NVIDIA H100 SXM5 vs NVIDIA GeForce RTX 3080 Ti Max-Q. Specifications

Comparison of basic technical data of graphics cards NVIDIA H100 SXM5 and NVIDIA GeForce RTX 3080 Ti Max-Q, chip, information processing units.

• Architecture
  Hopper Ampere
• Execution Units
  8448 7424
• GPU Chip
  GH100 GN20-E8-A1
• L1 Cache
  192 KB -
• L2 Cache
  50 MB -
• Render Output Units
  24 96
• Texture Units
  528 232
• Based on
  - GA103S

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 384 GB/s
• Memory Bus
  5120 bit 256 bit
• Memory Clock
  1500 MHz -
• Memory Data Rate
  3 Gbps -
• Memory Size
  80 GB 16 GB
• Memory Type
  HBM3 GDDR6
• Memory Speed (legacy frequency label)
  - 1500 MHz

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

• Base Clock
  1065 MHz 600 MHz
• Boost Clock
  1780 MHz 1125 MHz
• Overclocking
  - No

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

• Board Number
  PG520 -
• System Power Supply
  1100 W (PSU recommendation) -
• TDP
  700 W 80 W

Connectivity and connections.

• Max. Displays
  1 -
• PCIe Interface
  PCIe 5.0 x16 PCIe 4.0 x16
• Power Connectors
  8-pin EPS No

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

• CUDA
  9.0 -
• DirectX
  No 12_2
• Max. resolution
  3840x2160@60Hz 7680 × 4320
• OpenCL
  3.0 3.0
• OpenGL
  No 4.6
• Shader Model
  No 6.5
• Vulkan
  No 1.3

Comparison of interfaces and release dates for GPU NVIDIA GeForce RTX 3080 Ti Max-Q and NVIDIA H100 SXM5.

• Process Technology
  4 nm 8 nm
• Release date
  Q2/2022 -

• FP16 (theoretical)
  120.3 TFLOPS 16.7 TFLOPS
• FP32 (theoretical)
  30.07 TFLOPS 16.7 TFLOPS
• FP64 (theoretical)
  15.04 TFLOPS 0.261 TFLOPS
• Pixel Rate
  42.72 GPixel/s 108.0 GPixel/s
• Texture Rate
  939.8 GTexel/s 261.0 GTexel/s