显卡FP32浮点性能TFLOPS排行

显卡FP32浮点性能介绍

FP32 浮点性能指的是显卡在进行 32位单精度浮点数计算 时的处理能力,常用于图形渲染、科学计算、AI 推理、物理模拟等需要高精度数学计算的场景。“FP”是 Floating Point(浮点数) 的缩写 ,“32”表示它是 32位的单精度浮点数(相对于 FP16 是半精度,FP64 是双精度),是显卡重要的性能指标之一。

显卡FP32浮点性能排名(持续更新中)

  1. NVIDIA GeForce GTX 1080 Ti 354.4 TFLOPS
  2. NVIDIA GeForce RTX 5090 104.8 TFLOPS
  3. NVIDIA GeForce RTX 5090 D 104.8 TFLOPS
  4. NVIDIA GeForce RTX 5090 DD 104.8 TFLOPS
  5. NVIDIA GeForce RTX 4090 82.58 TFLOPS
  6. NVIDIA GeForce RTX 4090 D 73.54 TFLOPS
  7. NVIDIA GeForce RTX 4080 Ti 73.50 TFLOPS
  8. AMD Radeon RX 7900 XTX 61.39 TFLOPS
  9. NVIDIA GeForce RTX 5080 SUPER 56.28 TFLOPS
  10. NVIDIA GeForce RTX 5080 56.28 TFLOPS
  11. NVIDIA GeForce RTX 4080 SUPER 52.22 TFLOPS
  12. AMD Radeon RX 7900 XT 51.48 TFLOPS
  13. NVIDIA GeForce RTX 4080 48.74 TFLOPS
  14. AMD Radeon RX 9070 XT 48.66 TFLOPS
  15. AMD Radeon RX 7900 GRE 45.98 TFLOPS
  16. NVIDIA GeForce RTX 4070 Ti SUPER 44.1 TFLOPS
  17. NVIDIA GeForce RTX 5070 Ti SUPER 43.94 TFLOPS
  18. NVIDIA GeForce RTX 5070 Ti 43.94 TFLOPS
  19. NVIDIA GeForce RTX 4080 12GB 40.09 TFLOPS
  20. NVIDIA GeForce RTX 4070 Ti 40.09 TFLOPS
  21. NVIDIA GeForce RTX 3090 Ti 40 TFLOPS
  22. AMD Radeon RX 7900M 38.52 TFLOPS
  23. AMD Radeon RX 7800 XT 37.32 TFLOPS
  24. AMD Radeon RX 9070 36.13 TFLOPS
  25. AMD Radeon RX 7800M 35.87 TFLOPS
  26. NVIDIA GeForce RTX 3090 35.58 TFLOPS
  27. NVIDIA GeForce RTX 4070 SUPER 35.48 TFLOPS
  28. AMD Radeon RX 7700 XT 35.17 TFLOPS
  29. AMD Radeon RX 9070 GRE 34.28 TFLOPS
  30. NVIDIA GeForce RTX 3080 Ti 20 GB 34.10 TFLOPS
  31. NVIDIA GeForce RTX 3080 Ti 34.1 TFLOPS
  32. NVIDIA GeForce RTX 4090 Mobile 32.98 TFLOPS
  33. NVIDIA GeForce RTX 5070 SUPER 32.15 TFLOPS
  34. AMD Radeon RX 7700 31.95 TFLOPS
  35. NVIDIA GeForce RTX 5090 Mobile 31.80 TFLOPS
  36. NVIDIA GeForce RTX 5070 30.87 TFLOPS
  37. NVIDIA GeForce RTX 3080 12GB 30.64 TFLOPS
  38. NVIDIA GeForce RTX 3080 29.77 TFLOPS
  39. NVIDIA GeForce RTX 4070 29.15 TFLOPS
  40. NVIDIA GeForce RTX 4090 Max-Q 28.31 TFLOPS
  41. AMD Radeon RX 9060 XT 8GB 25.64 TFLOPS
  42. AMD Radeon RX 9060 XT 16GB 25.64 TFLOPS
