AMD Radeon Pro Duo Polaris vs NVIDIA GeForce RTX 3070 Ti Full Specs

Summary

4,608 Shaders 1.24GHz	6,144 Shaders 1.77GHz
32GB GDDR5448GB/s	8GB GDDR6X608.3GB/s
Peak·FP32·FP16 11.46 TFLOPS	Peak·FP32·FP16 21.75 TFLOPS
Form Factor PCIe Card	Form Factor PCIe Card
TDP 250W	TDP 290W
Power Connectors - 1x 8-Pin 1x 6-Pin	Power Connectors 1x 12-Pin - -

Highlights

Peak OPS

Peak OPS FP32 FP16 Tensor FP16-16 FP16-16 Tensor Sparse Tensor FP16-32 FP16-32 Tensor Sparse FP64 BF16 Tensor BF16 BF16 Tensor Sparse Tensor TF32 Tensor INT8 Tensor INT4 Memory Bandwidth TDP

Radeon Pro Duo Polaris11.46 TFLOPSFP16 (2x 5.73 TFLOPS)

x1

GeForce RTX 3070 Ti696 TOPSINT4 Tensor Sparse

x60.76

Show Top Score

Theoretical Performance

Clock Speed1243MHz UC·Boost·OC	Clock Speed1770MHz UC·Base·Boost·OC
Peak OPS 11.46 TFLOPSFP16	Peak OPS 696 TOPSINT4 Tensor Sparse
-	Tensor FP16-16 87 TFLOPS FP16-16 Tensor Sparse 174 TFLOPS Tensor FP16-32 43.5 TFLOPS FP16-32 Tensor Sparse 87 TFLOPS
-	BF16 21.75 TFLOPS Tensor BF16 43.5 TFLOPS BF16 Tensor Sparse 87 TFLOPS
Tensor TF32 -	Tensor TF32 21.75 TFLOPS
FP32 11.46 TFLOPS	FP32 21.75 TFLOPS
FP64 716 GFLOPS	FP64 339.8 GFLOPS
Tensor INT4 -	Tensor INT4 348 TOPS
-	Tensor INT8 174 TOPS
Ray -	Ray 32.8 TOPS
Pixel Rate 39.8 GPixel/s	Pixel Rate 169.9 GPixel/s
Texture Rate 159.1 GTexel/s	Texture Rate 339.8 GTexel/s

Core Config

Shaders 4,608 Shaders	Shaders 6,144 Shaders
TMUs 256 TMUs	TMUs 192 TMUs
ROPs 64 ROPs	ROPs 96 ROPs
Tensor Cores -	Tensor Cores 192 T-Cores
RT Cores -	RT Cores 48 RT-Cores
CUs 72 CUs	SMs 48 SMs

Clocks

Base Clock -	Base Clock 1.57GHz
Boost Clock 1.24GHz	Boost Clock 1.77GHz

Cache

L2 Cache 2MB shared

L2 Cache 4.1MB shared

Memory

32GB GDDR5	8GB GDDR6X
Memory Bus 256-bit	Memory Bus 256-bit
Memory Speed 7GT/s	Memory Speed 19GT/s
Memory Bandwidth 448GB/s	Memory Bandwidth 608.3GB/s

Power & Thermals

TDP 250W	TDP 290W
Max Temp -	Max Temp 93°C Max

Video Output

Multi-Monitor 3	Multi-Monitor 4
HDCP HDCP 2.2	HDCP HDCP 2.3

Ports

3x DisplayPort 1.4	3x DisplayPort 1.4
-	1x HDMI 2.1
1x HDMI 2.0	-

Encoder

Encoder Model VCE 3.0

Encoder Model NVENC 7

Decoder

Decoder Model UVD 6.3

Decoder Model NVDEC 5

Card

Form Factor PCIe Card	Form Factor PCIe Card
PCIe 2-Slots	PCIe 2.1-Slots
Height 111 mm (4.37") Width 305 mm (12.01") Depth 37 mm (1.46")	Height 112 mm (4.41") Width 267 mm (10.51") Depth 43 mm (1.69")
Cooling Blower 1x Fan	Cooling Open-Air 2x Fans
Power Connectors 1x 8-Pin, 1x 6-Pin	Power Connectors 1x 12-Pin

Chip

Manufacturer AMD	Manufacturer NVIDIA
Chip Designer AMD	Chip Designer NVIDIA
Architecture GCN 4	Architecture Ampere
Family Radeon Pro	Family GeForce 30
Branding	Branding
Codename Ellesmere	Codename NV174
Chip Variant Polaris 10 XL	Chip Variant GA104-150-A1
Market Segment Workstation	Market Segment Desktop
Release Date Apr 24, 2017	Release Date Jun 10, 2021

Fabrication

Foundry GlobalFoundries	Foundry Samsung
Fabrication Node 14LPP	Fabrication Node 8N
Die Size 232mm²	Die Size 393mm²
Transistor Count 5.7 Billion	Transistor Count 17.4 Billion
Transistor Density 24.57 MTr/mm²	Transistor Density 44.33 MTr/mm²

Images

No images available

No images available

Used In

Desktops -

Desktops XPS(XD8950, Late 2021)Aurora(R13, Late 2021)Aurora(R14, Late 2021) Show 9 more

Competitors

AMDRadeon Pro Duo Polaris vs GPU AMDRadeon PRO W7700 AMDRadeon Pro Duo Polaris vs GPU AMDRadeon PRO W7600 AMDRadeon Pro Duo Polaris vs GPU MSIGeForce RTX 3050 LP 6G OC Show 12 more

NVIDIAGeForce RTX 3070 Ti vs GPU AMDRadeon RX 7900 GRE NVIDIAGeForce RTX 3070 Ti vs GPU AMDRadeon RX 9070 XT NVIDIAGeForce RTX 3070 Ti vs GPU AMDRadeon RX 7800 XT Show 12 more