NVIDIA L40 对 NVIDIA GeForce GTX 1630. 规格

显卡的基本技术数据比较 NVIDIA L40 和 NVIDIA GeForce GTX 1630, 芯片，信息处理单元.

• 建筑学
  Ada Lovelace Turing
• Execution Units
  18176 512
• GPU芯片
  AD102 TU117
• L1 Cache
  128 KB 64 KB
• L2 Cache
  96 MB 1024 KB
• 渲染输出单位
  192 16
• RT Cores
  142 -
• 纹理单位
  568 32
• CUDA Cores
  - 512

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

• Memory Bandwidth
  864 GB/s 96 GB/s
• Memory Bus
  384 bit 64 bit
• Memory Clock
  2250 MHz 1500 MHz
• Memory Data Rate
  18 Gbps 12 Gbps
• 内存大小
  48 GB 4 GB
• 存储器类型
  GDDR6 GDDR6

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

• 基准频率
  735 MHz 1740 MHz
• 最大频率
  2490 MHz 1785 MHz

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

• System Power Supply
  700 W (PSU recommendation) 300 W (Minimum system power requirement)
• TDP
  300 W 75 W
• Board Number
  - PG174 SKU 0
• Graphics Card Power
  - 75 W
• Maximum GPU Temperature
  - 90-94 °C

连通性和连接.

• DisplayPort Ports
  4 × DisplayPort 1.4a 1 × DisplayPort 1.4a
• 最大。显示
  4 -
• PCIe Interface
  PCIe 4.0 x16 CIe 3.0 x8
• Power Connectors
  1x 16-pin 6-pin
• DVI端口
  - 1 × DVI
• HDMI端口
  - 1 × HDMI 2.0
• Other Display Outputs (unparsed)
  - HDMI 2.0b; DL-DVI-D; DP 1.4a

在电脑游戏中被充分使用的技术数据.

• CUDA
  8.9 7.5
• DirectX
  12.2 12.1
• 最大分辨率
  3840 × 2160 2560 × 1440
• OpenCL
  3.0 3.0
• OpenGL
  4.6 4.6
• Shader Model
  6.7 6.6
• Vulkan
  1.3 1.3

所比较的设备在尺寸、重量和插槽上的差异

• 长度
  267 mm -
• Slot Size
  Dual-slot -
• Width
  111 mm -

GPU的接口和发布日期的比较 NVIDIA GeForce GTX 1630 和 NVIDIA L40.

• Process Technology
  4 nm 12 nm
• 发布日期
  Q3/2022 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.
• Multi-monitor support
  - 是
• Verification scope
  - Explicit manufacturer fields refreshed; other original historical fields retained, not all reverified. Partial composite conflicts are preserved.

• FP16 (theoretical)
  90.52 TFLOPS (1:1) 3.656 TFLOPS
• FP32 (theoretical)
  90.52 TFLOPS 1.828 TFLOPS
• FP64 (theoretical)
  1.414 GFLOPS (1:64) 0.05712 TFLOPS
• Pixel Rate
  478.1 GPixel/s 28.56 GPixel/s
• Texture Rate
  1.414 GTexel/s 57.12 GTexel/s