NVIDIA L40G 对 NVIDIA A800 PCIe 40 GB. 规格

显卡的基本技术数据比较 NVIDIA L40G 和 NVIDIA A800 PCIe 40 GB, 芯片，信息处理单元.

• 建筑学
  Ada Lovelace Ampere
• Execution Units
  18176 6912
• GPU芯片
  AD102 GA100
• L1 Cache
  128 KB 192 KB
• L2 Cache
  48 MB 40 MB
• 渲染输出单位
  192 160
• RT Cores
  142 -
• 纹理单位
  568 432

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

• Memory Bandwidth
  864 GB/s 1,555 GB/s
• Memory Bus
  384 bit 5120 bit
• Memory Clock
  2250 MHz 1215 MHz
• Memory Data Rate
  18 Gbps 2.4 Gbps
• 内存大小
  24 GB 40 GB
• 存储器类型
  GDDR6 HBM2e

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

• 基准频率
  1005 MHz 765 MHz
• 最大频率
  2475 MHz 1410 MHz

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

• System Power Supply
  700 W (PSU recommendation) 600 W (PSU recommendation)
• TDP
  300 W 250 W

连通性和连接.

• DisplayPort Ports
  3 × DisplayPort 1.4a -
• HDMI端口
  1 × HDMI 2.1 -
• 最大。显示
  4 -
• PCIe Interface
  PCIe 4.0 x16 PCIe 4.0 x16
• Power Connectors
  1x 16-pin 8-pin EPS

在电脑游戏中被充分使用的技术数据.

• CUDA
  8.9 8.0
• DirectX
  12.2 12.2
• 最大分辨率
  3840 × 2160 3840 × 2160
• OpenCL
  3.0 3.0
• OpenGL
  4.6 4.6
• Shader Model
  6.7 -
• Vulkan
  1.3 -

所比较的设备在尺寸、重量和插槽上的差异

• 长度
  267 mm 267 mm
• Slot Size
  Dual-slot Dual-slot
• Width
  111 mm 112 mm

GPU的接口和发布日期的比较 NVIDIA A800 PCIe 40 GB 和 NVIDIA L40G.

• Process Technology
  4 nm 7 nm
• 发布日期
  Q3/2022 Q3/2022

• FP16 (theoretical)
  89.97 TFLOPS (1:1) 77.97 TFLOPS (4:1)
• FP32 (theoretical)
  89.97 TFLOPS 19.49 TFLOPS
• FP64 (theoretical)
  1,406 GFLOPS (1:64) 9.746 TFLOPS (1:2)
• Pixel Rate
  475.2 GPixel/s 225.6 GPixel/s
• Texture Rate
  1,406 GTexel/s 609.1 GTexel/s