<s>
Tegra	O
is	O
a	O
system	B-Architecture
on	I-Architecture
a	I-Architecture
chip	I-Architecture
(	O
SoC	O
)	O
series	O
developed	O
by	O
Nvidia	O
for	O
mobile	O
devices	O
such	O
as	O
smartphones	B-Application
,	O
personal	B-Application
digital	I-Application
assistants	I-Application
,	O
and	O
mobile	B-Device
Internet	I-Device
devices	I-Device
.	O
</s>
<s>
The	O
Tegra	O
integrates	O
an	O
ARM	B-Architecture
architecture	I-Architecture
central	B-General_Concept
processing	I-General_Concept
unit	I-General_Concept
(	O
CPU	B-General_Concept
)	O
,	O
graphics	B-Architecture
processing	I-Architecture
unit	I-Architecture
(	O
GPU	B-Architecture
)	O
,	O
northbridge	B-Device
,	O
southbridge	B-Device
,	O
and	O
memory	B-General_Concept
controller	I-General_Concept
onto	O
one	O
package	O
.	O
</s>
<s>
The	O
Tegra-line	O
evolved	O
to	O
emphasize	O
performance	O
for	O
gaming	O
and	O
machine	O
learning	O
applications	O
without	O
sacrificing	O
power	O
efficiency	O
,	O
before	O
taking	O
a	O
drastic	O
shift	O
in	O
direction	O
towards	O
platforms	O
that	O
provide	O
vehicular	O
automation	O
with	O
the	O
applied	O
"	O
Nvidia	O
Drive	O
"	O
brand	O
name	O
on	O
reference	O
boards	O
and	O
its	O
semiconductors	O
;	O
and	O
with	O
the	O
"	O
Nvidia	O
Jetson	O
"	O
brand	O
name	O
for	O
boards	O
adequate	O
for	O
AI	B-Application
applications	I-Application
within	O
e.g.	O
</s>
<s>
The	O
APX	O
chips	O
were	O
designed	O
for	O
smartphones	B-Application
,	O
while	O
the	O
Tegra	O
600	O
and	O
650	O
chips	O
were	O
intended	O
for	O
smartbooks	B-Device
and	O
mobile	B-Device
Internet	I-Device
devices	I-Device
(	O
MID	O
)	O
.	O
</s>
<s>
The	O
first	O
product	O
to	O
use	O
the	O
Tegra	O
was	O
Microsoft	O
's	O
Zune	O
HD	O
media	O
player	O
in	O
September	O
2009	O
,	O
followed	O
by	O
the	O
Samsung	B-Application
M1	O
.	O
</s>
<s>
Microsoft	O
's	O
Kin	B-Device
was	O
the	O
first	O
cellular	O
phone	O
to	O
use	O
the	O
Tegra	O
;	O
however	O
,	O
the	O
phone	O
did	O
not	O
have	O
an	O
app	O
store	O
,	O
so	O
the	O
Tegra	O
's	O
power	O
did	O
not	O
provide	O
much	O
advantage	O
.	O
</s>
<s>
In	O
September	O
2008	O
,	O
Nvidia	O
and	O
Opera	B-Application
Software	B-Architecture
announced	O
that	O
they	O
would	O
produce	O
a	O
version	O
of	O
the	O
Opera	B-Application
9.5	I-Application
browser	O
optimized	O
for	O
the	O
Tegra	O
on	O
Windows	B-Operating_System
Mobile	I-Operating_System
and	O
Windows	B-Operating_System
