Tesla D870 vs Data Center GPU Max Subsystem
We compared two Professional market GPUs: 1536MB VRAM Tesla D870 and 128GB VRAM Data Center GPU Max Subsystem to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
Tesla D870 's Advantages
Lower TDP (520W vs 2400W)
Data Center GPU Max Subsystem 's Advantages
Released 15 years and 8 months late
Boost Clock1600MHz
More VRAM (128GB vs 1536GB)
Larger VRAM bandwidth (3205GB/s vs 76.80GB/s)
16256 additional rendering cores
Score
Benchmark
FP32 (float)
Tesla D870
346
Data Center GPU Max Subsystem
+15053%
52430
Graphics Card
May 2007
Release Date
Jan 2023
Tesla
Generation
Data Center GPU
Professional
Type
Professional
PCIe 1.0 x16
Bus Interface
PCIe 5.0 x16
Clock Speeds
-
Base Clock
900 MHz
-
Boost Clock
1600 MHz
800 MHz
Memory Clock
1565 MHz
Memory
1536MB
Memory Size
128GB
GDDR3
Memory Type
HBM2e
384bit
Memory Bus
8192bit
76.80GB/s
Bandwidth
3205GB/s
Render Config
16
SM Count
-
128
Shading Units
16384
32
TMUs
1024
24
ROPs
0
-
Tensor Cores
1024
-
RT Cores
128
-
L1 Cache
64 KB (per EU)
96 KB
L2 Cache
408 MB
Theoretical Performance
14.40 GPixel/s
Pixel Rate
0 MPixel/s
38.40 GTexel/s
Texture Rate
1638 GTexel/s
-
FP16 (half)
52.43 TFLOPS
345.6 GFLOPS
FP32 (float)
52.43 TFLOPS
-
FP64 (double)
52.43 TFLOPS
Board Design
520W
TDP
2400W
900 W
Suggested PSU
2800 W
No outputs
Outputs
No outputs
-
Power Connectors
1x 16-pin
Graphics Processor
G80
GPU Name
Ponte Vecchio
Tesla
Architecture
Generation 12.5
TSMC
Foundry
Intel
90 nm
Process Size
10 nm
0.681 billion
Transistors
100 billion
484 mm²
Die Size
1280 mm²
Graphics Features
11.1 (10_0)
DirectX
12 (12_1)
3.3
OpenGL
4.6
1.1 (1.0)
OpenCL
3.0
N/A
Vulkan
N/A
1.0
CUDA
-
4.0
Shader Model
6.6
