not fully asynchronous.
  • Gzip low-level interface only provides decompression.
  • The device API only supports the ANS format
  • Gdeflate API test and ANS device test is not working on H100 with CTK 12.x, will be excluded in the X86_64 build.

Download

  • You can download the appropriate built binary packages from the nvCOMP Developer Page
  • Choose the package that corresponds to your CUDA toolkit version, operating system, and arch
  • For example, on linux, the package includes
include/ 
  nvcomp/ #nvcomp API headers
  gdeflate/ #Gdeflate CPU library headers
lib/
  libnvcomp.so
  <Other nvcomp libraries that are used internally by nvcomp's APIs>
  libnvcomp_gdeflate_cpu.so # CPU library for gdeflate
  cmake/ <Package files to allow use through cmake>
bin/ 
  <benchmark scripts>

Requirements

  • Pascal (sm60) or higher GPU architecture is required. Volta (sm70)+ GPU architecture is recommended for best results.
  • To compile using nvCOMP as a dependency, you need a compiler with C++ 11 support (e.g. GCC 4.4, Clang 3.3, MSVC 2017).
    • To use the packages provided for Linux, you'll need to compile translation units that will be linked against nvcomp using the old ABI, i.e. for GCC -D_GLIBCXX_USE_CXX11_ABI=0

nvCOMP library API Descriptions

Please view the following guides for information on how to use the two APIs provided by the library. Each of these guides links to a compilable example for further reference.

GPU Benchmarking

GPU Benchmark source are included in the binary releases. Source code for the benchmarks is also provided here on Github to provide additional examples on how to use nvCOMP. For further information on how to execute the benchmarks, please view Benchmarks Page

CPU compression examples

We provide some examples of how you might use CPU compression and GPU decompression or vice versa for LZ4 GDeflate and Deflate. These require some external dependencies, namely:

  • zlib for the GDeflate and Deflate CPU compression/decompression example (zlib1g-dev on debian based systems)
  • LZ4 for the LZ4 CPU compression example (liblz4-dev on debian based systems)
  • libdeflate for the Deflate CPU compression/decompression example (libdeflate-dev on debian based systems)

The CPU example executables are:

gdeflate_cpu_compression {-f <input_file>}
lz4_cpu_compression {-f <input_file>}
lz4_cpu_decompression {-f <input_file>}
deflate_cpu_compression {-a <0 libdeflate, 1 zlib_compress2, 2 zlib_deflate> -f <input_file>}
deflate_cpu_decompression {-a <0 libdeflate, 1 zlib_inflate> -f <input_file>}
gzip_gpu_decompression {-f <input_file>}

Building CPU and GPU Examples, GPU Benchmarks provided on Github

To build only the examples, you'll need cmake >= 3.18 and an nvcomp artifact. Then, you can follow the following steps from the top-level of your clone of nvCOMP from Github

cmake .. -DCMAKE_PREFIX_PATH=<path_to_nvcomp_install>

The path_to_nvcomp_install is the directory where you extracted the nvcomp artifact.

To compile the benchmarks too, you can add -DBUILD_BENCHMARKS=1, but note this is only provided for an additional example of building against the artifacts. The benchmarks are already provided in the artifact bin/ folder.

Logging

To enable logging, set the NVCOMP_LOG_LEVEL environment variable to an integer:

  • 0 for no logging
  • 1 for only error messages
  • 2 for error and warning messages
  • 3 for errors, warnings, and information logged for every low-level interface API call
  • 4 or 5 for debug information, not yet supported

By default, log messages will be written to a file named nvcomp_yyyy-mm-dd_hh-mm.log, with the date and time filled in. If the NVCOMP_LOG_FILE environment variable is set to a valid file path, messages will be logged to that file. Specifying stdout or stderr as the file will log to the console via the appropriate pipe, with color.

More Repositories

1

nvidia-docker

Build and run Docker containers leveraging NVIDIA GPUs
16,896
star
2

open-gpu-kernel-modules

NVIDIA Linux open GPU kernel module source
C
13,784
star
3

DeepLearningExamples

State-of-the-Art Deep Learning scripts organized by models - easy to train and deploy with reproducible accuracy and performance on enterprise-grade infrastructure.
Jupyter Notebook
12,579
star
4

FastPhotoStyle

Style transfer, deep learning, feature transform
Python
11,020
star
5

NeMo

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
Python
10,077
star
6

TensorRT

NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.
C++
9,059
star
7

vid2vid

Pytorch implementation of our method for high-resolution (e.g. 2048x1024) photorealistic video-to-video translation.
Python
8,482
star
8

Megatron-LM

Ongoing research training transformer models at scale
Python
8,169
star
9

apex

A PyTorch Extension: Tools for easy mixed precision and distributed training in Pytorch
Python
7,915
star
10

pix2pixHD

Synthesizing and manipulating 2048x1024 images with conditional GANs
Python
6,488
star
11

