← CUDA, Triton & Kernel Engineering / 34

hardNew

A GEMM launches 96 tiles into 80 execution slots. Should you split K?

Derive the ideal gain from filling a partial wave, then account for partial outputs, reduction order, workspace and the epilogue before selecting split-K.

Updated Oct 2026 · Learn how AI infrastructure works through explanations, worked examples and diagrams, then practise applying it to interview questions.

Derive the ideal gain from filling a partial wave, then account for partial outputs, reduction order, workspace and the epilogue before selecting split-K.

The full answer is part of Premium. A free account includes more answers in each topic, but does not unlock this one.

20 answers per topic instead of 10, plus saved progress and bookmarks · no cardor unlock all 316 remaining answers · ₹2,000 / $25

Work through the numbers

Change the assumptions in a free calculator, then explain what the result establishes and what still needs measurement.

Roofline calculator

The concepts behind this question

Use these concept explanations to review the ideas behind this question.

⚡ Kernels & Compilers🔒 Premium

GEMM Work Partitioning and Wave QuantizationOutput tiles determine how much independent matrix work is available. Split-K and work-centric schedules can reduce an underfilled final wave, while adding partial-result reduction and workspace costs.

⚡ Kernels & Compilers🔒 Premium

Tiled Matrix MultiplicationTiled matrix multiplication reuses input values across many output calculations. Shared-memory and register tiles reduce repeated data movement while consuming limited on-chip resources.

⚡ Kernels & Compilers🔒 Premium

CUTLASS and Tensor Core KernelsCUTLASS provides components for building GPU matrix kernels. It is useful when the required precision, batching or fused operations need more control than a library call offers.

⚡ Kernels & Compilers🔒 Premium

CUDA Compilation: PTX, Cubins and SASSCUDA source, PTX and native GPU instructions describe different stages of execution. Inspect the selected binary and its resource use before explaining a performance change.

Explore the full curriculum →

UP NEXT ON YOUR JOURNEY

Next in this trackEach GPU block computes correct local offsets, but output ranges overlap. What is missing?Next in this trackAn AI-generated kernel passes random tests and claims a 3x speedup. What would you verify?Closest to this topicA CUDA upgrade improves warm runs but slows the first request. Would you ship it?

DISCUSSION · 0

No comments yet — be the first to share your approach.