Flopper.io
Flopper.io GPU Comparison Sheet

Baidu Kunlun P800 Kunlun III vs NVIDIA L40S 48GB

Side by side datacenter GPU specifications from the Flopper.io database.

Overview

Baidu Kunlun P800 Kunlun IIINVIDIA L40S 48GB
FP32 TFLOPS—91.6
Memory—48 GB
Memory Bandwidth—864 GB/s
TDP—350 W

Performance by Precision

PrecisionBaidu Kunlun P800 Kunlun III peak TFLOPS (dense)NVIDIA L40S 48GB peak TFLOPS (dense)
FP64 —1.4
FP32 —91.6
TF32 —183.0
BF16 —362.1
FP16 345.0362.1
FP8 —733.0
FP6 ——
FP4 ——
INT8 —733.0
INT4 —733.0

Specifications

Baidu Kunlun P800 Kunlun IIINVIDIA L40S 48GB
ArchitectureXPU-P (3rd-gen)Ada Lovelace
Process Node7nm4nm
Launch Year20242023
Form Factor—PCIe
Memory TypeGDDR6—
Interface Width——
Interconnect——
TDP—350 W
FP32 Efficiency—0.26 TFLOPS/W