我們接受加密貨幣付款 — 讓您持有的幣種也能買到真正的硬體設備。以加密貨幣付款,直接購入真正的硬體設備。 全球訂單滿 US$399,即享免費保險運送。簡約外箱,不印任何品牌標誌,從德國出貨。訂單滿 US$399,即享免費保險運送。 結帳一開始,您的匯率即鎖定 30 分鐘。匯率鎖定 30 分鐘。 每件商品皆原廠密封、序號經過核對,並享有 24 個月保固。原廠密封,保固 24 個月。

追蹤清單 登入

HPE

HP T4Number graphics card 16 GB

SKU CH-APTJ-HAC9T MPN T4Number

NVIDIA T4 GPU accelerator with 16 GB memory for deep learning and HPC workloads

  • Chipset ManufacturerNVIDIA
  • GPUT4
  • Memory Size16GB
  • 16 GB memory enables large-model inference
  • NVIDIA T4 GPU delivers thousands of CUDA cores for parallel compute
  • Single-slot 15" length fits dense server chassis
  • Low-profile 11.5" height suits 1U and 2U platforms
  • Light 1.28 lb weight reduces rack load

Professional accelerator for servers

The HP T4Number graphics card delivers 16 GB of memory for compute workloads. It uses the NVIDIA T4 chipset with CUDA cores to accelerate deep learning and high-performance computing tasks. The single-slot form factor fits compatible HPE ProLiant servers. It targets IT teams that need efficient inference and training performance.

Suited for inference; not for display output

This card serves data-centre workloads such as AI inference, virtual desktop infrastructure, and scientific simulation. It lacks video ports, so it cannot drive monitors directly. Buyers who need a GPU for rendering, gaming, or multi-monitor setups should look at a different product class.

16 GB memory capacity

The 16 GB frame buffer determines the maximum model size and batch count that fit on the device. Workloads requiring larger datasets need a card with more memory. Verify that your server platform supports the card's physical dimensions and power envelope before purchase.

商品特色

  • 16 GB memory enables large-model inference
  • NVIDIA T4 GPU delivers thousands of CUDA cores for parallel compute
  • Single-slot 15" length fits dense server chassis
  • Low-profile 11.5" height suits 1U and 2U platforms
  • Light 1.28 lb weight reduces rack load

商品規格

BrandHPE
Chipset ManufacturerNVIDIA
GPUT4
Memory Size16GB

商品問答

Which HPE ProLiant server models accept this 16 GB T4 card?

The card is designed for select HPE ProLiant server families; check the server quick-specs for the exact supported models and any required riser or power cable.

Will this single-slot T4 fit in a legacy PCIe 3.0 x16 slot?

Yes, it installs in a standard PCIe 3.0 x16 slot and operates at PCIe 3.0 speeds; the 16 GB memory and CUDA cores remain fully functional.

How can I confirm the 16 GB T4 is running correctly after installation?

Run nvidia-smi or the HPE iLO GPU health view; both should show the T4 with 16 GB memory, driver version, temperature and utilisation metrics.

同類商品

顧客也比較過

同一分類,同一套結帳流程 — 支援八種加密貨幣,匯率鎖定 30 分鐘。