Tesla D870 vs GRID A100A
We compared two Professional market GPUs: 1536MB VRAM Tesla D870 and 48GB VRAM GRID A100A to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
GRID A100A 's Advantages
Released 13 years late
Boost Clock1005MHz
More VRAM (48GB vs 1536GB)
Larger VRAM bandwidth (1866GB/s vs 76.80GB/s)
6784 additional rendering cores
Lower TDP (400W vs 520W)
Score
Benchmark
FP32 (float)
Tesla D870
346
GRID A100A
+3914%
13890
Graphics Card
May 2007
Release Date
May 2020
Tesla
Generation
GRID
Professional
Type
Professional
PCIe 1.0 x16
Bus Interface
PCIe 4.0 x16
Clock Speeds
-
Base Clock
900 MHz
-
Boost Clock
1005 MHz
800 MHz
Memory Clock
1215 MHz
Memory
1536MB
Memory Size
48GB
GDDR3
Memory Type
HBM2e
384bit
Memory Bus
6144bit
76.80GB/s
Bandwidth
1866GB/s
Render Config
16
SM Count
108
128
Shading Units
6912
32
TMUs
432
24
ROPs
192
-
Tensor Cores
432
-
L1 Cache
192 KB (per SM)
96 KB
L2 Cache
48 MB
Theoretical Performance
14.40 GPixel/s
Pixel Rate
193.0 GPixel/s
38.40 GTexel/s
Texture Rate
434.2 GTexel/s
-
FP16 (half)
55.57 TFLOPS
345.6 GFLOPS
FP32 (float)
13.89 TFLOPS
-
FP64 (double)
6.947 TFLOPS
Board Design
520W
TDP
400W
900 W
Suggested PSU
800 W
No outputs
Outputs
No outputs
-
Power Connectors
None
Graphics Processor
G80
GPU Name
GA100
Tesla
Architecture
Ampere
TSMC
Foundry
TSMC
90 nm
Process Size
7 nm
0.681 billion
Transistors
54.2 billion
484 mm²
Die Size
826 mm²
Graphics Features
11.1 (10_0)
DirectX
N/A
3.3
OpenGL
N/A
1.1 (1.0)
OpenCL
3.0
N/A
Vulkan
N/A
1.0
CUDA
8.0
4.0
Shader Model
N/A
