TITAN RTX is binned like a workstation card, not like a GeForce one. The Ampere GA102 whitepaper prints 130.5 TFLOPS for BOTH of its FP16 tensor accumulate modes, exactly as the Quadro RTX 6000 does, so it escapes the halving that GeForce Turing suffers on FP16 with FP32 accumulate. Its FP16 row is therefore 8x FP32 while every GeForce Turing row here is 4x, and that is correct rather than a copy defect. NVIDIA's live product URL now redirects to an RTX 6000 Ada page, so the specification source is an archived capture.

NVIDIA Titan RTX 24GB

Turing PCIe 2018 12nm Spec confidence: Official
FP32
16
TFLOPS
VRAM
24 GB
GDDR6
Bandwidth
672 GB/s
memory
TDP
280 W
58.2 TFLOPS/kW

Overview

TITAN RTX is the full 72-SM TU102 die with 24 GB of GDDR6, and it was the last Titan before NVIDIA folded the tier into the GeForce RTX 3090. For AI it matters because of two things the GeForce cards of its generation did not have: the capacity, 24 GB against the RTX 2080 Ti's 11 GB, and the workstation bin, which runs FP16 tensor at 130.5 TFLOPS whether you accumulate in FP16 or FP32 rather than halving the FP32-accumulate mode. Two cards bridged over NVLink present 48 GB at 100 GB/s. Turing tensor cores cover FP16, INT8 and INT4 and stop there, so there is no BF16, TF32 or FP8 here, and no structured sparsity either, which arrived with Ampere. That makes the 130.5 TFLOPS figure a dense number directly comparable with everything else on this site.

Performance

Peak theoretical throughput by precision type

PrecisionPeak
FP64
64-bit floating point
0.5TFLOPS
FP32
32-bit floating point
16TFLOPS
TF32
No verified data available
BF16
No verified data available
FP16
16-bit floating point
131TFLOPS
FP8
No verified data available
FP6
No verified data available
FP4
No verified data available
INT8
8-bit integer
261TOPS
INT4
4-bit integer
522TOPS

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.

The Titan RTX 24GB in the GPU landscape

Peak FP16 TFLOPS (dense) against TDP, single-GPU parts tracked by Flopper

06001,2001,8002,4003,0000 W250 W500 W750 W1000 W1250 W1500 WInstinct MI355XTitan RTX 24GB

Higher and further left is better: more half-precision throughput for less power.

Specifications

Architecture

Turing

Form Factor

PCIe

Launch Year

2018

Process Node

12nm

Memory

24 GB GDDR6

Bandwidth

672 GB/s

TDP

280 W

Max Power

280 W

Transistors

18.6 billion

Interconnect

100 GB/s NVLink

CUDA Cores

4,608

Spec Confidence

Official

Full Specifications

Compute Engine
CUDA Cores 4,608
Tensor Cores 576 (Turing)
Streaming Multiprocessors 72
Base Clock 1.35 GHz
Boost Clock 1.77 GHz
Chip Design
Transistors 18.6 billion
Die Size 754 mm²
Process Node 12nm
Memory
VRAM 24 GB
Memory Type GDDR6
Bandwidth 672 GB/s
Interface Width 384-bit
Memory Clock 14 GT/s
Interconnect & I/O
GPU-to-GPU NVLink
Interconnect Bandwidth 100 GB/s
Interconnect Links 2
PCIe 3.0 x16
Cache
L2 Cache 6 MB
Power & Thermal
TDP 280 W
Max Board Power 280 W
Power Connector Two 8-pin
Cooling Active
Enterprise Features
ECC Memory No
Sparsity No
Compute APIs CUDA, DirectX 12, Vulkan, OpenGL
Physical & Media
Width 2-slot
General
Form Factor PCIe
Architecture Turing
Process Node 12nm
Launch Year 2018

Datasheet & Resources

Data Provenance

Every figure traced to a source

Primary Source

Publisher
NVIDIA
Published
2018-12-03

Data Quality

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

Rent this GPU

$0.14 / 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

Frequently Asked Questions

How many TFLOPS does the NVIDIA Titan RTX have?

The NVIDIA Titan RTX delivers 16 TFLOPS FP32 and 131 TFLOPS FP16 at peak. Flopper does not currently have a verified FP8 throughput figure for it.

What is the power consumption of the NVIDIA Titan RTX?

The NVIDIA Titan RTX has a TDP (Thermal Design Power) rating of 280 watts.

How much memory does the NVIDIA Titan RTX have?

The NVIDIA Titan RTX is equipped with 24 GB of VRAM with 672 GB/s of memory bandwidth.

What architecture is the NVIDIA Titan RTX based on?

The NVIDIA Titan RTX is based on the Turing architecture, launched in 2018.

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