Skip to content
H100
H100
H100
H100
H100
H100
NVIDIA H100 / H800 Data Center GPUs – CNX, PCIe 80GB / 96GB, SXM5 64GB / 80GB / 96GB, H800 PCIe 80GB, H800 SXM5 – Professional Graphics Cards - Image 7

NVIDIA H100 / H800 Data Center GPUs – CNX, PCIe 80GB / 96GB, SXM5 64GB / 80GB / 96GB, H800 PCIe 80GB, H800 SXM5 – Professional Graphics Cards

  1. H100 CNX:
    • A custom form factor GPU designed for specific data center deployments.
    • Features 80GB HBM3 memory and a ~700W TDP.
    • Ideal for custom HPC and AI workloads.
  2. H100 PCIe 80 GB:
    • A PCIe form factor GPU with 80GB HBM3 memory.
    • Operates at a ~350W TDP, making it suitable for standard server deployments.
  3. H100 PCIe 96 GB:
    • A PCIe form factor GPU with 96GB HBM3 memory.
    • Offers more memory capacity for memory-intensive workloads.
  4. H100 SXM5 64 GB:
    • An SXM5 form factor GPU with 64GB HBM3 memory.
    • Operates at a ~700W TDP, designed for high-performance data center deployments.
  5. H100 SXM5 80 GB:
    • An SXM5 form factor GPU with 80GB HBM3 memory.
    • Offers a balance of memory capacity and performance for HPC and AI workloads.
  6. H100 SXM5 96 GB:
    • An SXM5 form factor GPU with 96GB HBM3 memory.
    • Provides the highest memory capacity in the H100 SXM5 series for memory-intensive workloads.
  7. H800 PCIe 80 GB:
    • A PCIe form factor GPU with 80GB HBM3 memory.
    • Operates at a ~350W TDP, making it suitable for standard server deployments.
  8. H800 SXM5:
    • An SXM5 form factor GPU with 80GB HBM3 memory.
    • Operates at a ~700W TDP, designed for high-performance data center deployments.

Summary:

  • H100 CNX: A custom form factor GPU for specific data center deployments.
  • H100 PCIe 80 GB: A PCIe form factor GPU with 80GB HBM3 memory.
  • H100 PCIe 96 GB: A PCIe form factor GPU with 96GB HBM3 memory.
  • H100 SXM5 64 GB: An SXM5 form factor GPU with 64GB HBM3 memory.
  • H100 SXM5 80 GB: An SXM5 form factor GPU with 80GB HBM3 memory.
  • H100 SXM5 96 GB: An SXM5 form factor GPU with 96GB HBM3 memory.
  • H800 PCIe 80 GB: A PCIe form factor GPU with 80GB HBM3 memory.
  • H800 SXM5: An SXM5 form factor GPU with 80GB HBM3 memory.
Request Best Price
SKU: N/A Categories: ,

NVIDIA H100 / H800 Data Center GPUs – CNX, PCIe 80GB / 96GB, SXM5 64GB / 80GB / 96GB, H800 PCIe 80GB, H800 SXM5 – Professional Graphics Cards

Here’s a detailed comparison table for the GPUs: H100 CNX, H100 PCIe 80 GB, H100 PCIe 96 GB, H100 SXM5 64 GB, H100 SXM5 80 GB, H100 SXM5 96 GB, H800 PCIe 80 GB, and H800 SXM5. These GPUs are part of NVIDIA’s H100 and H800 series, designed for high-performance computing (HPC), AI, and data center workloads.

Feature H100 CNX H100 PCIe 80 GB H100 PCIe 96 GB H100 SXM5 64 GB H100 SXM5 80 GB H100 SXM5 96 GB H800 PCIe 80 GB H800 SXM5
Architecture Hopper Hopper Hopper Hopper Hopper Hopper Hopper Hopper
FP32 Performance ~60 TFLOPs ~60 TFLOPs ~60 TFLOPs ~60 TFLOPs ~60 TFLOPs ~60 TFLOPs ~60 TFLOPs ~60 TFLOPs
FP64 Performance ~30 TFLOPs ~30 TFLOPs ~30 TFLOPs ~30 TFLOPs ~30 TFLOPs ~30 TFLOPs ~30 TFLOPs ~30 TFLOPs
Tensor Core Performance ~1,000 TFLOPs ~1,000 TFLOPs ~1,000 TFLOPs ~1,000 TFLOPs ~1,000 TFLOPs ~1,000 TFLOPs ~1,000 TFLOPs ~1,000 TFLOPs
Memory 80GB HBM3 80GB HBM3 96GB HBM3 64GB HBM3 80GB HBM3 96GB HBM3 80GB HBM3 80GB HBM3
Memory Bandwidth ~3 TB/s ~3 TB/s ~3 TB/s ~3 TB/s ~3 TB/s ~3 TB/s ~3 TB/s ~3 TB/s
Form Factor CNX (Custom) PCIe PCIe SXM5 SXM5 SXM5 PCIe SXM5
TDP ~700W ~350W ~350W ~700W ~700W ~700W ~350W ~700W
Use Case HPC, AI, ML HPC, AI, ML HPC, AI, ML HPC, AI, ML HPC, AI, ML HPC, AI, ML HPC, AI, ML HPC, AI, ML
Key Feature Custom form factor PCIe form factor PCIe form factor SXM5 form factor SXM5 form factor SXM5 form factor PCIe form factor SXM5 form factor

1. NVIDIA H100 CNX

  • Architecture: Built on NVIDIA’s Hopper architecture, optimized for custom data center deployments.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Equipped with 80GB of HBM3 memory, providing high bandwidth for data-intensive workloads.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring fast data access and processing.
  • Form Factor: CNX (Custom), designed for specific data center configurations.
  • TDP: Operates at a ~700W TDP, making it suitable for high-performance custom deployments.
  • Use Case: Designed for custom HPC and AI workloads, targeting specific data center needs.
  • Key Feature: A custom form factor GPU for tailored data center solutions.

