# SPEC CPU 2017 Rate 可通过[交互式图表](https://jia.je/benchmark/viewer.html)查看和筛选本文的测试数据，也可以访问[数据图表](https://jia.je/benchmark/spec-cpu-2017-rate-figures/index.md)查看静态图表。 ## SPEC INT 2017 Rate-1 下面贴出自己测的数据（SPECint2017，Estimated，rate，base，1 copy），不保证满足 SPEC 的要求，仅供参考。总运行时间（秒）基本和分数成反比，乘积按 5e4 估算。 ### 原始数据 #### Debian Trixie 桌面平台（LTO + Jemalloc）： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3 -flto -ljemalloc`）: [9.31](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_Ryzen_7_5700X_O3-flto-ljemalloc_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3 -flto -ljemalloc`）: [3.58](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M1_E-Core_O3-flto-ljemalloc_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3 -flto -ljemalloc`）: [9.14](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M1_P-Core_O3-flto-ljemalloc_001.txt) - Apple M2 E-Core @ 2.4 GHz Blizzard（`-O3 -flto -ljemalloc`）: [4.77](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M2_E-Core_O3-flto-ljemalloc_001.txt) - Apple M2 P-Core @ 3.5 GHz Avalanche（`-O3 -flto -ljemalloc`）: [10.6](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M2_P-Core_O3-flto-ljemalloc_001.txt) - Huawei Kirin X90 VM P-Core @ 2.3 GHz（`-O3 -flto -ljemalloc`）: [5.79](https://jia.je/benchmark/data-trixie/int2017_rate1/Huawei_Kirin_X90_VM_P-Core_O3-flto-ljemalloc_001.txt) - Intel Core i5-1135G7 @ 4.2 GHz Willow Cove（`-O3 -flto -ljemalloc`）: [7.28](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i5-1135G7_O3-flto-ljemalloc_001.txt) - Intel Core i7-13700K E-Core @ 4.2 GHz Gracemont（`-O3 -flto -ljemalloc`）: [7.43](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i7-13700K_E-Core_O3-flto-ljemalloc_001.txt) - Intel Core i7-13700K P-Core @ 5.4 GHz Raptor Cove（`-O3 -flto -ljemalloc`）: [10.9](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i7-13700K_P-Core_O3-flto-ljemalloc_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3 -flto -ljemalloc`）: [6.96](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-10980XE_O3-flto-ljemalloc_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3 -flto -ljemalloc`）: [6.63](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-12900KS_E-Core_O3-flto-ljemalloc_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3 -flto -ljemalloc`）: [10.6](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-12900KS_P-Core_O3-flto-ljemalloc_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3 -flto -ljemalloc`）: [7.90](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-14900K_E-Core_O3-flto-ljemalloc_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3 -flto -ljemalloc`）: [12.6](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3-flto-ljemalloc_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3 -flto -ljemalloc`）: [8.96](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_w9-3595X_O3-flto-ljemalloc_001.txt) - Loongson 3A6000 @ 2.5 GHz LA664（`-O3 -flto -ljemalloc`）: [4.86](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3-flto-ljemalloc_001.txt) 桌面平台（LTO）： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3 -flto`）: [8.57](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_Ryzen_7_5700X_O3-flto_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3 -flto`）: [3.34](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M1_E-Core_O3-flto_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3 -flto`）: [8.40](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M1_P-Core_O3-flto_001.txt) - Apple M2 E-Core @ 2.4 GHz Blizzard（`-O3 -flto`）: [4.45](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M2_E-Core_O3-flto_001.txt) - Apple M2 P-Core @ 3.5 GHz Avalanche（`-O3 -flto`）: [9.66](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M2_P-Core_O3-flto_001.txt) - Huawei Kirin X90 VM P-Core @ 2.3 GHz（`-O3 -flto`）: [5.32](https://jia.je/benchmark/data-trixie/int2017_rate1/Huawei_Kirin_X90_VM_P-Core_O3-flto_001.txt) - Intel Core i5-1135G7 @ 4.2 GHz Willow Cove（`-O3 -flto`）: [6.80](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i5-1135G7_O3-flto_001.txt) - Intel Core i7-13700K E-Core @ 4.2 GHz Gracemont（`-O3 -flto`）: [6.97](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i7-13700K_E-Core_O3-flto_001.txt) - Intel Core i7-13700K P-Core @ 5.4 GHz Raptor Cove（`-O3 -flto`）: [10.3](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i7-13700K_P-Core_O3-flto_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3 -flto`）: [6.57](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-10980XE_O3-flto_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3 -flto`）: [6.31](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-12900KS_E-Core_O3-flto_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3 -flto`）: [10.0](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-12900KS_P-Core_O3-flto_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3 -flto`）: [7.43](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-14900K_E-Core_O3-flto_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3 -flto`）: [12.1](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3-flto_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3 -flto`）: [8.60](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_w9-3595X_O3-flto_001.txt) - Loongson 3A6000 @ 2.5 GHz LA664（`-O3 -flto`）: [4.56](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3-flto_001.txt) 桌面平台： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3`）: [8.19](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_Ryzen_7_5700X_O3_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3`）: [3.20](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M1_E-Core_O3_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3`）: [7.97](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M1_P-Core_O3_001.txt) - Apple M2 E-Core @ 2.4 GHz Blizzard（`-O3`）: [4.25](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M2_E-Core_O3_001.txt) - Apple M2 P-Core @ 3.5 GHz Avalanche（`-O3`）: [9.21](https://jia.je/benchmark/data-trixie/int2017_rate1/Apple_M2_P-Core_O3_001.txt) - Huawei Kirin X90 VM P-Core @ 2.3 GHz（`-O3`）: [5.12](https://jia.je/benchmark/data-trixie/int2017_rate1/Huawei_Kirin_X90_VM_P-Core_O3_001.txt) - Intel Core i5-1135G7 @ 4.2 GHz Willow Cove（`-O3`）: [6.58](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i5-1135G7_O3_001.txt) - Intel Core i7-13700K E-Core @ 4.2 GHz Gracemont（`-O3`）: [6.72](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i7-13700K_E-Core_O3_001.txt) - Intel Core i7-13700K P-Core @ 5.4 GHz Raptor Cove（`-O3`）: [9.85](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i7-13700K_P-Core_O3_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3`）: [6.31](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-10980XE_O3_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3`）: [6.10](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-12900KS_E-Core_O3_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3`）: [9.74](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-12900KS_P-Core_O3_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3`）: [7.18](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-14900K_E-Core_O3_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3`）: [11.6](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3`）: [8.29](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_w9-3595X_O3_001.txt) - Loongson 3A6000 @ 2.5 GHz LA664（`-O3`）: [4.35](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3_001.txt) [4.39](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3_002.txt) 服务器平台（LTO + Jemalloc）： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3 -flto -ljemalloc`）: [3.49](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_7551_O3-flto-ljemalloc_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3 -flto -ljemalloc`）: [5.48](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_7742_O3-flto-ljemalloc_001.txt) - AMD EPYC 9R45 @ 4.5 GHz Zen 5（`-O3 -flto -ljemalloc`）: [10.3](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_9R45_O3-flto-ljemalloc_001.txt) - AMD EPYC 9T95 @ 3.7 GHz Zen 5c（`-O3 -flto -ljemalloc`）: [8.80](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_9T95_O3-flto-ljemalloc_001.txt) - AWS Graviton 5 @ 3.3 GHz Neoverse V3（`-O3 -flto -ljemalloc`）: [9.89](https://jia.je/benchmark/data-trixie/int2017_rate1/AWS_Graviton_5_O3-flto-ljemalloc_001.txt) - Google Axion C4A @ Neoverse V2（`-O3 -flto -ljemalloc`）: [8.23](https://jia.je/benchmark/data-trixie/int2017_rate1/Google_Axion_C4A_O3-flto-ljemalloc_001.txt) - Google Axion N4A @ Neoverse N3（`-O3 -flto -ljemalloc`）: [7.97](https://jia.je/benchmark/data-trixie/int2017_rate1/Google_Axion_N4A_O3-flto-ljemalloc_001.txt) - IBM POWER8 @ 3.2 GHz POWER8（`-O3 -flto -ljemalloc`）: [3.63](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER8_O3-flto-ljemalloc_001.txt) - IBM POWER9 3.2 GHz @ 3.2 GHz