<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE erlref SYSTEM "erlref.dtd">
<erlref>
<header>
<copyright>
<year>2018</year><year>2018</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
</legalnotice>
<title>persistent_term</title>
<prepared></prepared>
<docno></docno>
<date></date>
<rev></rev>
<file>persistent_term.xml</file>
</header>
<module>persistent_term</module>
<modulesummary>Persistent terms.</modulesummary>
<description>
<p>This module is similar to <seealso
marker="stdlib:ets"><c>ets</c></seealso> in that it provides a
storage for Erlang terms that can be accessed in constant time,
but with the difference that <c>persistent_term</c> has been
highly optimized for reading terms at the expense of writing and
updating terms. When a persistent term is updated or deleted, a
global garbage collection pass is run to scan all processes for
the deleted term, and to copy it into each process that still uses
it. Therefore, <c>persistent_term</c> is suitable for storing
Erlang terms that are frequently accessed but never or
infrequently updated.</p>
<warning><p>Persistent terms is an advanced feature and is not a
general replacement for ETS tables. Before using persistent terms,
make sure to fully understand the consequence to system
performance when updating or deleting persistent terms.</p></warning>
<p>Term lookup (using <seealso
marker="#get/1"><c>get/1</c></seealso>), is done in constant time
and without taking any locks, and the term is <strong>not</strong>
copied to the heap (as is the case with terms stored in ETS
tables).</p>
<p>Storing or updating a term (using <seealso
marker="#put/2"><c>put/2</c></seealso>) is proportional to the
number of already created persistent terms because the hash table
holding the keys will be copied. In addition, the term itself will
be copied.</p>
<p>When a (complex) term is deleted (using <seealso
marker="#erase/1"><c>erase/1</c></seealso>) or replaced by another
(using <seealso marker="#put/2"><c>put/2</c></seealso>), a global
garbage collection is initiated. It works like this:</p>
<list>
<item><p>All processes in the system will be scheduled to run a
scan of their heaps for the term that has been deleted. While
such scan is relatively light-weight, if there are many
processes, the system can become less responsive until all
process have scanned theirs heaps.</p></item>
<item><p>If the deleted term (or any part of it) is still used
by a process, that process will do a major (fullsweep) garbage
collection and copy the term into the process. However, at most
two processes at a time will be scheduled to do that kind of
garbage collection.</p></item>
</list>
<p>Deletion of atoms and other terms that fit in one machine word
is specially optimized to avoid doing a global GC. It is still not
recommended to update persistent terms with such values too
frequently because the hash table holding the keys is copied every
time a persistent term is updated.</p>
<p>Some examples are suitable uses for persistent terms are:</p>
<list>
<item><p>Storing of configuration data that must be easily
accessible by all processes.</p></item>
<item><p>Storing of references for NIF resources.</p></item>
<item><p>Storing of references for efficient counters.</p></item>
<item><p>Storing an atom to indicate a logging level or whether debugging
is turned on.</p></item>
</list>
</description>
<section>
<title>Storing Huge Persistent Terms</title>
<p>The current implementation of persistent terms uses the literal
<seealso marker="erts_alloc">allocator</seealso> also used for
literals (constant terms) in BEAM code. By default, 1 GB of
virtual address space is reserved for literals in BEAM code and
persistent terms. The amount of virtual address space reserved for
literals can be changed by using the <seealso
marker="erts_alloc#MIscs"><c>+MIscs option</c></seealso> when
starting the emulator.</p>
<p>Here is an example how the reserved virtual address space for literals
can be raised to 2 GB (2048 MB):</p>
<pre>
erl +MIscs 2048</pre>
</section>
<section>
<title>Warning For Many Persistent Terms</title>
<p>The runtime system will send a warning report to the
error logger if more than 20000 persistent terms have been
created. It will look like this:</p>
<pre>
More than 20000 persistent terms have been created.
