GPU/CPU浮点性能峰值参考

AMD,NVIDIA,CPU架构和编程方法不同,不能直接用标称浮点和流处理器比较实际处理性能.
所有数据来自网络,多数型号官网并无浮点说明.
LAPACK
LINPACK
Intel® Optimized LINPACK Benchmark

型号 核心代号 单精度(GFLOPS) 双精度(GFLOPS) 显存带宽(GB/s) 显存容量(GB) CUDA核心
GeForce RTX 3060 Laptop GPU GA106 13271 235.6 6 3840
Tesla K80 2 x Kepler GK210 5600 1870 480 24 4992
Tesla K40 1 x Kepler GK110B 4290 1430 288 12 2880
GeForce GTX TITAN X GM200 6144 192 336 12 3072
GeForce GTX 980 GM204 4612 144 224 4 2048
GeForce GTX 970 GM204 3494 109 224 4 1664
GeForce GTX 960 GM206 2308 72 112 2 5760
GeForce GTX TITAN Z 2 x GK110 8122 2707 672 12 5760
GeForce GTX Titan Black GK110 5121 1707 336 6 2880
GeForce GTX TITAN GK110 4500 1500 288 6 2688
GeForce GTX 780 Ti GK110 5046 210 336 3 2304
GeForce GTX 780 GK110 3977 166 288 3 2304
GeForce GTX 770 GK104 3213 134 224 2 1536
GeForce GTX 760 GK104 2258 94 192 2 1152
GeForce GTX 690 2 × GK104-355-A2 5622 234 384 4 3072
GeForce GTX 590 GF110 x 2 2460 1244 327 3 1024
GeForce GTX 580 GF110 1577 790 192 1.5 512
型号 核心代号 单精度(GFLOPS) 双精度(GFLOPS) 显存带宽(GB/s) 显存容量(GB) 流处理器
Radeon R9 295X2 Vesuvius 11466 1433 640 8 5632
Radeon R9 290X Hawaii XT 5632 704 352 8 2816
Radeon R9 290 Hawaii PRO 4848 606 320 4 2560
Radeon R9 280X Tahiti XT2 Tahiti XTL 3481 870 288 3 2048
Radeon R9 280 Tahiti PRO 3046 761 240 3 1792
Radeon R9 270X Curacao XT 2560 160 179 4 1280
Radeon R9 270 Curacao Pro 2304 144 179 2 1280
Radeon R7 260X Bonaire XTX 1971 123 104 2 896
Radeon R7 260 Bonaire 1536 96 96 2 768
Radeon HD 7990 New Zealand 8200 1894 288 x 2 6 4096
Radeon HD 7970 GHz Edition Tahiti XT2 4300 1075 288 3 1792
Radeon HD 7970 Tahiti XT 3788 947 264 3 1792
Radeon HD 7950 Boost Tahiti PRO2 Tahiti PRO-H2 3315 828 240 3 1792
Radeon HD 7950 Tahiti PRO Tahiti PRO-H 2867 717 240 3 1792
Radeon HD 7870 XT Tahiti LE 2995 748 192 2 1536
Radeon HD 7870 GHz Edition Pitcairn XT 2560 160 153 2 1280
Radeon HD 7850 Pitcairn PRO 1761 110 153 2 1024
Radeon HD 7790 Bonaire XT 1790 128 96 1 896
Radeon HD 7770 GHz Edition Cape Verde XT 1280 80 72 2 640

Linpack benchmark using the Intel MKL optimizations

型号 架构 单精度(GFLOPS) 双精度(GFLOPS)
Xeon(R) Gold 6244 x2 1149
Xeon(R) Gold 6144 x2 1100
Xeon(R) E5-2667v3 x2 515
Core i7 5960X Haswell E 354
Xeon E5 2687W 345
Core i7-10870H 320
Core i7 5930K Haswell E 289
Xeon E5 2650 262
Core(TM) i7-9700 230
Core(TM) i7-1165G7 203
Core i7 4770K Haswell 182
Xeon E3 1245v3 Haswell 170
Core i7 4960X Ivy Bridge-E 165
Core(TM) i5-6400 163
Core i7 4790 136
Core i5 3570 Ivy Bridge 105
Core i7 920 Nehalem Bloomfield 40
Core i5 3317u Ivy Bridge 33

https://www.netlib.org/benchmark/linpackc.new

型号 架构 单精度(GFLOPS) 双精度(GFLOPS)
Raspberry Pi 4 Model B Rev 1.1 0.962 / 2.0 (NEON) 0.867

Leave a Reply

Your email address will not be published.

Time limit is exhausted. Please reload the CAPTCHA.

Proudly powered by WordPress   Premium Style Theme by www.gopiplus.com