Graphics card NVIDIA H100 PCIe, specifications and benchmarks

This section outlines the key features of the graphics card NVIDIA H100 PCIe.

• GPU Chip
  GH100
• Execution units
  7296
• Render Output Units
  24
• Texture Units
  456

Specifications for onboard video memory, including type (e.g., GDDR6), capacity, and bandwidth. Higher capacity and bandwidth contribute to improved gaming and rendering performance.

• Memory Size
  80 GB
• Memory Type
  HBM2e
• Memory Clock
  1 GHz
• Memory Speed
  2 Gbps
• Memory Bandwith
  1,280 GB/s
• Memory Interface
  5120 bit

Information on the base and boost clock speeds of the graphics card, which directly affect its performance in gaming and rendering applications. NVIDIA H100 PCIe.

• Base Clock
  1.065 GHz
• Boost Clock
  1.65 GHz

Details on the thermal design power (TDP) of the graphics card, including the maximum thermal output under standard and overclocked conditions, guiding the selection of cooling solutions.

• TDP
  350 W
• Power Connectors
  8-pin EPS
• Suggested PSU
  750 W

An overview of the hardware capabilities of the graphics card, including support for technologies like ray tracing, DLSS, and Vulkan, which enhance graphical fidelity and performance.

• Max. Displays
  1
• DirectX
  No
• OpenGL
  No
• OpenCL
  3.0
• Vulkan
  No
• Shader Model
  No
• CUDA
  9.0

• Architecture
  Hopper
• GPU Interface
  PCIe 5.0 x16
• Structure size
  4 nm
• Release date
  Q2/2022

• Pixel Rate
  39.60 GPixel/s
• Texture Rate
  752.4 GTexel/s
• FP16 (half)
  96.31 TFLOPS
• FP32 (float)
  24.08 TFLOPS
• FP64 (double)
  12.04 TFLOPS

Common positions NVIDIA H100 PCIe GPU in popular benchmarks, for comparison with other models.