GP102 has no tensor cores. NVIDIA publishes FP32 for this card and nothing else, so FP64, FP16 and INT8 are deliberately absent rather than derived. Pascal GP102 runs FP16 at a sixty-fourth of the FP32 rate as a compatibility path, and its INT8 capability is the dp4a shader instruction rather than a matrix unit; NVIDIA prints neither figure for the TITAN Xp, so neither is recorded here. NVIDIA's live product URL now redirects, so the specification source is an archived capture.

NVIDIA Titan Xp 12GB

Pascal PCIe 2017 16nm Spec confidence: Official
FP32
12
TFLOPS
VRAM
12 GB
GDDR5X
Bandwidth
548 GB/s
memory
TDP
250 W
48.6 TFLOPS/kW

Overview

TITAN Xp was the last and fastest Pascal Titan, a full GP102 with 3840 CUDA cores and 12 GB of 11.4 Gbps GDDR5X. It belongs in an AI specification database mainly as a baseline: it is the generation immediately before tensor cores, so its 12.15 TFLOPS of FP32 is all it has, with no FP16 or INT8 matrix path to fall back on. NVIDIA's own copy put it plainly, "3840 NVIDIA CUDA cores running at 1.6 GHz and packs 12 TFLOPS of brute force". The 547.7 GB/s of memory bandwidth and 12 GB of capacity kept it useful for a long time in workstations, and it still appears in second-hand listings, but a modern card of the same nominal FP32 rating will be several times faster on anything that touches half precision.

Performance

Peak theoretical throughput by precision type

PrecisionPeak
FP64
No verified data available
FP32
32-bit floating point
12TFLOPS
TF32
No verified data available
BF16
No verified data available
FP16
No verified data available
FP8
No verified data available
FP6
No verified data available
FP4
No verified data available
INT8
No verified data available

Every figure here is dense. The vendor documents no structured (2:4) sparsity mode for this part, so there is no second number to quote.

Specifications

Architecture

Pascal

Form Factor

PCIe

Launch Year

2017

Process Node

16nm

Memory

12 GB GDDR5X

Bandwidth

547.7 GB/s

TDP

250 W

Max Power

250 W

Transistors

12.0 billion

CUDA Cores

3,840

Spec Confidence

Official

Full Specifications

Compute Engine
CUDA Cores 3,840
Streaming Multiprocessors 30
Base Clock 1.58 GHz
Boost Clock 1.58 GHz
Chip Design
Transistors 12.0 billion
Die Size 471 mm²
Process Node 16nm
Memory
VRAM 12 GB
Memory Type GDDR5X
Bandwidth 548 GB/s
Interface Width 384-bit
Memory Clock 11.4 GT/s
Interconnect & I/O
PCIe 3.0 x16
Power & Thermal
TDP 250 W
Max Board Power 250 W
Power Connector One 8-pin and One 6-pin
Cooling Active
Enterprise Features
ECC Memory No
Sparsity No
Compute APIs CUDA, DirectX 12, Vulkan, OpenGL
Physical & Media
Card Length 266.7 mm
Width 2-slot
General
Form Factor PCIe
Architecture Pascal
Process Node 16nm
Launch Year 2017

Datasheet & Resources

Data Provenance

Every figure traced to a source

Primary Source

Publisher
NVIDIA
Published
2017-04-06

Data Quality

Spec confidence
Official
Clock basis
Boost
Core precisions with figures
1 of 9
Normalization
All values in TFLOPS

Rent this GPU

$0.05 / GPU-hr

Lowest of 1 live listing across 1 provider

See all rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

Similar GPUs

NVIDIA Tesla P40 24GB

PCIe · 2016
FP32: 12 TFLOPS
Compare vs Tesla P40

NVIDIA P100 SXM2 16GB

SXM · 2016
FP32: 11 TFLOPS
Compare vs P100

NVIDIA P100 PCIe 16GB

PCIe · 2016
FP32: 9.3 TFLOPS
Compare vs P100

NVIDIA Tesla P4 8GB

PCIe · 2016
FP32: 5.5 TFLOPS
Compare vs Tesla P4

Frequently Asked Questions

How many TFLOPS does the NVIDIA Titan Xp have?

The NVIDIA Titan Xp delivers 12 TFLOPS FP32 at peak. Flopper does not currently have verified FP16 and FP8 throughput figures for it.

What is the power consumption of the NVIDIA Titan Xp?

The NVIDIA Titan Xp has a TDP (Thermal Design Power) rating of 250 watts.

How much memory does the NVIDIA Titan Xp have?

The NVIDIA Titan Xp is equipped with 12 GB of VRAM with 547.7 GB/s of memory bandwidth.

What architecture is the NVIDIA Titan Xp based on?

The NVIDIA Titan Xp is based on the Pascal architecture, launched in 2017.

Stay Updated on GPU Releases

Get notified when new GPUs are added or specifications are updated.

Loading verification...

No spam, unsubscribe anytime.

Back to GPUs