Free U.S. shipping | Same-day shipping before 5 PM ET

R0W29C - Refurbished - HPE NVIDIA Tesla T4 16GB Computational Accelerator

$849.00 USD

Ready to ship  -  1 left

Warranty: Lifetime

Condition: Refurbished

The HPE NVIDIA Tesla T4 16GB Computational Accelerator (Part Number: R0W29C) is a versatile, energy-efficient GPU designed for AI inference, machine learning, and virtual desktop infrastructure (VDI) workloads. Built on NVIDIA's Turing architecture, it offers a balance of performance and efficiency for data center deployments.

Specifications

  • GPU Architecture: NVIDIA Turing

  • CUDA Cores: 2,560

  • Tensor Cores: 320

  • Base Clock: 585 MHz

  • Memory: 16 GB GDDR6

  • Memory Interface: 256-bit

  • Memory Bandwidth: 320 GB/s

  • Interface: PCI Express 3.0 x16

  • Form Factor: Low-profile, single-slot

  • Cooling: Passive (requires adequate system airflow)

  • Power Consumption: 70W (no external power connector required)

  • Display Outputs: None

Performance Highlights

  • FP32 (Single Precision): 8.1 TFLOPS

  • FP16 (Half Precision): 65 TFLOPS

  • INT8 Inference Performance: 130 TOPS

  • INT4 Inference Performance: 260 TOPS

Compatibility & Deployment

The Tesla T4 is compatible with systems that have a PCIe 3.0 x16 slot and sufficient airflow for passive cooling. It has been used in various server models, including:

  • HPE ProLiant DL360 Gen10

  • HPE ProLiant DL380 Gen10

  • HPE ProLiant DL385 Gen10

  • HPE ProLiant DL325 Gen10

  • HPE ML350 Gen10

  • HPE Synergy 480 Gen10

Due to its passive cooling design, it's essential to ensure that the host system provides adequate airflow to maintain optimal operating temperatures.
>
  • 5 stars: 1 (100%)
  • 4 stars: 0 (0%)
  • 3 stars: 0 (0%)
  • 2 stars: 0 (0%)
  • 1 star: 0 (0%)
P
Pradeep Krishnan
HPE Tesla T4 16GB — brilliant for low-profile GPU inference

Deployed this refurb HPE Tesla T4 16GB in a DL380 Gen10 for machine learning inference. The T4 is perfect for this use case and the low-profile passive cooled design fits our 2U chassis without any issue. NVIDIA driver and CUDA recognition were immediate under Ubuntu. INT8 inference throughput is impressive. Great refurb value at $849 versus new pricing.

Thank you Pradeep! The T4 INT8 inference performance in a 2U low-profile chassis is a great combination. Glad CUDA recognition was immediate. Enjoy the inference throughput!