It is recommended to avoid creating an excessive number of
persistent terms, as creation and deletion of persistent terms
will be slower as the number of persistent terms increases.</pre>
</section>
<section>
<title>Best Practices for Using Persistent Terms</title>
<p>It is recommended to use keys like <c>?MODULE</c> or
<c>{?MODULE,SubKey}</c> to avoid name collisions.</p>
<p>Prefer creating a few large persistent terms to creating many
small persistent terms. The execution time for storing a
persistent term is proportional to the number of already existing
terms.</p>
<p>Updating a persistent term with the same value as it already
has is specially optimized to do nothing quickly; thus, there is
no need compare the old and new values and avoid calling
<seealso marker="#put/2"><c>put/2</c></seealso> if the values
are equal.</p>
<p>When atoms or other terms that fit in one machine word are
deleted, no global GC is needed. Therefore, persistent terms that
have atoms as their values can be updated more frequently, but
note that updating such persistent terms is still much more
expensive than reading them.</p>
<p>Updating or deleting a persistent term will trigger a global GC
if the term does not fit in one machine word. Processes will be
scheduled as usual, but all processes will be made runnable at
once, which will make the system less responsive until all process
have run and scanned their heaps for the deleted terms. One way to
minimize the effects on responsiveness could be to minimize the
number of processes on the node before updating or deleting a
persistent term. It would also be wise to avoid updating terms
when the system is at peak load.</p>
<p>Avoid storing a retrieved persistent term in a process if that
persistent term could be deleted or updated in the future. If a
process holds a reference to a persistent term when the term is
deleted, the process will be garbage collected and the term copied
to process.</p>
<p>Avoid updating or deleting more than one persistent term at a
time. Each deleted term will trigger its own global GC. That
means that deleting N terms will make the system less responsive N
times longer than deleting a single persistent term. Therefore,
terms that are to be updated at the same time should be collected
into a larger term, for example, a map or a tuple.</p>
</section>
<section>
<title>Example</title>
<p>The following example shows how lock contention for ETS tables
can be minimized by having one ETS table for each scheduler. The
table identifiers for the ETS tables are stored as a single
persistent term:</p>
<pre>
%% There is one ETS table for each scheduler.
Sid = erlang:system_info(scheduler_id),
Tid = element(Sid, persistent_term:get(?MODULE)),
ets:update_counter(Tid, Key, 1).</pre>
</section>
<datatypes>
<datatype>
<name name="key"/>
<desc>
<p>Any Erlang term.</p>
</desc>
</datatype>
<datatype>
<name name="value"/>
<desc>
<p>Any Erlang term.</p>
</desc>
</datatype>
</datatypes>
<funcs>
<func>
<name name="erase" arity="1"/>
<fsummary>Erase the name for a persistent term.</fsummary>
<desc>
<p>Erase the name for the persistent term with key
<c><anno>Key</anno></c>. The return value will be <c>true</c>
if there was a persistent term with the key
<c><anno>Key</anno></c>, and <c>false</c> if there was no
persistent term associated with the key.</p>
<p>If there existed a previous persistent term associated with
key <c><anno>Key</anno></c>, a global GC has been initiated
when <c>erase/1</c> returns. See <seealso
marker="#description">Description</seealso>.</p>
</desc>
</func>
<func>
<name name="get" arity="0"/>
<fsummary>Get all persistent terms.</fsummary>
<desc>
<p>Retrieve the keys and values for all persistent terms.
The keys will be copied to the heap for the process calling
<c>get/0</c>, but the values will not.</p>
</desc>
</func>
<func>
<name name="get" arity="1"/>
<fsummary>Get the value for a persistent term.</fsummary>
<desc>
<p>Retrieve the value for the persistent term associated with
the key <c><anno>Key</anno></c>. The lookup will be made in
constant time and the value will not be copied to the heap
of the calling process.</p>
<p>This function fails with a <c>badarg</c> exception if no
term has been stored with the key
<c><anno>Key</anno></c>.</p>
<p>If the calling process holds on to the value of the
persistent term and the persistent term is deleted in the future,
the term will be copied to the process.</p>
</desc>
</func>
<func>
<name name="info" arity="0"/>
<fsummary>Get information about persistent terms.</fsummary>
<desc>
<p>Return information about persistent terms in a map. The map
has the following keys:</p>
<taglist>
<tag><c>count</c></tag>
<item><p>The number of persistent terms.</p></item>
<tag><c>memory</c></tag>
<item><p>The total amount of memory (measured in bytes)
used by all persistent terms.</p></item>
</taglist>
</desc>
</func>
<func>
<name name="put" arity="2"/>
<fsummary>Store a term.</fsummary>
<desc>
<p>Store the value <c><anno>Value</anno></c> as a persistent term and
associate it with the key <c><anno>Key</anno></c>.</p>
<p>If the value <c><anno>Value</anno></c> is equal to the value
previously stored for the key, <c>put/2</c> will do nothing and return
quickly.</p>
<p>If there existed a previous persistent term associated with
key <c><anno>Key</anno></c>, a global GC has been initiated
when <c>put/2</c> returns. See <seealso
marker="#description">Description</seealso>.</p>
</desc>
</func>
</funcs>
</erlref>