TensorRT-LLM

TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM also contains components to create Python and C++ runtimes that execute those TensorRT engines.
C++
6,429
star
12

FasterTransformer

Transformer related optimization, including BERT, GPT
C++
5,313
star
13

cuda-samples

Samples for CUDA Developers which demonstrates features in CUDA Toolkit
C
5,203
star
14

thrust

[ARCHIVED] The C++ parallel algorithms library. See https://github.com/NVIDIA/cccl
C++
4,845
star
15

DALI

A GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.
C++
4,839
star
16

tacotron2

Tacotron 2 - PyTorch implementation with faster-than-realtime inference
Jupyter Notebook
4,562
star
17

cutlass

CUDA Templates for Linear Algebra Subroutines
C++
4,278
star
18

DIGITS

Deep Learning GPU Training System
HTML
4,105
star
19

NeMo-Guardrails

NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.
Python
3,309
star
20

flownet2-pytorch

Pytorch implementation of FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks
Python
2,938
star
21

nccl

Optimized primitives for collective multi-GPU communication
C++
2,786
star
22

libcudacxx

[ARCHIVED] The C++ Standard Library for your entire system. See https://github.com/NVIDIA/cccl
C++
2,286
star
23

k8s-device-plugin

NVIDIA device plugin for Kubernetes
Go
2,269
star
24

waveglow

A Flow-based Generative Network for Speech Synthesis
Python
2,133
star
25

trt-llm-rag-windows

A developer reference project for creating Retrieval Augmented Generation (RAG) chatbots on Windows using TensorRT-LLM
Python
2,011
star
26

MinkowskiEngine

Minkowski Engine is an auto-diff neural network library for high-dimensional sparse tensors
Python
2,007
star
27

semantic-segmentation

Nvidia Semantic Segmentation monorepo
Python
1,746
star
28

DeepRecommender

Deep learning for recommender systems
Python
1,662
star
29

Stable-Diffusion-WebUI-TensorRT

TensorRT Extension for Stable Diffusion Web UI
Python
1,660
star
30

cub

[ARCHIVED] Cooperative primitives for CUDA C++. See https://github.com/NVIDIA/cccl
Cuda
1,645
star
31

warp

A Python framework for high performance GPU simulation and graphics
Python
1,573
star
32

OpenSeq2Seq

Toolkit for efficient experimentation with Speech Recognition, Text2Speech and NLP
Python
1,511
star
33

GenerativeAIExamples

Generative AI reference workflows optimized for accelerated infrastructure and microservice architecture.
Python
1,450
star
34

TransformerEngine

A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.
Python
1,400
star
35

VideoProcessingFramework

Set of Python bindings to C++ libraries which provides full HW acceleration for video decoding, encoding and GPU-accelerated color space and pixel format conversions
C++
1,253
star
36

nvidia-container-toolkit

Build and run containers leveraging NVIDIA GPUs
Go
1,239
star
37

trt-samples-for-hackathon-cn

Simple samples for TensorRT programming
Python
1,211
star
38

Q2RTX

NVIDIA’s implementation of RTX ray-tracing in Quake II
C
1,201
star
39

open-gpu-doc

Documentation of NVIDIA chip/hardware interfaces
C
1,193
star
40

stdexec

`std::execution`, the proposed C++ framework for asynchronous and parallel programming.
C++
1,182
star
41

deepops

Tools for building GPU clusters
Shell
1,165
star
42

partialconv

A New Padding Scheme: Partial Convolution based Padding
Python
1,145
star
43

CUDALibrarySamples

CUDA Library Samples
Cuda
1,122
star
44

gpu-operator

NVIDIA GPU Operator creates/configures/manages GPUs atop Kubernetes
Go
1,117
star
45

MatX

An efficient C++17 GPU numerical computing library with Python-like syntax
C++
1,104
star
46

aistore

AIStore: scalable storage for AI applications
Go
1,074
star
47

sentiment-discovery

Unsupervised Language Modeling at scale for robust sentiment classification
Python
1,055
star
48

nvidia-container-runtime

NVIDIA container runtime
Makefile
1,035
star
49

gpu-monitoring-tools

Tools for monitoring NVIDIA GPUs on Linux
C
974
star
50

retinanet-examples

Fast and accurate object detection with end-to-end GPU optimization
Python
876
star
51

flowtron

Flowtron is an auto-regressive flow-based generative network for text to speech synthesis with control over speech variation and style transfer
Jupyter Notebook
867
star
52

mellotron

Mellotron: a multispeaker voice synthesis model based on Tacotron 2 GST that can make a voice emote and sing without emotive or singing training data
Jupyter Notebook
842
star
53

jetson-gpio

A Python library that enables the use of Jetson's GPIOs
Python
834
star
54

gdrcopy

A fast GPU memory copy library based on NVIDIA GPUDirect RDMA technology
C++
766
star
55

nv-wavenet

Reference implementation of real-time autoregressive wavenet inference
Cuda
728
star
56