CE	I-Operating_System
.	O
</s>
<s>
At	O
Mobile	O
World	O
Congress	O
2009	O
,	O
Nvidia	O
introduced	O
its	O
port	O
of	O
Google	B-Application
's	I-Application
Android	I-Application
to	O
the	O
Tegra	O
.	O
</s>
<s>
On	O
January	O
7	O
,	O
2010	O
,	O
Nvidia	O
officially	O
announced	O
and	O
demonstrated	O
its	O
next	O
generation	O
Tegra	O
system-on-a-chip	B-Architecture
,	O
the	O
Nvidia	O
Tegra	O
250	O
,	O
at	O
Consumer	O
Electronics	O
Show	O
2010	O
.	O
</s>
<s>
Nvidia	O
primarily	O
supports	O
Android	B-Application
on	O
Tegra	O
2	O
,	O
but	O
booting	B-Operating_System
other	O
ARM-supporting	O
operating	O
systems	O
is	O
possible	O
on	O
devices	O
where	O
the	O
bootloader	B-Operating_System
is	O
accessible	O
.	O
</s>
<s>
Tegra	O
2	O
support	O
for	O
the	O
Ubuntu	B-Operating_System
Linux	I-Operating_System
distribution	I-Operating_System
was	O
also	O
announced	O
on	O
the	O
Nvidia	O
developer	O
forum	O
.	O
</s>
<s>
Nvidia	O
announced	O
the	O
first	O
quad-core	B-Architecture
SoC	O
at	O
the	O
February	O
2011	O
Mobile	O
World	O
Congress	O
event	O
in	O
Barcelona	O
.	O
</s>
<s>
One	O
VCM	O
powers	O
the	O
17-inch	O
touchscreen	B-Device
infotainment	O
system	O
,	O
and	O
one	O
drives	O
the	O
12.3-inch	O
all	O
digital	O
instrument	O
cluster.	O
"	O
</s>
<s>
At	O
the	O
announcement	O
event	O
,	O
Nvidia	O
showed	O
off	O
Epic	O
Games	O
 '	O
Unreal	B-Operating_System
Engine	I-Operating_System
4	O
"	O
Elemental	O
"	O
demo	O
,	O
running	O
on	O
a	O
Tegra	O
X1	O
.	O
</s>
<s>
On	O
October	O
20	O
,	O
2016	O
,	O
Nvidia	O
announced	O
that	O
the	O
Nintendo	B-Application
Switch	I-Application
hybrid	O
video	O
game	O
console	O
will	O
be	O
powered	O
by	O
Tegra	O
hardware	O
.	O
</s>
<s>
On	O
March	O
15	O
,	O
2017	O
,	O
TechInsights	O
revealed	O
the	O
Nintendo	B-Application
Switch	I-Application
is	O
powered	O
by	O
a	O
custom	O
Tegra	O
X1	O
(	O
model	O
T210	O
)	O
,	O
with	O
lower	O
clockspeeds	O
.	O
</s>
<s>
The	O
second	O
generation	O
Tegra	O
SoC	O
has	O
a	O
dual-core	B-Architecture
ARM	B-Application
Cortex-A9	I-Application
CPU	B-General_Concept
,	O
an	O
ultra	O
low	O
power	O
(	O
ULP	O
)	O
GeForce	B-Application
GPU	B-Architecture
,	O
a	O
32-bit	O
memory	B-General_Concept
controller	I-General_Concept
with	O
either	O
LPDDR2-600	O
or	O
DDR2-667	O
memory	O
,	O
a	O
32KB/32KB	O
L1	O
cache	O
per	O
core	O
and	O
a	O
shared	O
1MB	O
L2	O
cache	O
.	O
</s>
<s>
Tegra	O
2	O
's	O
