Dell PowerEdge R760xa GPU Server | Dual Xeon Platinum 8568Y+ | 1x NVIDIA H100 NVL 94GB | 256GB DDR5
  • 商品カテゴリー:サーバー
  • 品番:Dell PowerEdge R760xa
  • 在庫状況:In Stock
  • 状態:新品
  • 製品機能:即発送可能
  • 最小注文:1 単位
  • 定価:$67,899.00
  • 販売価格: $58,441.00 節約額 $9,458.00
  • 今すぐチャット メールを送信

ご安心ください。返品を承ります。

配送:国際発送の場合、関税手続きや追加料金が発生する場合があります。 詳細を見る

配達:国際配送が税関手続きの対象となる場合は、追加のお時間をいただく場合があります。 詳細を見る

返品:14日以内の返品が可能です。返品送料は販売者が負担します。 詳細を見る

送料無料。当社はNET 30 Daysの購入注文を受け付けています。信用情報に影響を与えることなく、数秒で決定を得られます。

大量の Dell PowerEdge R760xa 製品が必要な場合は、フリーダイヤル Whatsapp: (+86) 151-0113-5020 までお電話いただくか、ライブチャットでお見積もりをご依頼ください。担当営業マネージャーがすぐにご連絡いたします。

Dell PowerEdge R760xa GPU Server | Dual Xeon Platinum 8568Y+ | 1x NVIDIA H100 NVL 94GB | 256GB DDR5

Keywords

Dell PowerEdge R760xa, NVIDIA H100 NVL, Intel Xeon Platinum 8568Y+, AI Inference Server, GPU Accelerated Server, 256GB DDR5 RAM, H100 94GB, Buy Dell GPU Server

Description

Dominate the landscape of Large Language Models (LLMs) and generative AI with the purpose-built Dell PowerEdge R760xa. The "xa" designation stands for eXtreme Acceleration, indicating a chassis optimized specifically for thermal management and power delivery to high-performance GPUs. This configuration is driven by dual Intel Xeon Platinum 8568Y+ processors. As 5th Generation Intel Scalable CPUs (Emerald Rapids), each processor boasts 48 cores and massive cache, providing a combined 96 cores of orchestration power to manage heavy data preprocessing pipelines before they reach the accelerator.

The crown jewel of this system is the NVIDIA H100 NVL 94GB GPU. Designed specifically for LLM deployment, the H100 NVL offers a massive 94GB of HBM3e memory, significantly higher than the standard 80GB PCIe version. This extra VRAM allows for larger models (like Llama-3 70B or Falcon 180B layers) to be loaded into memory with higher batch sizes, drastically improving inference throughput and reducing latency. Supported by 256GB DDR5 RAM, the system ensures that the CPU and GPU are never starved of data.

Storage reliability is guaranteed by the enterprise-grade PERC H755 controller managing four 1.92TB SAS SSDs. This setup provides nearly 8TB of fast, redundant storage for models, datasets, and logs. The server is housed in an 8-bay 2.5" chassis, leaving room for future expansion. Connectivity is provided by a quad-port 1GbE OCP adapter for management and standard traffic. To handle the immense power draw of the H100 NVL and Platinum CPUs, the system is equipped with dual 2800W Titanium Power Supplies, offering the highest efficiency rating available and 1+1 redundancy.

Management is streamlined via iDRAC9 Enterprise, offering advanced telemetry for thermal and power usage, essential for GPU Accelerated Server deployments. Whether used for fine-tuning foundation models or serving high-traffic AI APIs, the R760xa provides the thermal headroom and power density required for sustained peak performance. When you Buy Dell GPU Server hardware of this caliber, you are investing in an infrastructure capable of handling the next generation of AI parameters.

Key Features

  • Extreme Acceleration: 1x NVIDIA H100 NVL GPU with 94GB HBM3e memory for superior AI inference.
  • Elite Processing: 2x Intel Xeon Platinum 8568Y+ (48 Cores/96 Threads each, 5th Gen) for maximum compute.
  • High-Bandwidth Memory: 256GB DDR5 memory ensures rapid data transfer to the GPU.
  • Robust Storage: 4x 1.92TB SAS SSDs managed by PERC H755 for reliable, high-speed access.
  • Power Density: Dual 2800W Titanium PSUs (1+1 Redundant) designed for high-TDP workloads.
  • Thermal Optimization: R760xa chassis features specialized airflow zoning for GPU cooling.
  • Management: iDRAC9 Enterprise includes advanced power and thermal controls.
  • Deployment Ready: Includes ReadyRails for easy rack installation.

