Tesla P100 PCIe 12 GB and Tesla C870 - video card comparison
Tesla P100 PCIe 12 GB
Tesla C870
Architecture
Pascal (2016−2021)
Tesla (2006−2010)
GPU code name
GP100
G80
Market segment
Workstation
Workstation
Release date
June 2016
May 2007
Interface
PCIe 3.0 x16
PCIe 1.0 x16
Supplementary power connectors
1x 8-pin
2x 6-pin
Boost clock speed
1329 MHz
-
Core clock speed
1190 MHz
600 MHz
Floating-point performance
10,609 gflops
345.6 gflops
Manufacturing process technology
16 nm
90 nm
Number of transistors
15,300 million
681 million
Pipelines / CUDA cores
3584
128
Power consumption (TDP)
250 Watt
171 Watt
Texture fill rate
297.7
38.40
Memory type
HBM2
GDDR3
Maximum RAM amount
12 GB
1536 MB
Memory bus width
3072 Bit
384 Bit
Memory clock speed
1430 MHz
1600 MHz
Memory bandwidth
549.1 GB/s
76.8 GB/s
Display Connectors
No outputs
No outputs
DirectX
12 (12_1)
11.1 (10_0)
CUDA
6.0
+
OpenCL
1.2
1.1
OpenGL
4.6
3.3
Shader Model
6.4
4.0
Vulkan
1.2.131
N/A