Cortex	B-Application
A9	I-Application
implementation	O
does	O
not	O
include	O
ARM	O
's	O
SIMD	O
extension	O
,	O
NEON	O
.	O
</s>
<s>
There	O
is	O
a	O
version	O
of	O
the	O
Tegra	O
2	O
SoC	O
supporting	O
3D	O
displays	O
;	O
this	O
SoC	O
uses	O
a	O
higher	O
clocked	O
CPU	B-General_Concept
and	O
GPU	B-Architecture
.	O
</s>
<s>
The	O
lack	O
of	O
support	O
for	O
high-profile	O
H.264	B-Application
is	O
particularly	O
troublesome	O
when	O
using	O
online	O
video	O
streaming	O
services	O
.	O
</s>
<s>
Modelnumber	O
CPU	B-General_Concept
GPU	B-Architecture
Memory	O
Adoption	O
Processor	O
Cores	O
Frequency	O
Micro-architecture	O
Coreconfiguration1	O
Frequency	O
Type	O
Amount	O
Buswidth	O
Band-width	O
Availability	O
AP20H	O
(	O
Ventana/Unknown	O
)	O
Cortex-A9	B-Application
2	O
1.0	O
GHz	O
VLIW-basedVEC4	O
units	O
4:4:4:4	O
300	O
MHz	O
LPDDR2	O
300	O
MHzDDR2	O
333	O
MHz	O
?	O
</s>
<s>
Model	O
Devices	O
AP20H	O
Motorola	B-Application
Atrix	I-Application
4G	I-Application
,	O
Motorola	O
Droid	O
X2	O
,	O
Motorola	B-Application
Photon	I-Application
,	O
LG	B-Application
Optimus	I-Application
2X	I-Application
/	O
LG	O
Optimus	O
Dual	O
P990	O
/	O
Optimus	O
2x	O
SU660	O
(	O
?	O
</s>
<s>
NVIDIA	O
's	O
Tegra	O
3	O
(	O
codenamed	O
"	O
Kal-El	O
"	O
)	O
is	O
functionally	O
a	O
SoC	O
with	O
a	O
quad-core	B-Architecture
ARM	B-Application
Cortex-A9	I-Application
MPCore	I-Application
CPU	B-General_Concept
,	O
but	O
includes	O
a	O
fifth	O
"	O
companion	O
"	O
core	O
in	O
what	O
Nvidia	O
refers	O
to	O
as	O
a	O
"	O
variable	O
SMP	B-Operating_System
architecture	O
"	O
.	O
</s>
<s>
While	O
all	O
cores	O
are	O
Cortex-A9s	B-Application
,	O
the	O
companion	O
core	O
is	O
manufactured	O
with	O
a	O
low-power	O
silicon	O
process	O
.	O
</s>
<s>
The	O
main	O
quad-core	B-Architecture
portion	O
of	O
the	O
CPU	B-General_Concept
powers	O
off	O
in	O
these	O
situations	O
.	O
</s>
<s>
The	O
GPU	B-Architecture
in	O
Tegra	O
3	O
is	O
an	O
evolution	O
of	O
the	O
Tegra	O
2	O
GPU	B-Architecture
,	O
with	O
4	O
additional	O
pixel	O
shader	O
units	O
and	O
higher	O
clock	O
frequency	O
.	O
</s>
<s>
It	O
can	O
also	O
output	O
video	O
up	O
to	O
2560×1600	O
resolution	O
and	O
supports	O
1080p	O
MPEG-4	B-Application
AVC/h.264	I-Application
40	O
Mbit/s	O
High-Profile	O
,	O
VC1-AP	O
,	O
and	O
simpler	O
forms	O
of	O
MPEG-4	B-Algorithm
such	O
as	O
DivX	O
and	O
Xvid	O
.	O
</s>
<s>
Modelnumber	O
CPU	B-General_Concept
