Editing arm holdings/microarchitectures/cortex-a510

__NOEDITSECTION____NOTOC__
{| id="mw-blocked-text" style="border:2px solid #000; width:100%; margin:0 auto 6px auto; padding-top:10px;"
|
<div style="width:100%; margin:auto; text-align:center;"><div style="font-size:140%;">'''You are currently unable to edit pages on Wikichip. The reason provided below:'''</div>
'''''You can still read pages''', but you cannot edit or create them.''
</div>
<div style="margin:auto; width:70%;">
<div style="text-align:justify;">
Editing from $7 has been disabled by <tt>$1</tt> for the following reason(s):<br style="clear:both;" />
<div style="margin:1em; padding:5px; border:3px #666 double;">$2</div>
This block has been set to expire: '''$6'''.

Even if blocked, you will ''usually'' still be able to edit your [[Special:MyTalk|user discussion page]] and may also be able to contact other editors and administrators by e-mail.
</div>
</div>
|}

{| style="margin:6px auto; width:100%; background:transparent; border:2px solid #AAA; padding:5px;"
! style="text-align:left;" | <h2 style="border:none;">What does this mean?</h2>
|-
|
<big> What is a block? </big>

A [[Wikibooks:Blocking policy|block]] is a measure used to protect Wikibooks from possible improper use or modification in breach of [[Wikibooks:Policies and guidelines|editorial policies]]. Blocks can apply to a user account, an [[w:IP address|IP address]], or a range of IP addresses that is deemed responsible for or related to problematic activity. You may be an innocent victim of [[w:collateral damage|collateral damage]], whereby a block of some other activity has accidentally caused your account to be inaccessible. The reason for your block may be found above.

<big> Most common causes </big>

* Your [[w:IP address|IP address]] matches that used by another blocked account. Ask for further information and/or request unblocking.
* Your account or a connected IP has been used in a problematic way against Wikibooks' policies, or your username was unsuitable and you need to choose another. The reason should be in the box above.
|}

{| style="margin:6px auto; width:100%; background:transparent; border:2px solid #AAA; padding:5px;"
! style="text-align:left;" | <h2 style="border:none;">What do I do now?</h2>
|-
|
* If instructions are given above in the block message, follow them.
* If you have never edited Wikibooks before and/or do not have an account, [[Special:UserLogin/signup|consider creating one]].
* If you wish to appeal the block, or you believe you have been blocked by mistake, please see the following section.

<big> Appealing </big>

The current block will automatically expire $6. You have many routes to appeal the block or its duration. Firstly, if you are a registered user and have a valid e-mail address confirmed, you can contact $1 '''[[{{ns:special}}:Emailuser/$4|via e-mail]]'''. This feature can be removed if abused.

Alternatively, you may appeal the block by requesting that another [[Wikibooks:Administrators|administrator]] review the block. To do so, add
:<tt>{&#123;unblock|''your reason here''&#125;}</tt>
to the '''bottom''' of [[{{ns:Special}}:Mytalk|your user discussion page]] to request unblocking. You '''must''' state a reason for this, and the block can then be discussed.

If $7 is not blocked, your [[w:IP address|IP address]] ($3) or range may have been blocked. Please check  <span class="plainlinks">[{{fullurl:Special:BlockList|ip=$3}} here]</span>. If this is the case, please copy and paste the following text to the '''bottom''' of [[{{ns:Special}}:Mytalk|your user talk page]].