Configuration

Component Specification
Model Dell PowerEdge R760xa (GPU Optimized Chassis)
GPU Accelerator 1x NVIDIA H100 NVL 94GB HBM3e
Processor (CPU) 2x Intel Xeon Platinum 8568Y+ (48 Cores, 2.3GHz, 5th Gen)
Memory (RAM) 256GB DDR5 Total
Storage 4x 1.92TB SAS SSD (2.5")
Storage Controller PERC H755 (SAS/SATA RAID Controller)
Power Supply 2x 2800W Titanium Redundant (1+1) PSU
Network 4x 1GbE OCP 3.0 Adapter
Management iDRAC9 Enterprise
Rails ReadyRails Sliding Rails

Compatibility

The Dell PowerEdge R760xa is specifically engineered to support high-wattage GPUs and 5th Generation Intel Xeon processors. It is fully certified for the NVIDIA AI Enterprise software suite, ensuring drivers and CUDA toolkits work seamlessly with the H100 NVL. Supported operating systems include Ubuntu Server 22.04 LTS, Red Hat Enterprise Linux 9, and Windows Server 2022. The PERC H755 controller supports SAS and SATA drives, and the OCP 3.0 slot allows for future network upgrades to 10GbE, 25GbE, or 100GbE if needed.

Usage Scenarios

This configuration is the ultimate Large Language Model (LLM) Inference Node. The 94GB of VRAM on the H100 94GB card is a critical differentiator, allowing larger models (such as 70B parameter models) to fit comfortably in memory without needing to split across multiple GPUs or resort to heavy quantization. This results in faster tokens-per-second generation.

It is also ideal for Generative AI Fine-Tuning. While full training of massive models requires clusters, fine-tuning existing models (like Llama 3 or Mistral) on proprietary company data can be efficiently handled by this single powerful node, leveraging the 96 CPU cores for data preprocessing.

In the field of High-Performance Simulation, such as Computational Fluid Dynamics (CFD) or molecular modeling, the combination of high-clock-speed Platinum CPUs and the double-precision capability of the H100 allows researchers to run complex physics simulations significantly faster than on standard CPU servers.

Frequently Asked Questions

Q: What is the difference between H100 NVL and H100 PCIe?
A: The standard H100 PCIe typically comes with 80GB of memory. The NVIDIA H100 NVL is a specialized version often deployed in pairs with an NVLink bridge, featuring 94GB of HBM3e memory per GPU. This extra memory capacity and bandwidth make the NVL version superior for memory-bound AI workloads like LLM inference.

Q: Why does this server need 2800W power supplies?
A: The R760xa is a high-density GPU server. The H100 GPU can draw up to 350-400W, and the two Platinum 8568Y+ processors can draw 350W each. Combined with fans running at high speed to cool these components, the power draw is substantial. The 2800W Titanium PSUs ensure the system operates efficiently and has enough overhead to prevent shutdowns during power spikes.

Q: Can I add more GPUs to this chassis later?
A: Yes, the R760xa chassis supports up to 4x double-width GPUs. With only one installed in this configuration, you have significant room for expansion. However, you must ensure you have the correct power cables and risers available when adding future cards.

Q: Is the R760xa different from the standard R760?
A: Yes. The "xa" model is deeper and has a different internal layout optimized for airflow to cool multiple high-performance GPUs. It supports more GPUs and higher-wattage configurations than the standard R760.

この商品に関連する製品
Inspur NF5280M6 AI 対応デュアル Xeon サーバー (エンタープライズ ワークロード向け Tesla L2 GPU 搭載) おすすめ
Inspur NF5466M6 デュアル Intel Xeon 4314 エンタープライズ ストレージおよびコンピューティング サーバー おすすめ
Dell PowerEdge R760xs - デュアル Xeon Silver 4410Y エンタープライズ構成 おすすめ
Dell PowerEdge R760xs - Xeon Gold 6507P パフォーマンス サーバー おすすめ
Dell PowerEdge R660 1U ラック サーバー - デュアル Xeon Gold 6430、1TB RAM、25GbE、およびファイバー チャネル HBA おすすめ
HPE ProLiant DL380 Gen11 2U ラック サーバー |デュアル インテル Xeon ゴールド 6542Y 48 コア | 1TB DDR5 RAM | 2x 300GB SAS HDD おすすめ
Inspur NF8480M5 4U エンタープライズ ストレージ サーバー — 24× LFF SAS、クアッド Xeon Gold 6248R 高密度プラットフォーム おすすめ
Lenovo ThinkSystem SR850 V3 高性能 4 CPU サーバー (Xeon Gold 6448H およびエンタープライズ ネットワーキング搭載) おすすめ