  43. NVIDIA GeForce RTX 4080 Mobile 24.72 TFLOPS
  44. NVIDIA GeForce RTX 5060 Ti 16GB 24.7 TFLOPS
  45. AMD Radeon RX 6950 XT 23.8 TFLOPS
  46. NVIDIA GeForce RTX 5060 Ti 23.7 TFLOPS
  47. AMD Radeon RX 6900 XT 23.04 TFLOPS
  48. NVIDIA GeForce RTX 5080 Mobile 23.04 TFLOPS
  49. AMD Radeon RX 7600 XT 22.57 TFLOPS
  50. AMD Radeon RX 7650 GRE 22.08 TFLOPS
  51. NVIDIA GeForce RTX 4060 Ti 16GB 22.06 TFLOPS
  52. NVIDIA GeForce RTX 4060 Ti 22.06 TFLOPS
  53. NVIDIA GeForce RTX 3070 Ti 21.75 TFLOPS
  54. AMD Radeon RX 7600 21.75 TFLOPS
  55. AMD Radeon RX 6800 XT 20.74 TFLOPS
  56. AMD Radeon RX 7700S 20.48 TFLOPS
  57. NVIDIA GeForce RTX 3070 20.31 TFLOPS
  58. AMD Radeon RX 7600M XT 20.23 TFLOPS
  59. NVIDIA GeForce RTX 4080 Max-Q 20.04 TFLOPS
  60. NVIDIA GeForce RTX 5060 19.18 TFLOPS
  61. NVIDIA GeForce RTX 3080 Mobile 18.98 TFLOPS
  62. NVIDIA GeForce RTX 3080 Ti Mobile 18.71 TFLOPS
  63. AMD Playstation 5 Pro GPU 18.05 TFLOPS
  64. AMD Radeon RX 7600M 17.27 TFLOPS
  65. NVIDIA GeForce RTX 5070 Ti Mobile 17.04 TFLOPS
  66. NVIDIA GeForce RTX 3060 Ti 16.2 TFLOPS
  67. NVIDIA GeForce RTX 3060 Ti GDDR6X 16.20 TFLOPS
  68. AMD Radeon RX 6800 16.17 TFLOPS
  69. NVIDIA GeForce RTX 3070 Mobile 15.97 TFLOPS
  70. NVIDIA GeForce RTX 3070 Ti Mobile 15.88 TFLOPS
  71. AMD Radeon RX 7600S 15.77 TFLOPS
  72. NVIDIA GeForce RTX 4070 Mobile 15.62 TFLOPS
  73. NVIDIA GeForce RTX 3080 Max-Q 15.30 TFLOPS
  74. NVIDIA GeForce RTX 4060 15.11 TFLOPS
  75. NVIDIA GeForce RTX 4050 13.52 TFLOPS
  76. NVIDIA GeForce RTX 2080 Ti 13.45 TFLOPS
  77. AMD Radeon RX 6750 XT 13.31 TFLOPS
  78. AMD Radeon RX 6700 XT 13.21 TFLOPS
  79. AMD Radeon RX 6850M XT 13.21 TFLOPS
  80. NVIDIA GeForce RTX 3070 Max-Q 13.21 TFLOPS
  81. AMD Radeon RX 6750 GRE 12GB 13.21 TFLOPS
  82. NVIDIA GeForce RTX 5050 13.17 TFLOPS
  83. NVIDIA GeForce RTX 5070 Mobile 13.13 TFLOPS
  84. NVIDIA GeForce RTX 5050 Mobile 12.90 TFLOPS
  85. NVIDIA GeForce RTX 3060 12.74 TFLOPS
  86. NVIDIA GeForce RTX 3060 8GB 12.74 TFLOPS
  87. AMD Radeon RX 6800M 12.24 TFLOPS
  88. NVIDIA GeForce RTX 3070 Ti Max-Q 12.19 TFLOPS