:<tt>{&#123;unblock-ip|1=$3|2=&lt;nowiki&gt;<nowiki>$2</nowiki>&lt;/nowiki&gt;|3=$1&#125;}</tt>

<span style="color:#CC0000;">'''Please note:'''</span> While this block may be upsetting or unexpected, abuse of appeal processes, repeatedly using the unblock template when denied, [[Wikibooks:No personal attacks|personal attacks]], or [[Wikibooks:Be civil|impolite conduct]] may lead to the removal of your ability to edit your discussion page.
|}

{| style="margin:6px auto; width:100%; background:transparent; border:2px solid #AAA; padding:5px;"
! style="text-align:left;" | <h2 style="border:none;">Additional information</h2>
|-
|
<big> Online Chat </big>

You may also use the [[w:freenode|freenode]] [[Wikibooks:Contact_us#IRC|IRC]] network to discuss your block with administrators available at the time. If you have an IRC client, the channel can be found at <span class="plainlinks">[irc://irc.freenode.net/wikibooks #wikibooks], or you may use [http://webchat.freenode.net the web interface] (channel: #wikibooks). A list of administrators may be found [[Wikibooks:Administrators|here]]. 

<big> E-mailing us </big>

<div style="padding:5px; margin-bottom:10px; border:1px solid #AAA; background-color:#DFF7FF; text-align:center; font-size:115%;">
Due to the high volume of e-mail we receive, you are more likely to get a quick response if you try requesting an unblock via [[Special:MyTalk|your discussion page]] first.
</div>

Before you e-mail us, please make sure that your situation has not already been addressed by the above sections. [[Special:MyTalk|Your discussion page]] may also contain further details related to your block or [[w:IP address|IP address]], and we strongly advise that you review it before contacting us.

<span style="color:#CC0000;">'''When sending us an e-mail, please copy, paste, and fill out one of these forms.'''</span>
{| style="margin:auto; width:100%; background:transparent;"
| style="vertical-align:top;" |
<div style="margin:auto; padding:5px; width:95%; background:#FAFAFA; border:1px solid #AAA;">
'''To request unblocking:'''
* IP address: <span class="plainlinks">http://en.wikibooks.org/wiki/User_talk:$3</span>
* Blocking admin: $4
* Block reason: <nowiki>$2</nowiki>
* Block originally applied to: $7
* Block ID number: $5
* Your account name (if you have one):
* An explanation of why your block is unfair:
</div>
| style="vertical-align:top;" |
<div style="margin:auto; padding:5px; width:95%; background:#FAFAFA; border:1px solid #AAA;">
'''To request an account:'''
* IP address: <span class="plainlinks">http://en.wikibooks.org/wiki/User_talk:$3</span>
* Blocking admin: $4
* Block reason: <nowiki>$2</nowiki>
* Block originally applied to: $7
* Block ID number: $5
* Preferred account name:
'''Note:''' your preferred username must not be listed as already taken [[Special:ListUsers|here]].
</div>
|}
Even though e-mails sent to the address are viewable only by [[Wikibooks:Administrators|administrators]] and trusted volunteers, '''under no circumstances''' should you e-mail us your account's password or reveal it to anyone else even if asked to. When requesting an account, a temporary password will be generated and e-mailed to you. Please change the password on your new account as soon as possible. You may contact us at wikibooks{{@}}wikimedia.org.

'''IMPORTANT:''' Please do not e-mail us without providing the information requested above. We will always require that information to help you. Any request to make edits to books on your behalf will be disregarded. Correspondence containing legal threats, [[Wikibooks:No personal attacks|personal attacks]], or [[Wikibooks:Be civil|rudeness]] is likely to be ignored and may result in the extension of pre-existing blocks.
|}



