Jared Hoberock

stanford-cs193g-sp2010. This is an archive of materials produced for an introductory class on CUDA programming at Stanford University in 2010

tuple_utility. The missing C++ tuple functionality

hindley_milner. A C++11 implementation of Hindley-Milner type inference.

thrust-workshop. Introductory Thrust workshop materials

bulk. Launching collective tasks in bulk

future. Implementation of std::experimental::future from the C++ Concurrency TS

is_call_possible. C++03 functionality for checking for the existence of a member function with a given name and signature.

managed_allocator. A C++ allocator based on cudaMallocManaged

cuda_launch_config. Utilities for automatically selecting a CUDA kernel launch configuration

cudex. CUDA executors

variant. Standalone C++11 implementation of variant

cuda_graphs_executor_prototypes. Executor prototypes for interacting with traditional CUDA kernel launch and CUDA Graphs.

tuple. Freestanding std::tuple implementation

gotham. Photorealistic Renderer based on Unbiased Rendering Algorithms

newton. Thrust-accelerated numerics

shmalloc. Dynamic __shared__ memory allocation for CUDA

thrust. Development fork of Thrust

then. Experimental implementation of future::then()

compiler. toy compiler

cumem. CUDA C++ memory utilities

process. A std::thread-alike for processes

personal. odds & ends that don't belong anywhere else

thread_pool. A simple thread pool implementation

static_algorithm. A Statically-Unrolled C++ Algorithms Library

interprocess_future. A std::-like future & promise pair for inter-process communication

igloo. Physically-based Renderer Igloo

strange. Ranges for Thrust.

static_process_pool. A std::static_thread_pool-alike for processes.

ubu. C++

moderngpu. Design patterns for GPU computing

overload. Standalone CUDA-compatible C++11 implementation of overload

agency. Execution primitives for C++

coord. Utilities for navigating multi-dimensional iteration spaces

cuda_cpp_template. Jared's repository template

coordinate. A mathematical vector in an N-dimentional space

nvcc-scons. A SCons build tool for the NVIDIA compiler

gpu_algorithms. Implementing GPU algorithms various ways

simple_cuda_executor. A demonstration of the implementation of a very simple GPU executor in CUDA

variable. C++20 expression template for a variable whose value is unknown

thrust-simple-benchmarks. Simple Thrust benchmarks

lazy_cuda_executor. Toy implementation of a lazy CUDA executor

parallelism-ts. Technical Specification: Parallelism

bounding_box_hierarchy. A generic data structure for finding intersections between rays and geometric objects

omp_parallel_invoke. Porting parallel_invoke to OpenMP

migrate_hg_to_git. Migrates an existing Mercurial repository to Git

optional. Standalone C++11 implementation of optional

Alenka. GPU database engine

thrust-benchmarks. C

any. Generic container for objects of a discriminated type

recursive_variant. A std::variant-alike permitting alternatives that are incomplete types

mlir-option-dialect. An MLIR dialect for working with monadic option types

bounded. C++20 class template for a value bounded by a compile-time constant

nes_emulator. This is an emulator for the Nintendo Entertainment System (NES) written in modern C++.

constant. C++20 class template for a compile-time constant value

integer_sequence. Standalone C++11 implementation of integer_sequence and friends

always_ready_future. An as-lightweight-as-possible future-like type holding a value that is always ready

ndarray. A multidimensional array container.

active_message. Demo of active messages using OpenSHMEM