Unit 2.1.2 Outline Week 2
ΒΆ-
2.1 Opening Remarks
2.1.1 Launch
2.1.2 Outline Week 2
2.1.3 What you will learn
-
2.2 Blocked Matrix-Matrix Multiplication
2.2.1 Basic idea
2.2.2 Haven't we seen this before?
-
2.3 Blocking for Registers
2.3.1 A simple model of memory and registers
2.3.2 Simple blocking for registers
2.3.3 Streaming \(A_{i,p} \) and \(B_{p,j} \)
2.3.4 Combining loops
2.3.5 Alternative view
-
2.4 Optimizing the Micro-kernel
2.4.1 Vector registers and instructions
2.4.2 Implementing the micro-kernel with vector instructions
2.4.3 Details
2.4.4 More options
2.4.5 Optimally amortizing data movement
-
2.5 Enrichments
2.5.1 Lower bound on data movement
-
2.6 Wrap Up
2.6.1 Additional exercises
2.6.2 Summary