<br />ka<br />
{| class="wikitable"
|-
! Header text !! Header text !! Header text !! Header text !! Header text
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|-
| Example || Example || Example || Example || Example
|}
{{armh title|Cortex-A510|arch}}
{{microarchitecture
|atype=CPU
|name=Cortex-A510
|designer=ARM Holdings
|manufacturer=TSMC
|manufacturer 2=Samsung
|manufacturer 3=GlobalFoundries
|manufacturer 4=SMIC
|introduction=May 25, 2021
|process=7 nm
|process 2=6 nm
|process 3=5 nm
|cores=1
|cores 2=2
|type=In-order
|oooe=No
|speculative=Yes
|renaming=No
|decode=3-way
|isa=ARMv9.0
|feature=Hardware virtualization
|extension=FPU
|extension 2=NEON
|extension 3=SVE
|extension 4=SVE2
|extension 5=TrustZone
|l1i=32-64 KiB
|l1i per=core
|l1i desc=4-way set associative
|l1d=32-64 KiB
|l1d per=core
|l1d desc=4-way set associative
|l2=0-512 KiB
|l2 per=cluster
|l2 desc=4-way set associative
|predecessor=Cortex-A55
|predecessor link=arm_holdings/microarchitectures/cortex-a55
|successor=Hayes
|successor link=arm_holdings/microarchitectures/hayes
}}
'''Cortex-A510''' (codename '''Klein''') is an ultra-high efficiency [[microarchitecture]] designed by [[ARM Holdings]] as a successor to the {{armh|Cortex-A55|l=arch}}. The Cortex-A510, which implements the {{arm|ARMv9.0}} ISA, is typically found in smartphone and other embedded devices. Often A510 cores are combined with higher performance processors (e.g. based on {{armh|Cortex-A710|l=arch}}) in {{armh|DynamIQ big.LITTLE}} configuration to achieve better energy/performance.

Note that this microarchitecture is designed as a synthesizable [[IP core]] and is sold to other semiconductor companies to be implemented in their own chips.

== Process Technology ==
The Cortex-A510 was primarily designed to take advantage of [[TSMC]]'s [[N7|7 nm]], [[N6|6 nm]], [[N5|5 nm]] as well as Samsung's [[7LPP|7 nm]] and [[5LPP|5 nm]].

== Architecture ==
The Cortex-A510 is a brand new ground-up CPU design. It borrows advanced processor components from Arm's high-performance cores - such as the branch prediction and prefetchers - to extract high performance from a traditional [[in-order]] core design. The Cortex-A512 is also designed to seamlessly integrate along with higher-performance cores through [[Arm]]'s {{armh|DynamIQ big.LITTLE}} technology.

=== Key changes from {{\\|Cortex-A55}} ===
* Brand new ground-up design
** Higher performance (Arm claims: +35% IPC (SPECint 2006) / +50% IPC (SPECfp 2006)
** Lower power (Arm claims: -20% energy @ [[iso-performance]] / +10% performance @ [[iso-power]])
** Core Complex with a merged core architecture
*** Two independent cores
*** One shared vector unit
**** Configurable 64b or 128b pipe
** Front-End
*** Wider fetch (128b/cycle, up from 64b)
*** Wider decoder (3-way, up from 2-way)
*** New branch predictors
*** New prefetchers
** Back-End
*** In-order
** Memory Subsystem
*** Larger L1 (32-64 KiB)
*** Larger L2 (0-512 KiB, up from 0-256 KiB)
**** 2x bandwidth L2->L3
*** Wider loads (128b/cycle, up from 64b/cycle)
*** 2x loads (2 lds/cycle, up from 1/cycle)
** New ISA Support
*** ARMv9.0 ISA
*** SVE, SVE2 support

=== Block Diagram ===
==== Core Complex ====
:[[File:cortex-a510 block diagram.svg|1000px]]

=== Memory Hierarchy ===
The Cortex-A510 has a private L1I, L1D, and cluster-wide L2 cache.

* Cache
** L1I Cache
*** Private to core
*** 32 KiB OR 64 KiB, 4-way set associative
*** 64-byte cache lines
*** Virtually-indexed, physically-tagged (VIPT) behaving as physically-indexed, physically-tagged (PIPT)
*** Single Error Detect (SED) parity cache protection
*** Pseudo-random cache replacement policy
** L1D Cache
*** Private to core
*** 32KB or 64KB, 4-way set associative
*** Virtually-Indexed, Physically-Tagged (VIPT) behaving as Physically-Indexed, Physically-Tagged (PIPT)
*** Error Correcting Code (ECC) cache protection
*** 64-byte cache lines
*** Pseudo-random cache replacement policy
** L2 Cache
*** Private to complex
*** 128 KiB OR 192 KiB OR 256 KiB OR 384 KiB OR 512 KiB, 8-way set associative
*** 64-byte cache lines
*** Can be configured as 1-2 slices
**** Slice includes: data RAMs, L2 tags, L2 replacement RAM, and L1 duplicate tag RAMs
**** Slice can be configured as single/dual partitions for up to two concurrent accesses to different L2 ways


The Cortex-A510 features an instruction TLB (ITLB) and data TLB (DTLB) which are private to each core and an L2 TLB that is private to the core complex.