tensorflow

An Open Source Machine Learning Framework for Everyone
C++
719
star
57

spark-rapids

Spark RAPIDS plugin - accelerate Apache Spark with GPUs
Scala
717
star
58

cuda-python

CUDA Python Low-level Bindings
Python
695
star
59

libnvidia-container

NVIDIA container runtime library
C
679
star
60

cccl

CUDA C++ Core Libraries
C++
676
star
61

MAXINE-AR-SDK

NVIDIA AR SDK - API headers and sample applications
C
671
star
62

nvvl

A library that uses hardware acceleration to load sequences of video frames to facilitate machine learning training
C++
665
star
63

nccl-tests

NCCL Tests
Cuda
648
star
64

gvdb-voxels

Sparse volume compute and rendering on NVIDIA GPUs
C
643
star
65

modulus

Open-source deep-learning framework for building, training, and fine-tuning deep learning models using state-of-the-art Physics-ML methods
Python
636
star
66

BigVGAN

Official PyTorch implementation of BigVGAN (ICLR 2023)
Python
633
star
67

runx

Deep Learning Experiment Management
Python
630
star
68

DLSS

NVIDIA DLSS is a new and improved deep learning neural network that boosts frame rates and generates beautiful, sharp images for your games
C
588
star
69

dcgm-exporter

NVIDIA GPU metrics exporter for Prometheus leveraging DCGM
Go
551
star
70

Dataset_Synthesizer

NVIDIA Deep learning Dataset Synthesizer (NDDS)
C++
530
star
71

NVFlare

NVIDIA Federated Learning Application Runtime Environment
Python
528
star
72

jitify

A single-header C++ library for simplifying the use of CUDA Runtime Compilation (NVRTC).
C++
495
star
73

libglvnd

The GL Vendor-Neutral Dispatch library
C
462
star
74

enroot

A simple yet powerful tool to turn traditional container/OS images into unprivileged sandboxes.
Shell
459
star
75

multi-gpu-programming-models

Examples demonstrating available options to program multiple GPUs in a single node or a cluster
Cuda
438
star
76

MDL-SDK

NVIDIA Material Definition Language SDK
C++
438
star
77

PyProf

A GPU performance profiling tool for PyTorch models
Python
437
star
78

AMGX

Distributed multigrid linear solver library on GPU
Cuda
434
star
79

gpu-rest-engine

A REST API for Caffe using Docker and Go
C++
421
star
80

nvbench

CUDA Kernel Benchmarking Library
Cuda
413
star
81

framework-reproducibility

Providing reproducibility in deep learning frameworks
Python
412
star
82

cuCollections

C++
410
star
83

hpc-container-maker

HPC Container Maker
Python
404
star
84

NeMo-Framework-Launcher

NeMo Megatron launcher and tools
Python
391
star
85

NvPipe

NVIDIA-accelerated zero latency video compression library for interactive remoting applications
Cuda
384
star
86

cuda-quantum

C++ and Python support for the CUDA Quantum programming model for heterogeneous quantum-classical workflows
C++
363
star
87

data-science-stack

NVIDIA Data Science stack tools
Shell
317
star
88

cuQuantum

Home for cuQuantum Python & NVIDIA cuQuantum SDK C++ samples
Jupyter Notebook
305
star
89

ai-assisted-annotation-client

Client side integration example source code and libraries for AI-Assisted Annotation SDK
C++
302
star
90

video-sdk-samples

Samples demonstrating how to use various APIs of NVIDIA Video Codec SDK
C++
301
star
91

nvidia-settings

NVIDIA driver control panel
C
284
star
92

DCGM

NVIDIA Data Center GPU Manager (DCGM) is a project for gathering telemetry and measuring the health of NVIDIA GPUs
C++
282
star
93

cnmem

A simple memory manager for CUDA designed to help Deep Learning frameworks manage memory
C++
280
star
94

radtts

Provides training, inference and voice conversion recipes for RADTTS and RADTTS++: Flow-based TTS models with Robust Alignment Learning, Diverse Synthesis, and Generative Modeling and Fine-Grained Control over of Low Dimensional (F0 and Energy) Speech Attributes.
Roff
269
star
95

fsi-samples

A collection of open-source GPU accelerated Python tools and examples for quantitative analyst tasks and leverages RAPIDS AI project, Numba, cuDF, and Dask.
Jupyter Notebook
265
star
96

tensorrt-laboratory

Explore the Capabilities of the TensorRT Platform
C++
259
star
97

CleanUNet

Official PyTorch Implementation of CleanUNet (ICASSP 2022)
Python
258
star
98

gpu-feature-discovery

GPU plugin to the node feature discovery for Kubernetes
Go
255
star
99

torch-harmonics

Differentiable spherical harmonic transforms and spherical convolutions in PyTorch
Jupyter Notebook
246
star
100

egl-wayland

The EGLStream-based Wayland external platform
C
243
star