From 9dd44323421814562127d6ac372279116ffac982 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9ophile=20Bastian?= <contact@tobast.fr>
Date: Mon, 9 Oct 2023 18:13:30 +0200
Subject: [PATCH] Intro: add ref to https://arxiv.org/abs/2003.00532
---
manuscrit/10_introduction/main.tex | 8 ++++++--
manuscrit/biblio/misc.bib | 10 ++++++++++
2 files changed, 16 insertions(+), 2 deletions(-)
diff --git a/manuscrit/10_introduction/main.tex b/manuscrit/10_introduction/main.tex
index 1a5c205..9c927c5 100644
--- a/manuscrit/10_introduction/main.tex
+++ b/manuscrit/10_introduction/main.tex
@@ -56,8 +56,12 @@ slower than the former~\cite{rowmajor_repo}.
This, however, is still an optimization that holds for the vast majority of
CPUs. In many cases, transformations targeting a specific microarchitecture can
-be very beneficial. \qtodo{Insert number/ref \wrt{} matmult or some kernel of
-the like.} This kind of optimizations, however, requires manual effort, and a
+be very beneficial.
+For instance, Uday Bondhugula found out that manual tuning, through many
+techniques and tools, of a general matrix multiplication could multiply its
+throughput by roughly 13.5 compared to \texttt{gcc~-O3}, or even 130 times
+faster than \texttt{clang -O3}~\cite{dgemm_finetune}.
+This kind of optimizations, however, requires manual effort, and a
deep expert knowledge both in optimization techniques and on the specific
architecture targeted.
These techniques are only worth applying on the parts of a program that are
diff --git a/manuscrit/biblio/misc.bib b/manuscrit/biblio/misc.bib
index 35ccd2e..1139896 100644
--- a/manuscrit/biblio/misc.bib
+++ b/manuscrit/biblio/misc.bib
@@ -148,3 +148,13 @@
month=10,
howpublished={\url{https://gitlab.inria.fr/tbastian/rowmajor-measure}},
}
+
+@misc{dgemm_finetune,
+ title={High Performance Code Generation in MLIR: An Early Case Study
+ with GEMM},
+ author={Uday Bondhugula},
+ year={2020},
+ eprint={2003.00532},
+ archivePrefix={arXiv},
+ primaryClass={cs.PF}
+}