GPU Comparison
GEFORCE
NVIDIA L40
CORE STATE
AD102
VRAM
48 GB
CLOCK SPEED
2490 MHz
TDP
300 W
BUS WIDTH
384 bit
ARCHITECTURE
Ada Lovelace
PROCESS
5 nm
LAUNCH DATE
2022
VS
GEFORCE
RTX A5000 Mobile
CORE STATE
GA104
VRAM
16 GB
CLOCK SPEED
1575 MHz
TDP
150 W
BUS WIDTH
256 bit
ARCHITECTURE
Ampere
PROCESS
8 nm
LAUNCH DATE
2021
PERFORMANCE BENCHMARKS
geekbench_opencl
geekbench_vulkan
passmark_directx_10
passmark_directx_11
passmark_directx_12
passmark_directx_9
passmark_g2d
passmark_g3d
passmark_gpu_compute
DETAILED SPECIFICATIONS
SPECIFICATION
L40
RTX A5000 Mobile
Core Specs
Shading Units
18,176
6,144
-66.2%
Shaders
18,176
6,144
-66.2%
TMUs
568
192
-66.2%
ROPs
192
96
-50.0%
SM Count
142
48
-66.2%
Clocks
Base Clock
735 MHz
900 MHz
Boost Clock
2490 MHz
1575 MHz
Memory Clock
2250 MHz
18 Gbps effective
1750 MHz
14 Gbps effective
Memory
Memory Size
48 GB
16 GB
VRAM (MB)
49,152
16,384
-66.7%
Memory Type
GDDR6
GDDR6
Memory Bus
384 bit
256 bit
Bandwidth
864.0 GB/s
448.0 GB/s
Cache
L1 Cache
128 KB (per SM)
128 KB (per SM)
L2 Cache
96 MB
4 MB
Performance
Pixel Rate
478.1 GPixel/s
151.2 GPixel/s
Texture Rate
1,414.3 GTexel/s
302.4 GTexel/s
FP32 (TFLOPS)
90.52 TFLOPS
19.35 TFLOPS
FP64 (TFLOPS)
1,414.3 GFLOPS (1:64)
302.4 GFLOPS (1:64)
FP16 (TFLOPS)
90.52 TFLOPS (1:1)
19.35 TFLOPS (1:1)
AI/RT
RT Cores
142
48
-66.2%
Tensor Cores
568
192
-66.2%
Power
TDP
300 W
150 W
TDP (W)
300
150
-50.0%
Suggested PSU
700 W
—
Power Connectors
1x 16-pin
None
Architecture
Architecture
Ada Lovelace
Ampere
GPU Name
AD102
GA104
Generation
Server Ada
(Lxx)
Ampere-MW
(Ax000)
Process Size
5 nm
8 nm
Transistors
76,300 million
17,400 million
Die Size
609 mm²
392 mm²
Foundry
TSMC
Samsung
Density
125.3M / mm²
44.4M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
8.9
8.6
Shader Model
6.8
6.8
Physical
Slot Width
Dual-slot
—
Length
267 mm
10.5 inches
—
Height
111 mm
4.4 inches
—
Outputs
4x DisplayPort 1.4a
Portable Device Dependent
Bus Interface
PCIe 4.0 x16
PCIe 4.0 x16
Other
Production
End-of-life
End-of-life
Predecessor
Server Ampere
Quadro Turing-M
Successor
Server Hopper
Ada-MW