GPU	B-Architecture
Memory	O
Adoption	O
Processor	O
Cores	O
Frequency( multi	O
-	O
/	O
single-core	O
mode	O
)	O
Micro-architecture	O
Coreconfiguration1	O
Frequency	O
Type	O
Amount	O
Buswidth	O
Band-width	O
Availability	O
T30L	O
Cortex-A9	B-Application
4+1	O
1.2	O
GHz	O
/	O
up	O
to	O
1.3	O
GHz	O
VLIW-basedVEC4	O
units	O
8:4:8:8	O
416	O
MHz	O
DDR3-1333	O
?	O
</s>
<s>
The	O
Tegra4	O
(	O
codenamed	O
"	O
Wayne	O
"	O
)	O
was	O
announced	O
on	O
January	O
6	O
,	O
2013	O
,	O
and	O
is	O
a	O
SoC	O
with	O
a	O
quad-core	B-Architecture
CPU	B-General_Concept
,	O
but	O
includes	O
a	O
fifth	O
low-power	O
Cortex	O
A15	O
companion	O
core	O
which	O
is	O
invisible	O
to	O
the	O
OS	O
and	O
performs	O
background	O
tasks	O
to	O
save	O
power	O
.	O
</s>
<s>
This	O
power-saving	O
configuration	O
is	O
referred	O
to	O
as	O
"	O
variable	O
SMP	B-Operating_System
architecture	O
"	O
and	O
operates	O
like	O
the	O
similar	O
configuration	O
in	O
Tegra3	O
.	O
</s>
<s>
The	O
GeForce	B-Application
GPU	B-Architecture
in	O
Tegra4	O
is	O
again	O
an	O
evolution	O
of	O
its	O
predecessors	O
.	O
</s>
<s>
In	O
3D	O
tests	O
,	O
the	O
Tegra4	O
GPU	B-Architecture
is	O
typically	O
several	O
times	O
faster	O
than	O
that	O
of	O
Tegra3	O
.	O
</s>
<s>
Additionally	O
,	O
the	O
Tegra4	O
video	O
processor	O
has	O
full	O
support	O
for	O
hardware	O
decoding	O
and	O
encoding	O
of	O
WebM	B-Application
video	O
(	O
up	O
to	O
1080p	O
60Mbit/s	O
@	O
60fps	O
)	O
.	O
</s>
<s>
Along	O
with	O
Tegra4	O
,	O
Nvidia	O
also	O
introduced	O
i500	O
,	O
an	O
optional	O
software	B-Architecture
modem	O
based	O
on	O
Nvidia	O
's	O
acquisition	O
of	O
Icera	O
,	O
which	O
can	O
be	O
reprogrammed	O
to	O
support	O
new	O
network	O
standards	O
.	O
</s>
<s>
Modelnumber	O
CPU	B-General_Concept
GPU	B-Architecture
Memory	O
Adoption	O
Processor	O
Cores	O
Frequency	O
Microarchitecture	O
Coreconfiguration1	O
Frequency	O
Type	O
Amount	O
Buswidth	O
Band-width	O
Availability	O
T114	O
Cortex-A15	O
4+1	O
up	O
to	O
1.9	O
GHz	O
VLIW-based	O
VEC4	O
units	O
72	O
(	O
48:24:4	O
)	O
672	O
MHz	O
DDR3L	O
or	O
LPDDR3	O
?	O
</s>
<s>
With	O
hardware	O
support	O
for	O
the	O
same	O
audio	O
and	O
video	O
formats	O
,	O
but	O
using	O
Cortex-A9	B-Application
cores	O
instead	O
of	O
Cortex-A15	O
,	O
the	O
Tegra4i	O
is	O
a	O
low-power	O
variant	O
of	O
the	O
Tegra4	O
and	O
is	O
designed	O
for	O
phones	O
and	O
tablets	O
.	O
</s>
<s>
Modelnumber	O
CPU	B-General_Concept
