NVIDIA GeForce RTX 5070 Ti SUPER vs NVIDIA DGX Spark (MediaTek AHJ11488B)
GPU Specs Compare
Graphics Card
Release Date
Jan 2026
-
Generation
GeForce 50
-
Type
Desktop
-
Bus Interface
PCIe 5.0 x16
-
Clock Speeds
Base Clock
2295 MHz
-
Boost Clock
2452 MHz
-
Memory Clock
1750 MHz
-
Memory
Memory Size
16GB
-
Memory Type
GDDR7
-
Memory Bus
256bit
-
Bandwidth
896.0GB/s
-
Render Config
SM Count
70
-
Shading Units
8960
-
TMUs
280
-
ROPs
96
-
Tensor Cores
280
-
RT Cores
70
-
L1 Cache
128 KB (per SM)
-
L2 Cache
48 MB
-
Theoretical Performance
Pixel Rate
235.4 GPixel/s
-
Texture Rate
686.6 GTexel/s
-
FP16 (half)
43.94 TFLOPS
-
FP32 (float)
43.94 TFLOPS
-
FP64 (double)
686.6 GFLOPS
-
Graphics Processor
GPU Name
GB203
-
GPU Variant
GB203-350-A1
-
Architecture
Blackwell 2.0
-
Foundry
TSMC
-
Process Size
5 nm
-
Transistors
45.6 billion
-
Die Size
378 mm²
-
Board Design
TDP
350W
-
Suggested PSU
750 W
-
Outputs
1x HDMI 2.1b3x DisplayPort 2.1b
-
Power Connectors
1x 16-pin
-
Graphics Features
DirectX
12 Ultimate (12_2)
-
OpenGL
4.6
-
OpenCL
3.0
-
Vulkan
1.4
-
CUDA
12.0
-
Shader Model
6.8
-
General Parameters
Release Date
-
Oct 2025
Type
-
Laptop
Core Architecture
-
N1X
Integrated Graphics
-
GB20B "Blackwell"
Generation
-
RTX Spark (GB10 (N1X))
Package
Manufacturing Process
-
3 nm
Power Consumption
-
140 W
Foundry
-
TSMC
Die Size
-
208 mm²
Package
-
FC-BGA
CPU Performance
Performance Core Base Frequency
-
2.8 GHz
Performance Core Turbo Frequency
-
4 GHz
Efficiency Core Base Frequency
-
1.5 GHz
Efficiency Core Turbo Frequency
-
2.8 GHz
Total Core Count
-
20
Total Thread Count
-
20
Bus Frequency
-
100 MHz
Multiplier
-
28.0
L1 Cache
-
128 KB per core
L2 Cache
-
2 MB per core
L3 Cache
-
8 MB shared
Unlocked Multiplier
-
No
SMP
-
1
Memory Parameters
Memory Types
-
LPDDR5X-8533
Max Memory Channels
-
4
Max Memory Bandwidth
-
273.1 GB/s
ECC Memory Support
-
No
Miscellaneous
PCIe Version
-
0