* TLBs
** ITLB
*** 16-entries
*** fully associative
*** TLB hits return the PA to the instruction cache
** DTLB
*** 16-entries
*** fully associative
*** TLB hits return the PA to the data cache
** L2 TLB
*** 8-way set associative
*** Shared by both cores in the complex

== Overview ==
[[File:cortex-a510-complex-overview.svg|right|200px]]
The Cortex-A510 is [[Arm]]'s successor to the {{\\|Cortex-A55}} which was introduced four years earlier. Designed to be ultra-low-power and versatile, this core can be used as a standalone CPU in low-power SoCs or serve the efficient core as part of a [[DynamIQ big.LITTLE]] architecture using the {{\\|DSU-110}}. To maintain high efficiency, the Cortex-A510 remains an [[in-order architecture]]. However, by borrowing high-performance components such as state-of-the-art [[branch predictors]] and [[prefetchers]], the Cortex-A510 enjoys significantly higher performance over its predecessor through higher effective instruction stream throughput. The Cortex-A55 is the first [[small core]] from Arm to feature the [[Armv9]] ISA along with the [[Scalable Vector Extension]] (SVE) and SVE2 extensions.

The Cortex-A510 introduces the concept of a Core Complex along with a merged core architecture. A core complex tightly integrates two Cortex-A510 cores, sharing a single common [[level 2 cache]] and vector processing unit (VPU). Like any other Arm IP, the Cortex-A510 complex can be instantiated within a standard DSU cluster as any other core would. The only difference is that you are dealing with two cores at once in a single instance. Because the effective utilization of the vector unit on the small cores is quite low, by implementing a single vector unit for two A510 cores, the silicon area is maintained relatively low while still offering good peak performance when needed.

== Core ==
[[File:a510-fetch.svg|right|400px]]
The Cortex-A510 is a new ground-up design. Compared to its predecessor, the {{\\|Cortex-A55}}, the A510 is wider, but the underlying architecture remains an [[in-order pipeline]]. The choice to remain with an in-order pipeline design was driven by the overall simplicity of the design which maximizes power efficiency and area usage. Each Cortex-A510 core integrates its own private [[level 1 instruction cache]] and private [[level 1 data cache]].


The Cortex-A510 features a 10-stage pipeline.


:[[File:a510-pipeline.svg|700px]]

=== Front-end ===
The front-end of the Cortex-A510 fetches instructions from the private [[level 1 instruction cache]] and streams them to [[instruction decode|decode]]. The L1I$ on the A510 can be configured as 32 KiB or 64 KiB, both are 4-way [[set associative]].

==== Fetch & Decode ====
[[File:a510-decode.svg|right|250px]]
The A510 speculatively fetches instructions streams. The A510 can predict conditional, unconditional, A32/T32 branch switches, and indirect procedure call/return branches. Each cycle, the A510 can [[instruction fetch|fetch]] up to 128b (16B) from the [[instruction cache]]; twice the fetch bandwidth of the {{\\|Cortex-A55}}. This allows it to deliver up to four instructions to decode.

Because the A510 is an in-order design, the performance of the core largely depends on the ability of the hardware to correctly predict future instruction stream patterns and paths. Advanced [[branch predictors]] along with aggressive [[prefetchers]] that were taken from [[Arm]]'s big cores (i.e., {{\\|Cortex-A710}}) were integrated into the A510 in order to allow for higher effective instruction stream bandwidth.

From the fetch stream, up to three instructions are sent to decode. The Cortex-A510 features a 3-wide decode, 50% wider than the {{\\|Cortex-A55}}.