  89. AMD Xbox Series X 6nm GPU 12.15 TFLOPS
  90. AMD Xbox Series X GPU 12.15 TFLOPS
  91. AMD Radeon 8060S 11.96 TFLOPS
  92. NVIDIA GeForce RTX 4060 Mobile 11.61 TFLOPS
  93. Intel Arc B570 11.52 TFLOPS
  94. NVIDIA GeForce RTX 4070 Max-Q 11.34 TFLOPS
  95. AMD Radeon RX 6700 11.29 TFLOPS
  96. AMD Radeon RX 6750 GRE 10GB 11.29 TFLOPS
  97. NVIDIA GeForce RTX 2080 SUPER 11.15 TFLOPS
  98. AMD Radeon RX 6700M 11.06 TFLOPS
  99. NVIDIA GeForce RTX 3060 Mobile 10.94 TFLOPS
  100. AMD Radeon RX 6650 XT 10.79 TFLOPS
  101. AMD Radeon RX 6600 XT 10.6 TFLOPS
  102. AMD Playstation 5 GPU 10.29 TFLOPS
  103. AMD Radeon RX 5700 XT 50th Anniversary 10.14 TFLOPS
  104. NVIDIA GeForce RTX 2080 10.07 TFLOPS
  105. AMD Radeon RX 6650M XT 9.896 TFLOPS
  106. NVIDIA GeForce RTX 3060 Max-Q 9.846 TFLOPS
  107. AMD Radeon RX 5700 XT 9.754 TFLOPS
  108. NVIDIA GeForce RTX 5060 Mobile 9.684 TFLOPS
  109. NVIDIA GeForce RTX 2080 Super Mobile 9.585 TFLOPS
  110. AMD Radeon 8050S 9.564 TFLOPS
  111. NVIDIA GeForce RTX 2080 Mobile 9.362 TFLOPS
  112. NVIDIA GeForce RTX 3050 8GB 9.098 TFLOPS
  113. NVIDIA GeForce RTX 2070 SUPER 9.062 TFLOPS
  114. NVIDIA GeForce RTX 4060 Max-Q 9.032 TFLOPS
  115. NVIDIA GeForce RTX 3050 OEM 8.986 TFLOPS
  116. NVIDIA GeForce RTX 4050 Mobile 8.986 TFLOPS
  117. AMD Radeon RX 6600 LE 8.942 TFLOPS
  118. AMD Radeon RX 6600 8.928 TFLOPS
  119. NVIDIA GeForce GTX 1080 Mobile 8.878 TFLOPS
  120. NVIDIA GeForce GTX 1080 8.873 TFLOPS
  121. AMD Radeon RX 6600M 8.659 TFLOPS
  122. AMD Radeon RX 6650M 8.659 TFLOPS
  123. AMD Radeon RX 6800S 8.602 TFLOPS
  124. AMD Ryzen Z2 8.294 TFLOPS
  125. AMD Radeon 780M 8.294 TFLOPS
  126. AMD Ryzen Z1 Extreme 8.294 TFLOPS
  127. NVIDIA GeForce RTX 4050 Max-Q 8.218 TFLOPS
  128. NVIDIA GeForce GTX 1070 Ti 8.186 TFLOPS
  129. AMD Radeon RX 5700 7.949 TFLOPS
  130. AMD Radeon RX 5700M 7.926 TFLOPS
  131. NVIDIA GeForce RTX 3050 6GB Mobile 7.639 TFLOPS
  132. NVIDIA GeForce RTX 2070 7.465 TFLOPS
  133. AMD Radeon RX 5600 XT 7.188 TFLOPS
  134. NVIDIA GeForce RTX 2060 SUPER 7.181 TFLOPS
  135. NVIDIA GeForce RTX 2060 12 GB 7.181 TFLOPS
