NVIDIA H100 PCIe vs Intel Arc Pro A40. Specifications

Comparison of basic technical data of graphics cards NVIDIA H100 PCIe and Intel Arc Pro A40, chip, information processing units.

• Architecture
  Hopper Generation 12.7 / Alchemist (Pro-Series)
• Execution Units
  7296 1024
• GPU Chip
  GH100 DG2-128
• Render Output Units
  24 32
• Texture Units
  456 64
• L1 Cache
  - No
• L2 Cache
  - 4 MB
• RT Cores
  - 8
• Shader
  - 128

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

• Memory Bandwidth
  1,280 GB/s 192 GB/s
• Memory Bus
  5120 bit 96 bit
• Memory Clock
  1000 MHz 2000 MHz
• Memory Data Rate
  2 Gbps 16 Gbps
• Memory Size
  80 GB 6 GB
• Memory Type
  HBM2e GDDR6

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

• Base Clock
  1065 MHz 1500 MHz
• Boost Clock
  1650 MHz 1700 MHz

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

• System Power Supply
  750 W (PSU recommendation) 250 W (PSU recommendation)
• TDP
  350 W 50 W

Connectivity and connections.

• Max. Displays
  1 -
• PCIe Interface
  PCIe 5.0 x16 PCIe 4.0 x8
• Power Connectors
  8-pin EPS No
• DisplayPort Ports
  - 4x mini-DisplayPort 2.0

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

• CUDA
  9.0 -
• DirectX
  No 12_2
• OpenCL
  3.0 3.0
• OpenGL
  No 4.6
• Shader Model
  No 6.6
• Vulkan
  No 1.3
• Max. resolution
  - 3840 × 2160

Comparison of interfaces and release dates for GPU Intel Arc Pro A40 and NVIDIA H100 PCIe.

• Process Technology
  4 nm 6 nm
• Release date
  Q2/2022 Q3/2022

• FP16 (theoretical)
  96.31 TFLOPS 6.963 TFLOPS (2:1)
• FP32 (theoretical)
  24.08 TFLOPS 3.482 TFLOPS
• FP64 (theoretical)
  12.04 TFLOPS 0.8704 TFLOPS
• Pixel Rate
  39.60 GPixel/s 54.40 GPixel/s
• Texture Rate
  752.4 GTexel/s 108.8 GTexel/s