Completed
Assumptions for Probabilistic Analysis
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Numerical Stability of Algorithms at Extreme Scale and Low Precisions
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 TOP500: June 2022 Frontier at Oak Ridge AMD EPYC 64C 2GHz, AMD Radeon Instinct GPU. 8,730,112 cores
- 3 Growth of Problem Size in TOP500
- 4 Today's Floating-Point Arithmetics
- 5 Backward Error Analysis for LU Factorizatia
- 6 Low Precision in Deep Learning
- 7 The (Partial) Explanation
- 8 Blocked Inner Products: 2 Pieces
- 9 Block Summation
- 10 FABsum Error Bound
- 11 Random Uniform (0, 1), b = 128, fp32
- 12 Extended Precision Registers
- 13 Mixed Precision Block FMA
- 14 Block FMA Hardware
- 15 Error Analysis of Block FMAS Blanchard, H, Lopez, Mary, & Pranesh (2020). Analysis of algs for matrix mult c - AB based on block
- 16 NVIDIA V100
- 17 Probabilistic Error Analysis Rounding error bounds above are worst-case.
- 18 Statistical Effects
- 19 Standard Tool for Rounding Error Analysis Theorem If | sufori - 1 n and nu 1 then
- 20 Assumptions for Probabilistic Analysis
- 21 Probabilistic Analysis Theorem (Connolly, H & Mary, 2021)
- 22 Linear Systems
- 23 Real-Life Matrices
- 24 Probabilistic QR Error Bound Theorem (Connolly & H, 2022) Under Model Mand a technical assumption, for the
- 25 Stochastic Rounding Forsythe (1950). .... Croci et al. (2022).
- 26 Stagnation Harmonic sum 1/k in fp16.
- 27 Random Data
- 28 Putting It All Together
- 29 Conclusions