GPU	B-Architecture
Memory	O
Adoption	O
Processor	O
Cores	O
Frequency	O
Microarchitecture	O
Coreconfiguration1	O
Frequency	O
Type	O
Amount	O
Buswidth	O
Band-width	O
Availability	O
T148	O
?	O
</s>
<s>
Nvidia	O
's	O
Tegra	O
K1	O
(	O
codenamed	O
"	O
Logan	O
"	O
)	O
features	O
ARM	O
Cortex-A15	O
cores	O
in	O
a	O
4+1	O
configuration	O
similar	O
to	O
Tegra4	O
,	O
or	O
Nvidia	O
's	O
64-bit	B-Device
Project	O
Denver	O
dual-core	B-Architecture
processor	I-Architecture
as	O
well	O
as	O
a	O
Kepler	B-General_Concept
graphics	B-Architecture
processing	I-Architecture
unit	I-Architecture
with	O
support	O
for	O
Direct3D	O
12	O
,	O
OpenGL	B-Application
ES	I-Application
3.1	O
,	O
CUDA	B-Architecture
6.5	O
,	O
OpenGL	O
4.4/OpenGL	O
4.5	O
,	O
and	O
Vulkan	B-Application
.	O
</s>
<s>
In	O
late	O
April	O
2014	O
,	O
Nvidia	O
shipped	O
the	O
"	O
Jetson	O
TK1	O
"	O
development	O
board	O
containing	O
a	O
Tegra	O
K1	O
SoC	O
and	O
running	O
Ubuntu	B-Operating_System
Linux	I-Operating_System
.	O
</s>
<s>
Modelnumber	O
CPU	B-General_Concept
GPU	B-Architecture
Memory	O
Adoption	O
Processor	O
Cores	O
Frequency	O
Micro-architecture	O
Coreconfiguration1	O
Frequency	O
GFLOPS(FP32 )	O
Type	O
Amount	O
Buswidth	O
Band-width	O
Availability	O
T124	O
Cortex-A15	O
R3(32-bit )	O
4+1	O
up	O
to	O
2.3	O
GHz	O
GK20A(Kepler )	O
192:8:4	O
756	O
–	O
951	O
MHz	O
290	O
–	O
365	O
DDR3LLPDDR3	O
max	O
8	O
GBwith	O
40-bit	O
address	O
extension2	O
64	B-Device
bit	I-Device
17	O
GB/s	O
Q2	O
2014	O
T132	O
Denver(64-bit )	O
2	O
up	O
to	O
2.5	O
GHz	O
max	O
8	O
GB	O
?	O
</s>
<s>
Released	O
in	O
2015	O
,	O
Nvidia	O
's	O
Tegra	O
X1	O
(	O
codenamed	O
"	O
Erista	O
"	O
)	O
features	O
two	O
CPU	B-General_Concept
clusters	O
,	O
one	O
with	O
four	O
ARM	O
Cortex-A57	O
cores	O
and	O
the	O
other	O
with	O
four	O
ARM	O
Cortex-A53	O
cores	O
,	O
as	O
well	O
as	O
a	O
Maxwell-based	O
graphics	B-Architecture
processing	I-Architecture
unit	I-Architecture
.	O
</s>
<s>
Only	O
one	O
cluster	O
of	O
cores	O
can	O
be	O
active	O
at	O
once	O
,	O
with	O
the	O
cluster	O
switch	O
being	O
handled	O
by	O
software	B-Architecture
on	O
the	O
BPMP-L	O
.	O
Devices	O
utilizing	O
the	O
Tegra	O
X1	O
have	O
only	O
been	O
seen	O
to	O
utilize	O
the	O
cluster	O
with	O
the	O
more	O
powerful	O
ARM	O
Cortex-A57	O
cores	O
.	O
</s>
<s>
The	O
Tegra	O
X1	O
was	O
found	O
to	O
be	O
vulnerable	O
to	O
a	O
Fault	O
Injection	O
(	O
FI	O
)	O
voltage	O
glitching	O