=== Back-end ===
[[File:a510-eus.svg|right|400px]]
Each cycle, up to three instructions are decoded and sent to the back-end for execution. To match fetch and decode width, the A510 features a 3-wide issue and execute. The A510 is 50% wider than its predecessor, the {{\\|Cortex-A55}} which was 2-wide.

Each cycle, up to three μOPs may be sent to execution. The A510 features a 10-wide pipe. Compared to the {{\\|Cortex-A55}}, Arm added a third ALU pipeline to increase integer workload bandwidth. With a 3-wide issue, the A510 can issue up to three ALU instructions to the three pipelines each cycle. There is a separate pipe for complex ALU operations as well as a separate ALU for performing branch operations.

Due to the nature of an in-order pipeline, Arm says a lot of work went into the forwarding capabilities of the design and the multi-issue capabilities of the execution engines. Special care was taken to maintain the execution bandwidth up in more cases while being less sensitive to specific instruction scheduling. Due to the inter-instruction dependencies nature of an in-order pipeline, poorly scheduled code can create backups and significantly reduce the effective execution throughput. Arm stated that in the Cortex-A510, a lot of attention has gone into maximizing this concern while keeping everything in order for maximum efficiency.

=== Memory subsystem ===
[[File:a510-mem.svg|right|250px]]
The Cortex-A510 has been significantly improved over the A55. Like the {{\\|Cortex-A55}}, the Cortex-A510 maintains two memory pipelines. Whereas the A55 could perform one load on one pipe and one store on the other, the Cortex-A510 is capable of executing loads on both pipes. In other words, the A510 supports up to two loads per cycle or - alternatively - one load and one store per cycle. Arm also doubled the width of the data loads on the A510. Data loads are 128b-wide now, double the 64b-wide loads of the {{\\|Cortex-A55}}. Because the A510 supports up to two loads per cycle or 2x128b loads, the peak L1 cache bandwidth is now 4x compared to the A55.

The Cortex-A510 features a private L1 data cache which can be configured with a capacity of 32 KiB or 64 KiB. Both configurations are organized as a 4-way set associative. The bandwidth from the other caches was also doubled – the L2 and L3 bandwidths are doubled that of the Cortex-A55. 

Data prefetching on the A510 also received special attention. Arm stated that technologies that were originally developed for the {{\\|Cortex-X2}} were carried over to the Cortex-A510 design. There are multiple complementary data prefetching algorithms implemented on the Cortex-A510.

{{clear}}

== Core Complex ==
[[File:a510-vpu.svg|right|300px]]
That Cortex-A510 is integrated as part of a ''Complex''. Normally, each complex would incorporate two Cortex-A510 cores. Both cores share a complex-private L2 cache and a complex-private vector processing unit. The rationale for a complex is the ability for the A510 to fully support the [[Armv9.0]] ISA including the [[Scalable Vector Extension]] (SVE) and [[SVE2]]. With a single shared vector processing unit, the A510 can support peak vector performance when needed while conserving silicon area due to lower utilization when compared to the [[big cores]]. While the two-core complex is the default and encouraged configuration, customers can opt to use it in a single-core complex, allowing for a dedicated VPU and L2 cache but at higher silicon area.

=== Vector Processing Unit (VPU) ===
A major component of the core complex is the vector processing unit (VPU) which is private to the complex and is shared by both Cortex-A510's. Because typical vector workloads will usually be assigned to the [[big cores]], the VPU utilization on the [[small cores]] is expected to be relatively low. To maximize silicon area usage while providing sufficient vector performance, a single VPU instance is shared by both cores.

Once a μOP is issued to the VPU from one of the cores it is sent to the arbitration unit. Arbitration is completely hardware-based with execution sharing and utilization being entirely transparent to software. There is no user or OS-level software intervention for the arbitration of vector unit accesses between the cores. Furthermore, from a software perspective, each core appears to have dedicated access to the vector unit. Scheduling is very fine-grained between the cores to make sure there is minimal overhead caused by the sharing. There is no penalty for sharing or using the VPU. Arm says that they see very minimal performance overhead due to the sharing of the VPU. In their own internal testing, on the SPECfp 2006 benchmark, they see around a 1% performance degradation if both cores are executing FP/vector instructions at the same time as they compete to use the vector datapath.

