<s>
Nvidia	B-Device
Tesla	I-Device
was	O
the	O
name	O
of	O
Nvidia	O
's	O
line	O
of	O
products	O
targeted	O
at	O
stream	B-Application
processing	I-Application
or	O
general-purpose	O
graphics	O
processing	O
units	O
(	O
GPGPU	B-Architecture
)	O
,	O
named	O
after	O
pioneering	O
electrical	O
engineer	O
Nikola	O
Tesla	O
.	O
</s>
<s>
They	O
are	O
programmable	O
using	O
the	O
CUDA	B-Architecture
or	O
OpenCL	B-Application
APIs	B-Application
.	O
</s>
<s>
The	O
Nvidia	B-Device
Tesla	I-Device
product	O
line	O
competed	O
with	O
AMD	O
's	O
Radeon	B-Operating_System
Instinct	I-Operating_System
and	O
Intel	B-General_Concept
Xeon	I-General_Concept
Phi	I-General_Concept
lines	O
of	O
deep	O
learning	O
and	O
GPU	O
cards	O
.	O
</s>
<s>
Its	O
new	O
GPUs	O
are	O
branded	O
Nvidia	O
Data	O
Center	O
GPUs	O
,	O
as	O
in	O
the	O
Ampere	B-General_Concept
A100	O
GPU	O
.	O
</s>
<s>
Offering	O
computational	O
power	O
much	O
greater	O
than	O
traditional	O
microprocessors	B-General_Concept
,	O
the	O
Tesla	O
products	O
targeted	O
the	O
high-performance	B-Architecture
computing	I-Architecture
market	O
.	O
</s>
<s>
,	O
Nvidia	B-Device
Teslas	I-Device
power	O
some	O
of	O
the	O
world	O
's	O
fastest	O
supercomputers	B-Architecture
,	O
including	O
Summit	B-Device
at	O
Oak	O
Ridge	O
National	O
Laboratory	O
and	O
Tianhe-1A	O
,	O
in	O
Tianjin	O
,	O
China	O
.	O
</s>
<s>
Tesla	O
cards	O
have	O
four	O
times	O
the	O
double	O
precision	O
performance	O
of	O
a	O
Fermi-based	O
Nvidia	B-Application
GeForce	I-Application
card	O
of	O
similar	O
single	O
precision	O
performance	O
.	O
</s>
<s>
Unlike	O
Nvidia	O
's	O
consumer	O
GeForce	B-Application
cards	O
and	O
professional	O
Nvidia	B-Application
Quadro	I-Application
cards	O
,	O
Tesla	O
cards	O
were	O
originally	O
unable	O
to	O
output	O
images	O
to	O
a	O
display	B-Device
.	O
</s>
<s>
This	O
will	O
be	O
a	O
64-bit	O
follow-up	O
to	O
the	O
32-bit	O
Tegra	B-Operating_System
chips	O
.	O
</s>
<s>
The	O
Tesla	B-Device
P100	I-Device
uses	O
TSMC	O
's	O
16	O
nanometer	O
FinFET	O
semiconductor	B-Architecture
manufacturing	I-Architecture
process	I-Architecture
,	O
which	O
is	O
more	O
advanced	O
than	O
the	O
28-nanometer	O
process	O
previously	O
used	O
by	O
AMD	O
and	O
Nvidia	B-Architecture
GPUs	I-Architecture
between	O
2012	O
and	O
2016	O
.	O
</s>
<s>
The	O
P100	O
also	O
uses	O
Samsung	B-Application
's	O
HBM2	O
memory	O
.	O
</s>
