NVIDIA Tesla P40 24GB vs NVIDIA Tesla P4 8GB
The Tesla P40 delivers 2.2x the FP32 throughput of the Tesla P4 (12 vs 5.5 TFLOPS dense). The Tesla P40 also carries 16 GB more memory (24 GB vs 8 GB).
NVIDIA Tesla P40 24GB
Full specs →NVIDIA Tesla P4 8GB
Full specs →Specifications
Performance (TFLOPS)
FLOPS by Precision
What actually differs
Both GPUs launched in 2016: the Tesla P40 on Pascal and the Tesla P4 on Pascal.
Dense throughput for the Tesla P40 against the Tesla P4: FP32 12 vs 5.5 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.
Memory is 24 GB against 8 GB, fed at 346 GB/s versus 192 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.
Power budgets are 250 W for the Tesla P40 and 75 W for the Tesla P4. At FP32 that works out to 0.05 against 0.07 TFLOPS per watt.
For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.
Frequently asked questions
Is the Tesla P40 faster than the Tesla P4?
At FP32 precision the Tesla P40 reaches 12 TFLOPS dense against 5.5 TFLOPS for the Tesla P4. The performance table on this page lists every published precision for both GPUs.
Which has more memory, the Tesla P40 or the Tesla P4?
The Tesla P40 carries 24 GB of VRAM versus 8 GB for the Tesla P4. Memory bandwidth is 346 GB/s for the Tesla P40 and 192 GB/s for the Tesla P4.
How much power do the Tesla P40 and the Tesla P4 draw?
The Tesla P40 is rated at 250 W TDP and the Tesla P4 at 75 W. On FP32 throughput per watt, the Tesla P4 is the more efficient part.
Get Comparison Updates
New GPUs added weekly. Be the first to see how they compare.