NVIDIA H100 SXM5 vs NVIDIA H100 PCIe. Specifications

Comparison of basic technical data of graphics cards NVIDIA H100 SXM5 and NVIDIA H100 PCIe, chip, information processing units.

• Architecture
  Hopper Hopper
• Execution Units
  8448 7296
• GPU Chip
  GH100 GH100
• L1 Cache
  192 KB -
• L2 Cache
  50 MB -
• Render Output Units
  24 24
• Texture Units
  528 456

Video memory capacity, type, bus width and bandwidth describe different properties. These values alone do not determine graphics performance.

• Memory Bandwidth
  1,920 GB/s 1,280 GB/s
• Memory Bus
  5120 bit 5120 bit
• Memory Clock
  1500 MHz 1000 MHz
• Memory Data Rate
  3 Gbps 2 Gbps
• Memory Size
  80 GB 80 GB
• Memory Type
  HBM3 HBM2e

Published GPU clock frequencies describe different operating modes. Clock speed alone does not determine graphics performance.

• Base Clock
  1065 MHz 1065 MHz
• Boost Clock
  1780 MHz 1650 MHz

Graphics card power limits and recommended system power supply describe different specifications.

• Board Number
  PG520 -
• System Power Supply
  1100 W (PSU recommendation) 750 W (PSU recommendation)
• TDP
  700 W 350 W

Connectivity and connections.

• Max. Displays
  1 1
• PCIe Interface
  PCIe 5.0 x16 PCIe 5.0 x16
• Power Connectors
  8-pin EPS 8-pin EPS

Technical data that is used to its full potential in computer games.

• CUDA
  9.0 9.0
• DirectX
  No No
• Max. resolution
  3840x2160@60Hz -
• OpenCL
  3.0 3.0
• OpenGL
  No No
• Shader Model
  No No
• Vulkan
  No No

Comparison of interfaces and release dates for GPU NVIDIA H100 PCIe and NVIDIA H100 SXM5.

• Process Technology
  4 nm 4 nm
• Release date
  Q2/2022 Q2/2022

• FP16 (theoretical)
  120.3 TFLOPS 96.31 TFLOPS
• FP32 (theoretical)
  30.07 TFLOPS 24.08 TFLOPS
• FP64 (theoretical)
  15.04 TFLOPS 12.04 TFLOPS
• Pixel Rate
  42.72 GPixel/s 39.60 GPixel/s
• Texture Rate
  939.8 GTexel/s 752.4 GTexel/s