diff --git a/reports/benchmarks/benchmarks.tex b/reports/benchmarks/benchmarks.tex
index 2e96b6b..f874743 100644
--- a/reports/benchmarks/benchmarks.tex
+++ b/reports/benchmarks/benchmarks.tex
@@ -1,137 +1,138 @@
 \documentclass[11pt,a4paper]{article}
 
 \usepackage[colorlinks,urlcolor=blue,linkcolor=magenta,citecolor=red,linktocpage=true]{hyperref}
 \usepackage[english]{babel}
 \usepackage{a4wide}
 \usepackage{booktabs}
 \usepackage{minted}
 \usepackage{siunitx}
 
 \title{Google Summer of Code 2019}
 \author{Thibault Allançon}
 \date{August 23, 2019}
 
 \begin{document}
 
 \maketitle
 
-Benchmarks results on the compressed graph, using the API endpoints.
+Benchmarks results on the compressed graph, using the API endpoints. These were
+done on the Azure VM (16 vCPUs, 224 GiB of RAM).
 
 \section{Environment}
 
 For each benchmarks, 10000 nodes were randomly selected and used for each
 use-case as a starting point. All the results are then aggregated.
 
 See
 \href{https://docs.softwareheritage.org/devel/swh-graph/use-cases.html}{use-cases}
 for more details on each use-case.
 
 \section{Results}
 
 All the timings are in $\mu$s and normalized with respect to the number of edges
 accessed during the traversal. The goal is to evaluate the cost of accessing an
 edge on average for each use-case.
 
 \subsection{Browsing use-cases}
 
 \begin{center}
     \begin{tabular}{@{} l *3r @{}}
         \toprule
         \multicolumn{1}{c}{} &
             \textbf{ls} &
             \textbf{ls -R} &
             \textbf{git log} \\
         \midrule
         min value & 0.05 & 0.08 & 0.1 \\
         max value & 1791 & 100  & 58 \\
         average & 3.0 & 3.8 & 1.5 \\
         median & 1.6 & 1.7 & 3.2 \\
         standard deviation & 20 & 5.0 & 3.4 \\
         \bottomrule
     \end{tabular}
 \end{center}
 
 \vspace{0.5cm}
 
 \begin{center}
     \begin{tabular}{@{} l *3r @{}}
         \toprule
         \multicolumn{1}{c}{} &
             \textbf{ls} &
             \textbf{ls -R} &
             \textbf{git log} \\
         \midrule
         average edge accessed & 30 & 2566 & 21081 \\
         \bottomrule
     \end{tabular}
 \end{center}
 
 \newpage
 
 \subsection{Vault use-cases}
 
 \begin{center}
     \begin{tabular}{@{} l r @{}}
         \toprule
         \multicolumn{1}{c}{} &
             \textbf{git bundle} \\
         \midrule
         min value & 0.03 \\
         max value & 483 \\
         average & 2.9 \\
         median & 0.9 \\
         standard deviation & 0.2 \\
         \bottomrule
     \end{tabular}
 \end{center}
 
 \vspace{0.5cm}
 
 \begin{center}
     \begin{tabular}{@{} l r @{}}
         \toprule
         \multicolumn{1}{c}{} &
             \textbf{git bundle} \\
         \midrule
         average edge accessed & 456418 \\
         \bottomrule
     \end{tabular}
 \end{center}
 
 \subsection{Provenance use-cases}
 
 \begin{center}
     \begin{tabular}{@{} l *4r @{}}
         \toprule
         \multicolumn{1}{c}{} &
             \textbf{commit pr.} &
             \textbf{complete commit pr.} &
             \textbf{origin pr.} &
             \textbf{complete origin pr.} \\
         \midrule
         min value & 0 & 0.03 & 0 & 0.02 \\
         max value & 2600 & 136 & 571 & 57 \\
         average & 3.5 & 1.7 & 1.4 & 1.4 \\
         median & 2.4 & 1.7 & 1.1 & 1.4 \\
         standard deviation & 3.0 & 1.7 & 6.2 & 1.2 \\
         \bottomrule
     \end{tabular}
 \end{center}
 
 \vspace{0.5cm}
 
 \begin{center}
     \begin{tabular}{@{} l *4r @{}}
         \toprule
         \multicolumn{1}{c}{} &
             \textbf{commit pr.} &
             \textbf{complete commit pr.} &
             \textbf{origin pr.} &
             \textbf{complete origin pr.} \\
         \midrule
         average edge accessed & 38 & 12 & 2799 & 12 \\
         \bottomrule
     \end{tabular}
 \end{center}
 
 \end{document}