POWER9（`-O3 -flto -ljemalloc`）: [3.53](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER9_3.2_GHz_O3-flto-ljemalloc_001.txt) - IBM POWER9 3.8 GHz @ 3.8 GHz POWER9（`-O3 -flto -ljemalloc`）: [4.81](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER9_3.8_GHz_O3-flto-ljemalloc_001.txt) - Intel Xeon 6975P-C @ 3.9 GHz Redwood Cove（`-O3 -flto -ljemalloc`）: [8.03](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_6975P-C_O3-flto-ljemalloc_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3 -flto -ljemalloc`）: [4.95](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_E5-2680_v4_O3-flto-ljemalloc_001.txt) - Intel Xeon Gold 6430 @ 2.6 GHz Golden Cove（`-O3 -flto -ljemalloc`）: [5.39](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_Gold_6430_O3-flto-ljemalloc_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3 -flto -ljemalloc`）: [6.17](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_Platinum_8358P_O3-flto-ljemalloc_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3 -flto -ljemalloc`）: [3.65](https://jia.je/benchmark/data-trixie/int2017_rate1/Kunpeng_920_O3-flto-ljemalloc_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3 -flto -ljemalloc`）: [6.02](https://jia.je/benchmark/data-trixie/int2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3-flto-ljemalloc_001.txt) - Loongson 3C6000S @ 2.2 GHz LA664（`-O3 -flto -ljemalloc`）: [4.54](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3C6000S_O3-flto-ljemalloc_001.txt) 服务器平台（LTO）： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3 -flto`）: [3.28](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_7551_O3-flto_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3 -flto`）: [5.05](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_7742_O3-flto_001.txt) - AMD EPYC 9R45 @ 4.5 GHz Zen 5（`-O3 -flto`）: [9.49](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_9R45_O3-flto_001.txt) - AMD EPYC 9T95 @ 3.7 GHz Zen 5c（`-O3 -flto`）: [8.18](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_9T95_O3-flto_001.txt) - AWS Graviton 5 @ 3.3 GHz Neoverse V3（`-O3 -flto`）: [9.38](https://jia.je/benchmark/data-trixie/int2017_rate1/AWS_Graviton_5_O3-flto_001.txt) - Google Axion C4A @ Neoverse V2（`-O3 -flto`）: [7.68](https://jia.je/benchmark/data-trixie/int2017_rate1/Google_Axion_C4A_O3-flto_001.txt) - Google Axion N4A @ Neoverse N3（`-O3 -flto`）: [7.44](https://jia.je/benchmark/data-trixie/int2017_rate1/Google_Axion_N4A_O3-flto_001.txt) - IBM POWER8 @ 3.2 GHz POWER8（`-O3 -flto`）: [3.45](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER8_O3-flto_001.txt) - IBM POWER9 3.2 GHz @ 3.2 GHz POWER9（`-O3 -flto`）: [3.30](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER9_3.2_GHz_O3-flto_001.txt) - IBM POWER9 3.8 GHz @ 3.8 GHz POWER9（`-O3 -flto`）: [4.41](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER9_3.8_GHz_O3-flto_001.txt) - Intel Xeon 6975P-C @ 3.9 GHz Redwood Cove（`-O3 -flto`）: [7.65](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_6975P-C_O3-flto_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3 -flto`）: [4.59](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_E5-2680_v4_O3-flto_001.txt) - Intel Xeon Gold 6430 @ 2.6 GHz Golden Cove（`-O3 -flto`）: [5.16](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_Gold_6430_O3-flto_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3 -flto`）: [5.91](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_Platinum_8358P_O3-flto_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3 -flto`）: [3.32](https://jia.je/benchmark/data-trixie/int2017_rate1/Kunpeng_920_O3-flto_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3 -flto`）: [5.65](https://jia.je/benchmark/data-trixie/int2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3-flto_001.txt) - Loongson 3C6000S @ 2.2 GHz LA664（`-O3 -flto`）: [4.39](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3C6000S_O3-flto_001.txt) [4.37](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3C6000S_O3-flto_002.txt) 服务器平台： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3`）: [3.12](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_7551_O3_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3`）: [4.78](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_7742_O3_001.txt) - AMD EPYC 9R45 @ 4.5 GHz Zen 5（`-O3`）: [9.07](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_9R45_O3_001.txt) - AMD EPYC 9T95 @ 3.7 GHz Zen 5c（`-O3`）: [7.83](https://jia.je/benchmark/data-trixie/int2017_rate1/AMD_EPYC_9T95_O3_001.txt) - AWS Graviton 5 @ 3.3 GHz Neoverse V3（`-O3`）: [9.10](https://jia.je/benchmark/data-trixie/int2017_rate1/AWS_Graviton_5_O3_001.txt) - Google Axion C4A @ Neoverse V2（`-O3`）: [7.25](https://jia.je/benchmark/data-trixie/int2017_rate1/Google_Axion_C4A_O3_001.txt) - Google Axion N4A @ Neoverse N3（`-O3`）: [7.16](https://jia.je/benchmark/data-trixie/int2017_rate1/Google_Axion_N4A_O3_001.txt) - IBM POWER8 @ 3.2 GHz POWER8（`-O3`）: [3.24](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER8_O3_001.txt) - IBM POWER9 3.2 GHz @ 3.2 GHz POWER9（`-O3`）: [3.01](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER9_3.2_GHz_O3_001.txt) - IBM POWER9 3.8 GHz @ 3.8 GHz POWER9（`-O3`）: [4.14](https://jia.je/benchmark/data-trixie/int2017_rate1/IBM_POWER9_3.8_GHz_O3_001.txt) - Intel Xeon 6975P-C @ 3.9 GHz Redwood Cove（`-O3`）: [7.38](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_6975P-C_O3_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3`）: [4.39](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_E5-2680_v4_O3_001.txt) - Intel Xeon Gold 6430 @ 2.6 GHz Golden Cove（`-O3`）: [4.97](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_Gold_6430_O3_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3`）: [5.66](https://jia.je/benchmark/data-trixie/int2017_rate1/Intel_Xeon_Platinum_8358P_O3_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3`）: [3.17](https://jia.je/benchmark/data-trixie/int2017_rate1/Kunpeng_920_O3_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3`）: [5.43](https://jia.je/benchmark/data-trixie/int2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3_001.txt) - Loongson 3C5000 @ 2.2 GHz LA464（`-O3`）: [2.63](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3C5000_O3_001.txt) - Loongson 3C6000S @ 2.2 GHz LA664（`-O3`）: [4.19](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3C6000S_O3_001.txt) [4.14](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3C6000S_O3_002.txt) #### Debian Bookworm 桌面平台（LTO + Jemalloc）： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3 -flto -ljemalloc`）: [9.13](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_7_5700X_O3-flto-ljemalloc_001.txt) - AMD Ryzen 9 9950X @ 5.7 GHz Zen 5（`-O3 -flto -ljemalloc`）: [12.9](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_9_9950X_O3-flto-ljemalloc_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3 -flto -ljemalloc`）: [3.52](https://jia.je/benchmark/data-bookworm/int2017_rate1/Apple_M1_E-Core_O3-flto-ljemalloc_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3 -flto -ljemalloc`）: [8.93](https://jia.je/benchmark/data-bookworm/int2017_rate1/Apple_M1_P-Core_O3-flto-ljemalloc_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3 -flto -ljemalloc`）: [6.70](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-10980XE_O3-flto-ljemalloc_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3 -flto -ljemalloc`）: [10.7](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-12900KS_P-Core_O3-flto-ljemalloc_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3 -flto -ljemalloc`）: [12.1](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3-flto-ljemalloc_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3 -flto -ljemalloc`）: [8.71](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_w9-3595X_O3-flto-ljemalloc_001.txt) - Qualcomm X1E80100 @ 4.0 GHz X Elite（`-O3 -flto -ljemalloc`）: [9.25](https://jia.je/benchmark/data-bookworm/int2017_rate1/Qualcomm_X1E80100_O3-flto-ljemalloc_001.txt) 桌面平台（LTO）： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3 -flto`）: [8.44](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_7_5700X_O3-flto_001.txt) - AMD Ryzen 9 9950X @ 5.7 GHz Zen 5（`-O3 -flto`）: [11.7](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_9_9950X_O3-flto_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3 -flto`）: [3.29](https://jia.je/benchmark/data-bookworm/int2017_rate1/Apple_M1_E-Core_O3-flto_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3 -flto`）: [8.24](https://jia.je/benchmark/data-bookworm/int2017_rate1/Apple_M1_P-Core_O3-flto_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3 -flto`）: [6.37](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-10980XE_O3-flto_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3 -flto`）: [9.97](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-12900KS_P-Core_O3-flto_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3 -flto`）: [11.7](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3-flto_001.txt) [11.7](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3-flto_002.