2. NVIDIA H100 PCIe 80 GB

  • Architecture: Based on NVIDIA’s Hopper architecture, offering high performance in a standard PCIe form factor.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Comes with 80GB of HBM3 memory, providing ample memory for modern AI and HPC workloads.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring smooth performance in demanding applications.
  • Form Factor: PCIe, making it compatible with standard server configurations.
  • TDP: Operates at a ~350W TDP, balancing performance and power consumption.
  • Use Case: Ideal for HPC and AI workloads in standard server deployments.
  • Key Feature: A PCIe form factor GPU offering excellent performance for data centers.

3. NVIDIA H100 PCIe 96 GB

  • Architecture: Built on NVIDIA’s Hopper architecture, offering enhanced memory capacity in a PCIe form factor.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Equipped with 96GB of HBM3 memory, providing the highest memory capacity in the PCIe series.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring smooth performance in memory-intensive applications.
  • Form Factor: PCIe, making it compatible with standard server configurations.
  • TDP: Operates at a ~350W TDP, balancing performance and power consumption.
  • Use Case: Designed for memory-intensive HPC and AI workloads in standard server deployments.
  • Key Feature: A PCIe form factor GPU with the highest memory capacity in the series.

4. NVIDIA H100 SXM5 64 GB

  • Architecture: Based on NVIDIA’s Hopper architecture, offering high performance in an SXM5 form factor.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Comes with 64GB of HBM3 memory, providing high bandwidth for data-intensive workloads.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring fast data access and processing.
  • Form Factor: SXM5, designed for high-performance data center deployments.
  • TDP: Operates at a ~700W TDP, making it suitable for high-performance computing.
  • Use Case: Ideal for HPC and AI workloads in high-performance data centers.
  • Key Feature: An SXM5 form factor GPU offering excellent performance for data centers.

5. NVIDIA H100 SXM5 80 GB

  • Architecture: Built on NVIDIA’s Hopper architecture, offering a balance of memory capacity and performance in an SXM5 form factor.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Equipped with 80GB of HBM3 memory, providing ample memory for modern AI and HPC workloads.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring smooth performance in demanding applications.
  • Form Factor: SXM5, designed for high-performance data center deployments.
  • TDP: Operates at a ~700W TDP, making it suitable for high-performance computing.
  • Use Case: Designed for HPC and AI workloads in high-performance data centers.
  • Key Feature: An SXM5 form factor GPU offering a balance of memory capacity and performance.

6. NVIDIA H100 SXM5 96 GB

  • Architecture: Based on NVIDIA’s Hopper architecture, offering the highest memory capacity in the SXM5 series.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Comes with 96GB of HBM3 memory, providing the highest memory capacity in the series.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring smooth performance in memory-intensive applications.
  • Form Factor: SXM5, designed for high-performance data center deployments.
  • TDP: Operates at a ~700W TDP, making it suitable for high-performance computing.
  • Use Case: Ideal for memory-intensive HPC and AI workloads in high-performance data centers.
  • Key Feature: An SXM5 form factor GPU with the highest memory capacity in the series.

7. NVIDIA H800 PCIe 80 GB

  • Architecture: Built on NVIDIA’s Hopper architecture, offering high performance in a PCIe form factor.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Equipped with 80GB of HBM3 memory, providing ample memory for modern AI and HPC workloads.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring smooth performance in demanding applications.
  • Form Factor: PCIe, making it compatible with standard server configurations.
  • TDP: Operates at a ~350W TDP, balancing performance and power consumption.
  • Use Case: Designed for HPC and AI workloads in standard server deployments.
  • Key Feature: A PCIe form factor GPU offering excellent performance for data centers.

8. NVIDIA H800 SXM5

  • Architecture: Based on NVIDIA’s Hopper architecture, offering high performance in an SXM5 form factor.
  • Performance: Features ~60 TFLOPs of FP32 performance and ~30 TFLOPs of FP64 performance, with ~1,000 TFLOPs of Tensor Core performance for AI workloads.
  • Memory: Comes with 80GB of HBM3 memory, providing ample memory for modern AI and HPC workloads.
  • Memory Bandwidth: Offers ~3 TB/s, ensuring smooth performance in demanding applications.
  • Form Factor: SXM5, designed for high-performance data center deployments.
  • TDP: Operates at a ~700W TDP, making it suitable for high-performance computing.
  • Use Case: Ideal for HPC and AI workloads in high-performance data centers.
  • Key Feature: An SXM5 form factor GPU offering excellent performance for data centers.

Summary:

  • H100 CNX: A custom form factor GPU for specific data center deployments.
  • H100 PCIe 80 GB: A PCIe form factor GPU with 80GB HBM3 memory.
  • H100 PCIe 96 GB: A PCIe form factor GPU with 96GB HBM3 memory.
  • H100 SXM5 64 GB: An SXM5 form factor GPU with 64GB HBM3 memory.
  • H100 SXM5 80 GB: An SXM5 form factor GPU with 80GB HBM3 memory.
  • H100 SXM5 96 GB: An SXM5 form factor GPU with 96GB HBM3 memory.
  • H800 PCIe 80 GB: A PCIe form factor GPU with 80GB HBM3 memory.
  • H800 SXM5: An SXM5 form factor GPU with 80GB HBM3 memory.
NVIDIA H100 series

H100 CNX, H100 PCIe 80 GB, H100 PCIe 96 GB, H100 SXM5 64 GB, H100 SXM5 80 GB, H100 SXM5 96 GB, H800 PCIe 80 GB, H800 SXM5

Related products