NVIDIA announces A100 Tensor Core GPU specifications

NVIDIA Tesla A100 features 6912 CUDA Cores

The card features 7nm Ampere GA100 GPU with 6912 CUDA cores and 432 Tensor cores. This GPU has a die size of 826mm2  and 54-billion transistors.  The GPU is divided into 108 Streaming Multiprocessors. The GPU in Tesla A100 is clearly not the full chip.

The card features third-generation NVLINK with bi-directional bandwidth of 4.8 TB/s with the server and 600 GB/s in GPU to GPU interconnection.

Tesla A100 features 40GB of HBM2 memory across 5120 memory bus.

NVIDIA Tesla Series
VideoCardz.com	A100	Tesla V100s	Tesla V100	Tesla P100
Picture
GPU	7nm GA100	12nm GV100	12nm GV100	16nm GP100
Die Size	826 mm^2	815 mm^2	815 mm^2	610 mm^2
Transistors	54 billion	21.1 billion	21.1 billion	15.3 billion
SMs	108	80	80	56
CUDA Cores	6912	5120	5120	3840
Tensor Cores	432	640	640	NA
FP16 Compute	78 TFLOPS	32.8 TFLOPS	31.4 TFLOPS	21.2 TFLOPS
FP32 Compute	19.5 TFLOPS	16.4 TFLOPS	15.7 TFLOPS	10.6 TFLOPS
FP64 Compute	9.7 TFLOPS	8.2 TFLOPS	7.8 TFLOPS	5.3 TFLOPS
Boost Clock	~1410MHz	~1601 MHz	~1533 MHz	~1480MHz
Max. Memory Bandwidth	1134 GB/s	1134 GB/s	900 GB/s	721 GB/s
Eff. Memory Clock	2430 MHz	2214 MHz	1760 MHz	1408 MHz
Memory Config.	40GB HBM2e	32GB HBM2	16GB / 32GB HBM2	16GB HBM2
Memory Bus	5120-bit	4096-bit	4096-bit	4096-bit
TDP	400	250W	300W	300W
Form Factor	SXM4 / PCIe 4.0	PCIe 3.0	SXM2 / PCIe 3.0	SXM