diff options
author | Paul E. McKenney <paulmck@kernel.org> | 2023-06-25 20:48:09 -0700 |
---|---|---|
committer | Paul E. McKenney <paulmck@kernel.org> | 2023-06-25 20:48:09 -0700 |
commit | bef8d41efe53d227a5b0f886490a161608ae4ea5 (patch) | |
tree | 8275857db62249a8df6de57033311c5fe6f6d111 | |
parent | 636a610ab243d77b5b2446be6d3ed58b4f8de665 (diff) | |
download | perfbook-bef8d41efe53d227a5b0f886490a161608ae4ea5.tar.gz |
cpu/hwfreelunch: Add vacuum-gap transistors and wordsmith
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
-rw-r--r-- | bib/hw.bib | 9 | ||||
-rw-r--r-- | cpu/hwfreelunch.tex | 33 |
2 files changed, 29 insertions, 13 deletions
@@ -883,6 +883,15 @@ on Computer Architecture", lastchecked="November 13, 2016", } +@article{JinWooHan2014VacuumGapTransistor, + author="Jin-Woo Han and Meyya Mayyappan", + Title="Introducing the Vacuum Transistor: A Device Made of Nothing", + Year="2014", + Month="June", + pages="??--??", + Journal="IEEE Spectrum", +} + @unpublished{MichaelJKelly2017DeviceLevel, Author="Michael J. Kelly", Title="How Might the Manufacturability of the Hardware at Device Level Impact on Exascale Computing?", diff --git a/cpu/hwfreelunch.tex b/cpu/hwfreelunch.tex index db8795f4..76f048f1 100644 --- a/cpu/hwfreelunch.tex +++ b/cpu/hwfreelunch.tex @@ -110,6 +110,14 @@ to communicate between a CPU and main memory. up on electrodynamics. }\QuickQuizEnd +In fact, Stephen Hawking is said to have claimed that semiconductor +manufacturers have but two fundamental problems: +\begin{enumerate*}[(1)] +\item The finite speed of light and +\item The atomic nature of matter~\cite{BryanGardiner2007}. +\end{enumerate*} +That is right, light is too slow and atoms are too big!!! + There are nevertheless some technologies (both hardware and software) that might help improve matters: @@ -126,18 +134,11 @@ Each of these is described in one of the following sections. \subsection{Novel Materials and Processes} \label{sec:cpu:Novel Materials and Processes} -Stephen Hawking is said to have claimed that semiconductor manufacturers -have but two fundamental problems: -\begin{enumerate*}[(1)] -\item The finite speed of light and -\item The atomic nature of matter~\cite{BryanGardiner2007}. -\end{enumerate*} -That is right, light is too slow and atoms are too big!!! - -It is possible that semiconductor manufacturers are approaching the limits -implied by this pair of laws of physics, but there are nevertheless a -few avenues of research and development focused on working around these -fundamental limits. +It is possible that Stephen Hawking is right, and that semiconductor +manufacturers are approaching the limits implied by the finite speed of +light and the non-zero sizes of atoms. +However, there are a few avenues of research and development focused on +working around these fundamental laws of physics. One workaround for the atomic nature of matter are so-called ``high-K dielectric'' materials, which allow larger devices to mimic the @@ -153,6 +154,9 @@ reliably in production semiconductor devices. Another proposed workaround is the ``quantum dot'' approach that allows much smaller device sizes, but which is still in the research stage. +Yet another proposed workaround is to replace the atoms making up the +base of a traditional transistor with a vacuum, resulting in the +vacuum-gap transistor. One challenge is that many recent hardware-device-level breakthroughs require very tight control of which atoms are placed @@ -209,6 +213,8 @@ long horizontal electrical connections (which are both slow and power hungry) can be replaced by short vertical electrical connections, which are both faster and more power efficient. +If you cannot make light go faster, make your devices smaller! + However, delays due to levels of clocked logic will not be decreased by 3D integration, and significant manufacturing, testing, power-supply, and heat-dissipation problems must be solved for 3D integration to @@ -259,7 +265,8 @@ users to make the specialized hardware affordable. In part due to these sorts of economic considerations, specialized hardware has thus far appeared only for a few application areas, including graphics processing (GPUs), vector processors (MMX, SSE, -and VMX instructions), and, to a lesser extent, encryption. +and VMX instructions), and, to a lesser extent, encryption and +compression. And even in these areas, it is not always easy to realize the expected performance gains, for example, due to thermal throttling~\cite{VladKrasnov2017SIMDfreqscale,DanielLemire2018SIMDfreqscale,TravisDowns2020SIMDfreqscale}. |