NVIDIA L4 vs NVIDIA GRID RTX T10 4
We compared two Professional market GPUs: 24GB VRAM L4 and 4GB VRAM GRID RTX T10 4 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA L4 's Advantages
Released 3 years and 2 months late
Boost Clock has increased by 46% (2040MHz vs 1395MHz)
More VRAM (24GB vs 4GB)
3840 additional rendering cores
Lower TDP (72W vs 150W)
NVIDIA GRID RTX T10 4 's Advantages
Larger VRAM bandwidth (672.0GB/s vs 300.1GB/s)
Score
Benchmark
FP32 (float)
L4
+202%
30.29 TFLOPS
GRID RTX T10 4
9.999 TFLOPS
Graphics Card
Mar 2023
Release Date
Jan 2020
Tesla Ada(Lxx)
Generation
GRID
Professional
Type
Professional
PCIe 4.0 x16
Bus Interface
PCIe 3.0 x16
Clock Speeds
795 MHz
Base Clock
1065 MHz
2040 MHz
Boost Clock
1395 MHz
1563 MHz
Memory Clock
1750 MHz
Memory
24GB
Memory Size
4GB
GDDR6
Memory Type
GDDR6
192bit
Memory Bus
384bit
300.1GB/s
Bandwidth
672.0GB/s
Render Config
-
-
-
60
SM Count
56
7424
Shading Units
3584
240
TMUs
224
80
ROPs
64
240
Tensor Cores
448
60
RT Cores
56
128 KB (per SM)
L1 Cache
64 KB (per SM)
48 MB
L2 Cache
6 MB
-
-
-
Theoretical Performance
163.2 GPixel/s
Pixel Rate
89.28 GPixel/s
489.6 GTexel/s
Texture Rate
312.5 GTexel/s
30.29 TFLOPS
FP16 (half)
20.00 TFLOPS
30.29 TFLOPS
FP32 (float)
9.999 TFLOPS
473.3 GFLOPS
FP64 (double)
312.5 GFLOPS
Board Design
72W
TDP
150W
250 W
Suggested PSU
450 W
No outputs
Outputs
No outputs
None
Power Connectors
1x 6-pin + 1x 8-pin
Graphics Processor
AD104
GPU Name
TU102
AD104-???-A1
GPU Variant
TU102-875-A1
Ada Lovelace
Architecture
Turing
TSMC
Foundry
TSMC
5 nm
Process Size
12 nm
35.8 billion
Transistors
18.6 billion
294 mm²
Die Size
754 mm²
Graphics Features
12 Ultimate (12_2)
DirectX
12 Ultimate (12_2)
4.6
OpenGL
4.6
3.0
OpenCL
3.0
1.3
Vulkan
1.3
8.9
CUDA
7.5
6.8
Shader Model
6.8