attack	O
,	O
which	O
allowed	O
for	O
arbitrary	O
code	O
execution	O
and	O
homebrew	O
software	B-Architecture
on	O
the	O
devices	O
it	O
was	O
implemented	O
in	O
.	O
</s>
<s>
CPU	B-General_Concept
:	O
ARMv8	O
ARM	O
Cortex-A57	O
quad-core	B-Architecture
(	O
64-bit	B-Device
)	O
+	O
(	O
unused	O
?	O
)	O
</s>
<s>
TSMC	O
16	B-Algorithm
nm	I-Algorithm
process	I-Algorithm
for	O
the	O
Tegra	O
X1+	O
.	O
</s>
<s>
TDP	B-General_Concept
:	O
</s>
<s>
TM671D-A2TSMC	O
20	O
nmCortex-A57	O
+	O
Cortex-A53	O
A57	O
:	O
4A53	O
:	O
4	O
A57	O
:	O
2.2	O
GHzTegra	O
T210	O
dfll	O
tableA53	O
:	O
1.3	O
GHz	O
GM20B(Maxwell )	O
256	O
:	O
16:16	O
1000	O
MHz	O
512	O
1024	O
LPDDR3/LPDDR4	O
8	O
GB	O
64	O
bit	O
25.6	O
GB/s	O
Q2	O
2015TM660M-A2A57	O
:	O
1.428	O
GHzA53	O
:	O
?	O
</s>
<s>
Nvidia	O
's	O
Tegra	O
X2	O
(	O
codenamed	O
"	O
Parker	B-Device
"	O
)	O
features	O
Nvidia	O
's	O
own	O
custom	O
general-purpose	O
ARMv8-compatible	O
core	O
Denver	O
2	O
as	O
well	O
as	O
code-named	O
Pascal	B-General_Concept
graphics	O
processing	O
core	O
with	O
GPGPU	B-Architecture
support	O
.	O
</s>
<s>
It	O
contains	O
7	O
billion	O
transistors	O
and	O
8	O
custom	O
ARMv8	O
cores	O
,	O
a	O
Volta	B-General_Concept
GPU	B-Architecture
with	O
512	O
CUDA	B-Architecture
cores	O
,	O
an	O
open	O
sourced	O
TPU	B-Device
(	O
Tensor	B-Device
Processing	I-Device
Unit	I-Device
)	O
called	O
DLA	B-General_Concept
(	O
Deep	B-General_Concept
Learning	I-General_Concept
Accelerator	I-General_Concept
)	O
.	O
</s>
<s>
It	O
is	O
able	O
to	O
encode	O
and	O
decode	O
8K	B-Architecture
Ultra	I-Architecture
HD	I-Architecture
(	O
7680×4320	O
)	O
.	O
</s>
<s>
Users	O
can	O
configure	O
operating	O
modes	O
at	O
10W	O
,	O
15W	O
,	O
and	O
30W	O
TDP	B-General_Concept
as	O
needed	O
and	O
the	O
die	O
size	O
is	O
350mm2	O
.	O
</s>
<s>
Model	O
SOC	O
VariantDevicesT194	O
Nvidia	O
Drive	O
Xavier	O
(	O
Drive	O
PX-series	O
)	O
(	O
formerly	O
named	O
Xavier	O
AI	O
Car	O
Supercomputer	O
)	O
Nvidia	O
Drive	O
Pegasus	B-Operating_System
(	O
Drive	O
PX-series	O
)	O
Nvidia	O
Drive	O
AGX	O
Xavier	O
Developer	O
Kit	O
Nvidia	O
Jetson	O
AGX	O
Xavier	O
Developer	O
Kit	O
Nvidia	O
Jetson	O
XavierTE860M-A2	O
Nvidia	O
Jetson	O
Xavier	O
NX	O
Nvidia	O
Clara	O
AGX	O
"	O
Clara	O
AGX	O
is	O
based	O
on	O
NVIDIA	O
Xavier	O
and	O
NVIDIA	O
Turing	O
GPUs.	O
"	O
</s>
<s>
On	O
the	O
Linux	B-Operating_System
Kernel	I-Operating_System
Mailing	O
List	O
,	O
a	O
Tegra194	O
