NVIDIA A30
NVIDIA · Ampere
Shipping
inferenceavailablemid-range
Memory
24GB HBM2
FP8 TFLOPS
N/A
Power
165W
Released
Apr 1, 2021
Technical Specifications
| Manufacturer | NVIDIA |
|---|---|
| Architecture | Ampere |
| Memory | 24GB HBM2 |
| Memory Bandwidth | 0.933 TB/s |
| FP16 TFLOPS | 165 |
| FP8 TFLOPS | N/A |
| Interconnect | NVLink3 (600GB/s) |
| Power (TDP) | 165W |
| Released | Apr 1, 2021 |
| Status | Shipping |
Availability
Available — mainstream inference GPU, lower cost than A100
Pricing
Retail (estimated)
3000-5000
Cloud Pricing ($/hr)
Key Specs
Memory: 24GB HBM2
Bandwidth: 0.933 TB/s
FP8: N/A TFLOPS
Power: 165W