liveBook · Manning

Kernel launch

A kernel launch is an operation that submits a GPU kernel for execution by GPU threads under a specified execution configuration.

How 2 Manning books explain it

CUDA for LLMs

A kernel launch is the operation that starts execution of a CUDA kernel on GPU threads.

How it is used
Orchestrates GPU computation after host-to-device input copies and before results return to the host.
How deep it goes
Defines launch syntax, thread counts, and thread indexing, while deferring fuller block details.
Examples
Demonstrates vector addition and neural-network forward and backward passes.
What you can do after
Apply CUDA launch syntax to invoke a device function with configured blocks and threads.

Read how CUDA for LLMs explains kernel launch →

Grokking GPU Programming

A kernel launch starts execution of a GPU kernel across a specified number of threads.

How it is used
Starts GPU work by running a kernel across threads that process data in parallel.
How deep it goes
Explains launch dimensions, asynchronous completion, bounds checks, and the failure caused by exceeding a block’s thread limit.
Examples
Uses <<<1, n>>> to launch n threads in one block and illustrates indexing and bounds checks.
What you can do after
Configure it by choosing a thread count, checking bounds, and synchronizing before relying on results.

Read how Grokking GPU Programming explains kernel launch →

What to know first

Related concepts

Questions about kernel launch

What is kernel launch?

A kernel launch is an operation that submits a GPU kernel for execution by GPU threads under a specified execution configuration.

Which Manning books explain kernel launch?

CUDA for LLMs and Grokking GPU Programming.

What should I know before learning kernel launch?

Manning books that teach kernel launch usually introduce CUDA first.

What is kernel launch part of?

Kernel launch is part of CUDA data lifecycle. The CUDA data lifecycle is the five-stage GPU workflow of allocating device memory, copying data in, computing, copying results out, and freeing device memory.

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.