based	O
development	O
board	O
with	O
type	O
ID	O
"	O
P2972-0000	O
"	O
got	O
reported	O
:	O
The	O
board	O
consists	O
of	O
the	O
P2888	O
compute	O
module	O
and	O
the	O
P2822	O
baseboard	O
.	O
</s>
<s>
Nvidia	O
announced	O
the	O
next-gen	O
SoC	O
codename	O
Orin	O
on	O
March	O
27	O
,	O
2018	O
,	O
at	O
GPU	B-Architecture
Technology	O
Conference	O
2018	O
.	O
</s>
<s>
GPU	B-Architecture
:	O
Ampere-based	O
,	O
2048	O
CUDA	B-Architecture
cores	O
and	O
64	O
tensor	O
cores	O
;	O
"	O
with	O
up	O
to	O
131	O
Sparse	O
TOPs	O
of	O
INT8	O
Tensor	O
compute	O
,	O
and	O
up	O
to	O
5.32	O
FP32	O
TFLOPs	O
of	O
CUDA	B-Architecture
compute.	O
"	O
</s>
<s>
Nvidia	O
announced	O
the	O
latest	O
member	O
of	O
the	O
family	O
,	O
"	O
Orin	O
Nano	O
"	O
in	O
September	O
2022	O
at	O
the	O
GPU	B-Architecture
Technology	O
Conference	O
2022	O
.	O
</s>
<s>
Nvidia	O
announced	O
the	O
next-gen	O
SoC	O
codename	O
Atlan	O
on	O
April	O
12	O
,	O
2021	O
,	O
at	O
GPU	B-Architecture
Technology	O
Conference	O
2021	O
.	O
</s>
<s>
Nvidia	O
announced	O
the	O
next-gen	O
SoC	O
codename	O
Thor	O
on	O
September	O
20	O
,	O
2022	O
,	O
at	O
GPU	B-Architecture
Technology	O
Conference	O
2022	O
,	O
replacing	O
the	O
cancelled	O
Atlan	O
.	O
</s>
<s>
?	O
CPUInstruction	O
setARMv7-A	O
(	O
32	O
bit	O
)	O
ARMv8-A	O
(	O
64	B-Device
bit	I-Device
)	O
ARMv8.2-A	O
(	O
64	B-Device
bit	I-Device
)	O
?	O
ARMv9.x-A	O
(	O
64	B-Device
bit	I-Device
)	O
Cores2	O
A94+1	O
A94+1	O
A154+1	O
A94+1	O
A152	O
Denver4	O
A53	O
(	O
disabled	O
)	O
+	O
4	O
A574	O
A572	O
Denver2	O
+	O
4	O
A578	O
Nvidia	O
Carmel12	O
Arm	O
Cortex-A78AE	O
(	O
Hercules	O
)	O
Nvidia	O
Grace-NextArm	O
Neoverse	O
Poseidon	O
AEL1	O
Cache	O
(	O
I	O
/	O
D	O
)	O
32	O
/	O
32	O
KB128	O
/	O
64	O
KB32	O
/	O
32	O
KB	O
+	O
64	O
/	O
32	O
KB128	O
/	O
64	O
KB	O
+	O
48	O
/	O
32	O
KB128	O
/	O
64	O
KB64	O
/	O
64	O
KB	O
?	O
</s>
<s>
*	O
VLIW-based	O
Vec4	O
:	O
Pixel	O
shaders	O
+	O
Vertex	O
shaders	O
.	O
</s>
<s>
Since	O
Kepler	B-General_Concept
,	O
Unified	O
shaders	O
are	O
used	O
.	O
</s>
<s>
FreeBSD	B-Operating_System
supports	O
a	O
number	O
of	O
different	O
Tegra	O
models	O
and	O
generations	O
,	O
ranging	O
from	O
Tegra	O
K1	O
,	O
to	O
Tegra	O
210	O
.	O
</s>
<s>
Nvidia	O
distributes	O
proprietary	O
device	O
drivers	O
for	O
Tegra	O
through	O
OEMs	O
and	O
as	O
part	O
of	O
its	O
"	O
Linux	B-Operating_System
for	O
Tegra	O
"	O
(	O
formerly	O
"	O
L4T	O
"	O
)	O
development	O