  136. AMD Radeon RX 6700S 7.168 TFLOPS
  137. AMD Radeon RX 6600S 7.168 TFLOPS
  138. NVIDIA GeForce RTX 3050 4 GB 7.127 TFLOPS
  139. NVIDIA GeForce RTX 2070 Super Mobile 7.066 TFLOPS
  140. NVIDIA GeForce GTX 1080 Max-Q 6.994 TFLOPS
  141. NVIDIA GeForce RTX 2070 Mobile 6.843 TFLOPS
  142. NVIDIA GeForce RTX 3050 6GB 6.774 TFLOPS
  143. NVIDIA GeForce GTX 1070 Mobile 6.738 TFLOPS
  144. NVIDIA GeForce RTX 2060 SUPER Mobile 6.659 TFLOPS
  145. NVIDIA GeForce RTX 2080 Super Max-Q 6.636 TFLOPS
  146. NVIDIA GeForce GTX 1070 6.463 TFLOPS
  147. NVIDIA GeForce RTX 2060 6.451 TFLOPS
  148. NVIDIA GeForce RTX 2080 Max-Q 6.447 TFLOPS
  149. NVIDIA GeForce RTX 3050 4GB Mobile 6.18 TFLOPS
  150. AMD Radeon 890M 5.939 TFLOPS
  151. NVIDIA GeForce RTX 2070 Super Max-Q 5.914 TFLOPS
  152. AMD Radeon RX 5600M 5.829 TFLOPS
  153. AMD Radeon RX 6550M 5.816 TFLOPS
  154. AMD Radeon RX 6500 XT 5.765 TFLOPS
  155. NVIDIA GeForce GTX 1070 Max-Q 5.648 TFLOPS
  156. AMD Ryzen AI Z2 Extreme 5.530 TFLOPS
  157. AMD Ryzen Z2 Extreme 5.530 TFLOPS
  158. NVIDIA GeForce RTX 3050 Mobile 5.501 TFLOPS
  159. NVIDIA GeForce RTX 2070 Max-Q 5.46 TFLOPS
  160. NVIDIA GeForce GTX 1660 Ti 5.437 TFLOPS
  161. AMD Radeon 760M 5.323 TFLOPS
  162. NVIDIA GeForce RTX 3050 Ti Mobile 5.299 TFLOPS
  163. AMD Radeon RX 5500 XT 5.196 TFLOPS
  164. AMD Radeon RX 5500 5.196 TFLOPS
  165. NVIDIA GeForce GTX 1660 Super 5.027 TFLOPS
  166. NVIDIA GeForce GTX 1660 5.027 TFLOPS
  167. AMD Radeon RX 6500M 4.915 TFLOPS
  168. NVIDIA GeForce GTX 1660 Ti Mobile 4.884 TFLOPS
  169. NVIDIA GeForce RTX 3050 A Mobile 4.813 TFLOPS
  170. AMD Radeon RX 5500M 4.632 TFLOPS
  171. NVIDIA GeForce RTX 2060 Mobile 4.608 TFLOPS
  172. NVIDIA GeForce RTX 2060 Max-Q 4.55 TFLOPS
  173. AMD Radeon 880M 4.454 TFLOPS
  174. NVIDIA GeForce GTX 1650 SUPER 4.416 TFLOPS
  175. NVIDIA GeForce GTX 1060 6 GB 4.375 TFLOPS
  176. NVIDIA GeForce GTX 1060 5GB 4.375 TFLOPS
  177. NVIDIA GeForce RTX 3050 Max-Q 4.329 TFLOPS
  178. NVIDIA GeForce GTX 1060 Mobile 4.278 TFLOPS
  179. NVIDIA GeForce GTX 1660 Ti Max-Q 4.101 TFLOPS