All scalar and floating-point NEON and SVE2 instructions are executed through the vector datapath. There is support for optional crypto as well. This datapath itself is configurable. Possible configurations include having two 64b or two 128b data pipes - twice the peak bandwidth delta between the two configurations. A configuration of 2x128b data pipes will offer twice the peak compute of the {{\\|Cortex-55}}. Depending on the configuration, designers have the flexibility in deciding between area and performance depending on their application and workload characterization. In the 2x128b configuration, the Cortex-A510 complex can achieve 16 FP16 FLOPS/clk. Note that regardless of the vector unit size or configuration, for SVE, there is a 128b vector register file.

== Performance claims ==
At the peak operating point of the {{\\|Cortex-A55}}, the Cortex-A510 is capable of reaching the same performance at 20% less energy. Alternatively, within the same power budget, the A510 achieves 10% higher performance. Compared to the {{\\|Cortex-A55}}, Arm says that the Cortex-A510 has around 35% higher IPC on [[SPECint 2006]]. On the SPECfp 2006 benchmark, Arm claims to see an even higher performance improvement at around 50%.

{|
| [[File:cortex-a510 perf claims.png|500px]] || [[File:cortex-a510-spec-perf.png|500px|DVFS of power vs performance at ISO-process on SPECint 2006 for both the Cortex-A55 and the new Cortex-A510.]]
|}

{{clear}}

== Die ==
Advanced implementation die floorplan.

:[[File:cortex-a510 die plot.jpg|500px|class=wikichip_ogimage]]

== All Cortex-A510 Processors ==
<!-- NOTE: 
           This table is generated automatically from the data in the actual articles.
           If a microprocessor is missing from the list, an appropriate article for it needs to be
           created and tagged accordingly.

           Missing a chip? please dump its name here: https://en.wikichip.org/wiki/WikiChip:wanted_chips
-->
{{comp table start}}
<table class="comptable sortable tc4 tc6 tc9">
{{comp table header|main|8:List of Cortex-A510-based Processors}}
{{comp table header|main|6:Main processor|2:Integrated Graphics}}
{{comp table header|cols|Family|Launched|Process|Arch|Cores|%Frequency|GPU|%Frequency}}
{{#ask: [[Category:all microprocessor models]] [[microarchitecture::Cortex-A510]]
 |?full page name
 |?model number
 |?family
 |?first launched
 |?process
 |?microarchitecture
 |?core count
 |?base frequency#GHz
 |?integrated gpu
 |?integrated gpu base frequency
 |format=template
 |template=proc table 3
 |userparam=10
 |mainlabel=-
 |valuesep=,
}}
{{comp table count|ask=[[Category:all microprocessor models]] [[microarchitecture::Cortex-A510]]}}
</table>
{{comp table end}}

== Bibliography ==
* Arm Tech Day, 2021
* Arm. personal communication. 2021.

== Documents ==
* [[:File:arm cortex a510 sog.pdf|Cortex-A510 Software Optimization Guide]]
* [[:File:arm cortex a510 trm.pdf|Cortex-A510 Technical Reference Manual]]
codename	Cortex-A510 +
core count	1 + and 2 +
designer	ARM Holdings +
first launched	May 25, 2021 +
full page name	arm holdings/microarchitectures/cortex-a510 +
instance of	microarchitecture +
instruction set architecture	ARMv9.0 +
manufacturer	TSMC +, Samsung +, GlobalFoundries + and SMIC +
microarchitecture type	CPU +
name	Cortex-A510 +
process	7 nm (0.007 μm, 7.0e-6 mm) +, 6 nm (0.006 μm, 6.0e-6 mm) + and 5 nm (0.005 μm, 5.0e-6 mm) +