L4

Ada Lovelace202372W TDPBF16FP8
NVIDIA Product Page

Specifications

CUDA Cores7,424
Tensor Cores232
AI TOPS485
FP32 TFLOPS30.3
Memory24 GB GDDR6
Memory Bandwidth300 GB/s
Memory Interface192-bit
TDP72W

Pricing

Proshop (cheapest)-
Vast.ai$0.25/hr
Google Cloud$0.81/hr

Efficiency

Power Efficiency
CUDA/Watt
103.11
Tensor/Watt
3.22
Bandwidth/Watt
4.17GB/s/W
Rental Value (CUDA/$/hr)
Vast.ai
29,696.00
GCP
9,165.43
Modal
9,280.00