NVIDIA L40 CNX vs Intel Arc B570

GPU · Specs Compare

Products
GPU NVIDIA L40 CNX NVIDIA · 2022-10
GPU Intel Arc B570 Intel · 2025-01
Graphics Card
Release Date
Oct 2022
Jan 2025
Generation
Tesla Ada
Battlemage
Type
Professional
Desktop
Bus Interface
PCIe 4.0 x16
PCIe 4.0 x8
Clock Speeds
Base Clock
1005 MHz
2500 MHz
Boost Clock
2475 MHz
2500 MHz
Memory Clock
2250 MHz
2375 MHz
Memory
Memory Size
24GB
10GB
Memory Type
GDDR6
GDDR6
Memory Bus
384bit
160bit
Bandwidth
864.0GB/s
380.0GB/s
Render Config
SM Count
142
—
Shading Units
18176
2304
TMUs
568
144
ROPs
192
80
Tensor Cores
568
144
RT Cores
142
18
L1 Cache
128 KB (per SM)
256 KB (per EU)
L2 Cache
48 MB
18 MB
Theoretical Performance
Pixel Rate
475.2 GPixel/s
200.0 GPixel/s
Texture Rate
1406 GTexel/s
360.0 GTexel/s
FP16 (half)
89.97 TFLOPS
23.04 TFLOPS
FP32 (float)
89.97 TFLOPS
11.52 TFLOPS
FP64 (double)
1406 GFLOPS
1440 GFLOPS
Graphics Processor
GPU Name
AD102
BMG-G21
GPU Variant
AD102-???-A1
SRPMG
Architecture
Ada Lovelace
Xe2
Foundry
TSMC
TSMC
Process Size
4 nm
5 nm
Transistors
76.3 billion
19.6 billion
Die Size
608 mm²
272 mm²
Board Design
TDP
300W
150W
Suggested PSU
700 W
450 W
Outputs
1x HDMI 2.13x DisplayPort 1.4a
1x HDMI 2.1a3x DisplayPort 2.1
Power Connectors
1x 16-pin
1x 8-pin
Graphics Features
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
OpenCL
3.0
3.0
Vulkan
1.3
1.4
CUDA
8.9
—
Shader Model
6.6
6.6
Benchmark
3DMark Time Spy
—
12,600
3DMark Time Spy Extreme
—
6,004
▲ Specs Compare Data sourced from official specifications and benchmark tests.