kit	O
,	O
also	O
Nvidia	O
provides	O
JetPack	O
SDK	O
with	O
"	O
Linux	B-Operating_System
for	O
Tegra	O
"	O
and	O
other	O
tools	O
with	O
it	O
.	O
</s>
<s>
The	O
newer	O
and	O
more	O
powerful	O
devices	O
of	O
the	O
Tegra	O
family	O
are	O
now	O
supported	O
by	O
Nvidia	O
's	O
own	O
Vibrante	B-Application
Linux	B-Operating_System
distribution	O
.	O
</s>
<s>
Vibrante	B-Application
comes	O
with	O
a	O
larger	O
set	O
of	O
Linux	B-Operating_System
tools	O
plus	O
several	O
Nvidia	O
provided	O
libraries	O
for	O
acceleration	O
in	O
the	O
area	O
of	O
data	O
processing	O
and	O
especially	O
image	O
processing	O
for	O
driving	O
safety	O
and	O
automated	O
driving	O
up	O
to	O
the	O
level	O
of	O
deep	B-Algorithm
learning	I-Algorithm
and	O
neuronal	O
networks	O
that	O
make	O
e.g.	O
</s>
<s>
heavy	O
use	O
of	O
the	O
CUDA	B-Architecture
capable	O
accelerator	O
blocks	O
,	O
and	O
via	O
OpenCV	B-Language
can	O
make	O
use	O
of	O
the	O
NEON	O
vector	O
extensions	O
of	O
the	O
ARM	O
cores	O
.	O
</s>
<s>
,	O
due	O
to	O
different	O
"	O
business	O
needs	O
"	O
from	O
that	O
of	O
their	O
GeForce	B-Application
line	O
of	O
graphics	O
cards	O
,	O
Nvidia	O
and	O
one	O
of	O
their	O
Embedded	O
Partners	O
,	O
Avionic	O
Design	O
GmbH	O
from	O
Germany	O
,	O
are	O
also	O
working	O
on	O
submitting	O
open-source	B-Application
drivers	O
for	O
Tegra	O
upstream	O
to	O
the	O
mainline	O
Linux	B-Operating_System
kernel	I-Operating_System
.	O
</s>
<s>
Nvidia	O
co-founder	O
&	O
CEO	O
laid	O
out	O
the	O
Tegra	O
processor	O
roadmap	O
using	O
Ubuntu	B-Operating_System
Unity	O
in	O
GPU	B-Architecture
Technology	O
Conference	O
2013	O
.	O
</s>
<s>
By	O
end	O
of	O
2018	O
it	O
is	O
evident	O
that	O
Nvidia	O
employees	O
have	O
contributed	O
substantial	O
code	O
parts	O
to	O
make	O
the	O
T186	O
and	O
T194	O
models	O
run	O
for	O
HDMI	B-Protocol
display	O
and	O
audio	O
with	O
the	O
upcoming	O
official	O
Linux	B-Operating_System
kernel	I-Operating_System
4.21	O
in	O
about	O
Q1	O
2019	O
.	O
</s>
<s>
The	O
affected	O
software	B-Architecture
modules	O
are	O
the	O
open	O
source	O
Nouveau	O
and	O
the	O
closed	O
source	O
Nvidia	O
graphics	O
drivers	O
along	O
with	O
the	O
Nvidia	O
proprietary	O
CUDA	B-Architecture
interface	O
.	O
</s>
<s>
The	O
Drive	O
PX2	O
board	O
was	O
announced	O
with	O
QNX	B-Operating_System
RTOS	O
support	O
at	O
the	O
April	O
2016	O
GPU	B-Architecture
Technology	O
Conference	O
.	O
</s>