  180. AMD Radeon RX 5300M 4.069 TFLOPS
  181. AMD Xbox Series S GPU 4.006 TFLOPS
  182. NVIDIA GeForce GTX 1060 3 GB 3.935 TFLOPS
  183. NVIDIA GeForce GTX 1060 Max-Q 3.789 TFLOPS
  184. AMD Radeon RX 6450M 3.779 TFLOPS
  185. AMD Radeon RX 6400 3.565 TFLOPS
  186. AMD Radeon 680M 3.379 TFLOPS
  187. NVIDIA GeForce GTX 1650 Mobile 3.195 TFLOPS
  188. AMD Radeon 860M 3.072 TFLOPS
  189. NVIDIA GeForce GTX 1650 Ti Mobile 3.041 TFLOPS
  190. NVIDIA GeForce GTX 1650 2.984 TFLOPS
  191. AMD Radeon 740M 2.867 TFLOPS
  192. NVIDIA GeForce MX550 2.703 TFLOPS
  193. AMD Ryzen Z1 2.560 TFLOPS
  194. NVIDIA GeForce GTX 1050 Ti Mobile 2.488 TFLOPS
  195. NVIDIA GeForce GTX 1650 Ti Max-Q 2.458 TFLOPS
  196. NVIDIA GeForce GTX 1050 3GB 2.332 TFLOPS
  197. NVIDIA GeForce GTX 1650 Max-Q 2.304 TFLOPS
  198. NVIDIA GeForce GTX 1050 3GB Mobile 2.215 TFLOPS
  199. NVIDIA GeForce GTX 1050 Ti 2.138 TFLOPS
  200. NVIDIA GeForce GTX 1050 Ti Max-Q 1.983 TFLOPS
  201. NVIDIA GeForce GTX 1050 Mobile 1.911 TFLOPS
  202. NVIDIA GeForce MX350 1.879 TFLOPS
  203. NVIDIA GeForce GTX 1050 1.862 TFLOPS
  204. NVIDIA GeForce GTX 1630 1.828 TFLOPS
  205. NVIDIA GeForce MX450 12W 1.667 TFLOPS
  206. AMD Steam Deck OLED GPU 1.638 TFLOPS
  207. AMD Steam Deck GPU 1.638 TFLOPS
  208. AMD Radeon 660M 1.459 TFLOPS
  209. NVIDIA GeForce MX230 0.783 TFLOPS
  210. AMD Radeon 610M 0.5632 TFLOPS
  211. NVIDIA GeForce GTX 1050 Max-Q 0.145 TFLOPS
  212. NVIDIA GeForce MX330 0.122 TFLOPS
  213. NVIDIA GeForce MX250 0.121 TFLOPS
  214. NVIDIA GeForce MX150 0.117 TFLOPS
  215. Intel UHD Graphics 730 N/A

显卡FP32浮点性能计算方法

显卡的 FP32 浮点性能(单位是 TFLOPS)可以大致通过以下公式估算:

FP32 性能(TFLOPS) = CUDA核心数 × 主频 × 每个时钟的操作数 × 2(如果是 FMA 指令) ÷ 1,000,000

以 NVIDIA RTX 3080 为例,它的 CUDA核心数:8704,主频:1.71 GHz,根据公式计算:8704 × 1.71 × 2 ≈ 29.77 TFLOPS,这就意味着 RTX 3080 每秒大约能进行 29.77 万亿次 FP32 运算。

显卡FP32浮点性能的意义

游戏更依赖的是 图形渲染能力(光栅/着色器等),FP32 性能虽然相关,但不是决定性因素,但它确实影响一些高负载的图形计算,比如光线追踪、物理模拟等。

AI/深度学习方面,FP32 性能直接影响神经网络的前向推理和训练速度,但现在更高效的是 FP16 或更低精度的混合精度计算(Tensor Core 支持)。

科学计算 / 工业仿真的高精度需求场景下,FP32 是基本门槛,有些任务还要求 FP64,更适合专业卡如 NVIDIA A100、Quadro 系列。