Dell Nvidia Tesla T4 Turing 16GB GDDR6 PCI-Express 3.0 x16 256-Bit Graphics Card - 0HCKJM

Dell NVIDIA Tesla T4 Turing 16GB GDDR6 PCIe 3.0 x16 256-Bit AI Inference GPU

The Dell NVIDIA Tesla T4 (0HCKJM) is a 16GB GDDR6 PCIe 3.0 x16 enterprise AI inference GPU based on NVIDIA's Turing architecture. Designed for data center AI inference, machine learning, and virtual GPU (vGPU) workloads, the T4 delivers Tensor Core-accelerated INT8 and FP16 inference performance in a low-profile, 70W passive-cooled form factor — purpose-built for dense server deployments.

Key Features

• 16GB GDDR6 — large GPU memory for AI inference and vGPU workloads
• NVIDIA Turing Tensor Cores for INT8/FP16 AI inference acceleration
• 70W TDP — passive cooling for dense server rack deployments
• PCIe 3.0 x16 — standard server GPU connectivity
• Multi-instance GPU (MIG-like) via vGPU for virtualized inference
• Dell OEM — validated for Dell PowerEdge server platforms

Product Specifications

• Part Number: 0HCKJM
• Manufacturer: Dell / NVIDIA
• OEM Status: OEM (Dell)
• GPU: NVIDIA Tesla T4 (Turing TU104)
• Memory: 16GB GDDR6
• Memory Bus: 256-bit
• Interface: PCIe 3.0 x16
• TDP: 70W

Component Specifications

• GPU Architecture: NVIDIA Turing (TU104)
• CUDA Cores: 2,560
• Tensor Cores: 320 (2nd Gen)
• Memory: 16GB GDDR6
• Memory Bandwidth: 300 GB/s
• Memory Bus: 256-bit
• Interface: PCIe 3.0 x16
• Form Factor: Low-profile (LP)

Performance Specifications

• FP32: 8.1 TFLOPS
• FP16: 65 TOPS (Tensor Core)
• INT8: 130 TOPS (Tensor Core)
• TDP: 70W (passive cooled)
• vGPU Support: Yes (NVIDIA vGPU software)

Use Cases

• AI inference acceleration for deep learning models in data center servers
• Virtual GPU (vGPU) for virtualized AI and GPU-accelerated desktop deployments
• Video analytics and computer vision inference at the edge and in the data center
• Dense multi-GPU server configurations leveraging 70W passive cooling

Compatibility

• Compatible with Dell PowerEdge servers with PCIe 3.0 x16 slots supporting low-profile GPUs
• Requires NVIDIA T4 drivers and CUDA 10.0+
• vGPU requires NVIDIA vGPU software license
• Passive cooling — requires adequate server airflow for thermal management

Frequently Asked Questions

What is the NVIDIA Tesla T4 used for?

The NVIDIA Tesla T4 is an AI inference GPU optimized for Tensor Core-accelerated INT8 and FP16 inference — ideal for deploying trained deep learning models at scale in data center server environments.

Does the T4 require external power connectors?

No — the NVIDIA Tesla T4 is powered entirely via the PCIe slot (70W TDP), requiring no external power connectors, simplifying server integration.

Does the T4 support vGPU?

Yes — the T4 supports NVIDIA vGPU software for virtualizing GPU resources across multiple VMs, enabling shared GPU access for AI inference and GPU-accelerated virtual desktops.

The Dell NVIDIA Tesla T4 (0HCKJM) delivers 16GB GDDR6 AI inference performance with Tensor Core acceleration in a 70W passive-cooled PCIe 3.0 form factor — purpose-built for Dell PowerEdge data center AI inference and vGPU deployments. Order today and accelerate your AI inference workloads with proven NVIDIA T4 performance.