NVIDIA H100 PCIe contro NVIDIA GeForce RTX 3090 Ti. Specifiche, prestazioni, test

Confronto dei dati tecnici di base delle schede grafiche NVIDIA H100 PCIe e NVIDIA GeForce RTX 3090 Ti, chip, unità di elaborazione delle informazioni.

• Architettura
  Hopper Ampere
• Execution Units
  7296 -
• Chip GPU
  GH100 GA102-350-A1
• Unità di uscita del rendering
  24 112
• Unità di texture
  456 336
• Basato su
  - GA102
• CUDA Cores
  - 10752
• L1 Cache
  - 128 KB
• L2 Cache
  - 6 MB
• Shader
  - 10752

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 1,008 GB/s
• Memory Bus
  5120 bit 384 bit
• Memory Clock
  1000 MHz -
• Memory Data Rate
  2 Gbps -
• Dimensione della memoria
  80 GB 24 GB
• Tipo di memoria
  HBM2e GDDR6X
• Memory Speed (legacy frequency label)
  - 1300 MHz

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

• Frequenza di base
  1065 MHz 1560 MHz
• Frequenza massima
  1650 MHz 1860 MHz
• Overclock
  - No

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

• System Power Supply
  750 W (PSU recommendation) 850 W (System power requirement)
• TDP
  350 W 450 W
• Graphics Card Power
  - 450 W
• Maximum GPU Temperature
  - 92 °C

Connettività e collegamenti.

• Max. Visualizzazioni
  1 4
• PCIe Interface
  PCIe 5.0 x16 PCIe 4.0 x16
• Power Connectors
  8-pin EPS 3x PCIe 8-pin cables (adapter in box) OR 450W or greater PCIe Gen 5 cable
• DisplayPort Ports
  - 3 × DisplayPort 1.4a
• HDCP Version
  - 2.3
• Porte HDMI
  - 1 × HDMI 2.1
• Other Display Outputs (unparsed)
  - HDMI

Dati tecnici utilizzati al massimo delle loro potenzialità nei giochi per computer.

• 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
• Risoluzione massima
  - 7680 × 4320

Confronto tra le interfacce e le date di rilascio per le GPU NVIDIA GeForce RTX 3090 Ti e NVIDIA H100 PCIe.

• Process Technology
  4 nm 8 nm
• Data di rilascio
  Q2/2022 -
• Manufacturer specification scope
  - NVIDIA Founders Edition/reference design. Clock specifications apply to gaming use. AIB shipping specifications and complete-system power needs may differ.

• FP16 (theoretical)
  96.31 TFLOPS 40 TFLOPS
• FP32 (theoretical)
  24.08 TFLOPS 40 TFLOPS
• FP64 (theoretical)
  12.04 TFLOPS 0.625 TFLOPS
• Pixel Rate
  39.60 GPixel/s 208.3 GPixel/s
• Texture Rate
  752.4 GTexel/s 625.0 GTexel/s