Editing qualcomm/microarchitectures/cloud ai 100

{{qualcomm title|Cloud AI 100|arch}}
{{microarchitecture
|atype=NPU
|name=Cloud AI 100
|designer=Qualcomm
|manufacturer=TSMC
|introduction=March, 2021
|process=7 nm
|processing elements=16
|type=VLIW
|decode=4-way
|l2=1 MiB
|l2 per=core
|side cache=8 MiB
|side cache per=core
}}
'''Cloud AI 100''' is an [[NPU]] microarchitecture designed by [[Qualcomm]] for the server and edge market. Those NPUs are sold under the {{qualcomm|Cloud AI}} brand.

== Process Technology ==
The Cloud AI 100 SoC is fabricated on TSMC's [[7-nanometer process]].

== Architecture ==

=== Key Features ===

== Block Diagram ==

=== SoC ===
:[[File:cloud ai 100 soc.svg|700px]]

=== AI Core ===
:[[File:cloud ai 100 ai core.svg|350px]]

== Memory Hierarchy ==
* L1D$ / L1I$
** Private per AI Core
* L2
** 1 MiB / AI Core
* Vector Tightly-Coupled Memory (VTCM)
** 8 MiB / AI Core
* DRAM
** 8-32 GiB
*** LPDDR4x-4266
**** 68.25 - 136.5 GB/s

== Overview ==

== AI Core ==

== Performance claims ==
Performance-per-watt was published by Quall based on an Int8 3×3 convolution operation with uniformly distributed weights and input action comprising 50% zeros which Qualcomm says is typical for Deep CNN with Relu operators. To that end, Qualcomm says the AI 100 can achieve up to ~150 TOPs at ~12 W at over 12 TOPS/W in edge cases and ~363 TOPs at under 70 W at 5.24 TOPs/W in data center uses. Numbers are at the SoC level.

{| class="wikitable"
! SoC Power
| 12.05 W || 19.74 W || 69.26 W
|-
! TOPS
| 149.01 || 196.94 || 363.02 
|-
! TOPS/W
| 12.37 || 9.98 || 5.24
|}

== Bibliography ==
* Linley Fall Processor Conference 2021
* {{bib|hc|33|Qualcomm}}
codename	Cloud AI 100 +
designer	Qualcomm +
first launched	March 2021 +
full page name	qualcomm/microarchitectures/cloud ai 100 +
instance of	microarchitecture +
manufacturer	TSMC +
name	Cloud AI 100 +
process	7 nm (0.007 μm, 7.0e-6 mm) +
processing element count	16 +