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3 -flto`）: [8.30](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_w9-3595X_O3-flto_001.txt) - Qualcomm X1E80100 @ 4.0 GHz X Elite（`-O3 -flto`）: [8.62](https://jia.je/benchmark/data-bookworm/int2017_rate1/Qualcomm_X1E80100_O3-flto_001.txt) 桌面平台： - AMD Ryzen 5 7500F @ 5.0 GHz Zen 4（`-O3`）: [9.51](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_5_7500F_O3_001.txt) - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3`）: [7.87](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_7_5700X_O3_001.txt) - AMD Ryzen 9 9950X @ 5.7 GHz Zen 5（`-O3`）: [11.2](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_9_9950X_O3_001.txt) [11.3](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_Ryzen_9_9950X_O3_002.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3`）: [3.15](https://jia.je/benchmark/data-bookworm/int2017_rate1/Apple_M1_E-Core_O3_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3`）: [7.85](https://jia.je/benchmark/data-bookworm/int2017_rate1/Apple_M1_P-Core_O3_001.txt) - Huawei Kirin X90 VM P-Core @ 2.3 GHz（`-O3`）: [4.07](https://jia.je/benchmark/data-bookworm/int2017_rate1/Huawei_Kirin_X90_VM_P-Core_O3_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3`）: [6.24](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-10980XE_O3_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3`）: [6.08](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-12900KS_E-Core_O3_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3`）: [9.62](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-12900KS_P-Core_O3_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3`）: [7.03](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-14900K_E-Core_O3_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3`）: [11.3](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Core_i9-14900K_P-Core_O3_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3`）: [8.05](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_w9-3595X_O3_001.txt) - Qualcomm 8cx Gen3 E-Core @ 2.4 GHz Cortex-A78C（`-O3`）: [4.11](https://jia.je/benchmark/data-bookworm/int2017_rate1/Qualcomm_8cx_Gen3_E-Core_O3_001.txt) - Qualcomm 8cx Gen3 P-Core @ 3.0 GHz Cortex-X1C（`-O3`）: [5.73](https://jia.je/benchmark/data-bookworm/int2017_rate1/Qualcomm_8cx_Gen3_P-Core_O3_001.txt) - Qualcomm X1E80100 @ 4.0 GHz X Elite（`-O3`）: [8.31](https://jia.je/benchmark/data-bookworm/int2017_rate1/Qualcomm_X1E80100_O3_001.txt) 服务器平台（LTO + Jemalloc）： - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3 -flto -ljemalloc`）: [5.33](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7742_O3-flto-ljemalloc_001.txt) - AMD EPYC 9754 @ 3.1 GHz Zen 4c（`-O3 -flto -ljemalloc`）: [5.79](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9754_O3-flto-ljemalloc_001.txt) - AMD EPYC 9755 @ 4.1 GHz Zen 5（`-O3 -flto -ljemalloc`）: [9.66](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9755_O3-flto-ljemalloc_001.txt) - AMD EPYC 9K65 @ 3.7 GHz Zen 5c（`-O3 -flto -ljemalloc`）: [8.19](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9K65_O3-flto-ljemalloc_001.txt) - AMD EPYC 9K85 @ 4.1 GHz Zen 5（`-O3 -flto -ljemalloc`）: [9.48](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9K85_O3-flto-ljemalloc_001.txt) - AMD EPYC 9R14 @ 3.7 GHz Zen 4（`-O3 -flto -ljemalloc`）: [7.21](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9R14_O3-flto-ljemalloc_001.txt) - AMD EPYC 9T24 @ 3.7 GHz Zen 4（`-O3 -flto -ljemalloc`）: [7.62](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9T24_O3-flto-ljemalloc_001.txt) - AWS Graviton 3 @ 2.6 GHz Neoverse V1（`-O3 -flto -ljemalloc`）: [5.24](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_3_O3-flto-ljemalloc_001.txt) - AWS Graviton 3E @ 2.6 GHz Neoverse V1（`-O3 -flto -ljemalloc`）: [6.17](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_3E_O3-flto-ljemalloc_001.txt) - AWS Graviton 4 @ 2.8 GHz Neoverse V2（`-O3 -flto -ljemalloc`）: [7.64](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3-flto-ljemalloc_001.txt) [7.41](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3-flto-ljemalloc_002.txt) - Hygon C86 7390（`-O3 -flto -ljemalloc`）: [3.29](https://jia.je/benchmark/data-bookworm/int2017_rate1/Hygon_C86_7390_O3-flto-ljemalloc_001.txt) - IBM POWER8NVL @ 4.0 GHz POWER8（`-O3 -flto -ljemalloc`）: [4.02](https://jia.je/benchmark/data-bookworm/int2017_rate1/IBM_POWER8NVL_O3-flto-ljemalloc_001.txt) - Intel Xeon 6981E Crestmont（`-O3 -flto -ljemalloc`）: [4.79](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_6981E_O3-flto-ljemalloc_001.txt) - Intel Xeon 6982P-C @ 3.6 GHz Redwood Cove（`-O3 -flto -ljemalloc`）: [7.22](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_6982P-C_O3-flto-ljemalloc_001.txt) - Intel Xeon Platinum 8581C @ 3.4 GHz Raptor Cove（`-O3 -flto -ljemalloc`）: [6.87](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_Platinum_8581C_O3-flto-ljemalloc_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3 -flto -ljemalloc`）: [3.57](https://jia.je/benchmark/data-bookworm/int2017_rate1/Kunpeng_920_O3-flto-ljemalloc_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3 -flto -ljemalloc`）: [6.03](https://jia.je/benchmark/data-bookworm/int2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3-flto-ljemalloc_001.txt) 服务器平台（LTO）： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3 -flto`）: [3.19](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7551_O3-flto_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3 -flto`）: [5.02](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7742_O3-flto_001.txt) - AMD EPYC 9754 @ 3.1 GHz Zen 4c（`-O3 -flto`）: [5.48](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9754_O3-flto_001.txt) - AMD EPYC 9755 @ 4.1 GHz Zen 5（`-O3 -flto`）: [8.97](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9755_O3-flto_001.txt) - AMD EPYC 9K65 @ 3.7 GHz Zen 5c（`-O3 -flto`）: [7.78](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9K65_O3-flto_001.txt) - AMD EPYC 9K85 @ 4.1 GHz Zen 5（`-O3 -flto`）: [8.83](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9K85_O3-flto_001.txt) - AMD EPYC 9R14 @ 3.7 GHz Zen 4（`-O3 -flto`）: [6.62](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9R14_O3-flto_001.txt) - AMD EPYC 9T24 @ 3.7 GHz Zen 4（`-O3 -flto`）: [7.14](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9T24_O3-flto_001.txt) - AWS Graviton 3 @ 2.6 GHz Neoverse V1（`-O3 -flto`）: [5.68](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_3_O3-flto_001.txt) - AWS Graviton 4 @ 2.8 GHz Neoverse V2（`-O3 -flto`）: [7.14](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3-flto_001.txt) [6.53](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3-flto_002.txt) [6.51](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3-flto_003.txt) - Hygon C86 7390（`-O3 -flto`）: [3.09](https://jia.je/benchmark/data-bookworm/int2017_rate1/Hygon_C86_7390_O3-flto_001.txt) - Intel Xeon 6981E Crestmont（`-O3 -flto`）: [4.62](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_6981E_O3-flto_001.txt) - Intel Xeon 6982P-C @ 3.6 GHz Redwood Cove（`-O3 -flto`）: [6.90](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_6982P-C_O3-flto_001.txt) - Intel Xeon Platinum 8581C @ 3.4 GHz Raptor Cove（`-O3 -flto`）: [6.67](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_Platinum_8581C_O3-flto_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3 -flto`）: [3.26](https://jia.je/benchmark/data-bookworm/int2017_rate1/Kunpeng_920_O3-flto_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3 -flto`）: [5.71](https://jia.je/benchmark/data-bookworm/int2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3-flto_001.txt) 服务器平台： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3`）: [3.07](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7551_O3_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3`）: [4.73](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7742_O3_001.txt) - AMD EPYC 7H12 @ 3.3 GHz Zen 2（`-O3`）: [4.23](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7H12_O3_001.txt) - AMD EPYC 7K83 Zen 3（`-O3`）: [5.18](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_7K83_O3_001.txt) - AMD EPYC 9754 @ 3.1 GHz Zen 4c（`-O3`）: [5.33](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9754_O3_001.txt) - AMD EPYC 9755 @ 4.1 GHz Zen 5（`-O3`）: [8.57](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9755_O3_001.txt) - AMD EPYC 9K65 @ 3.7 GHz Zen 5c（`-O3`）: [7.47](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9K65_O3_001.txt) - AMD EPYC 9K85 @ 4.1 GHz Zen 5（`-O3`）: [8.44](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9K85_O3_001.txt) - AMD EPYC 9R14 @ 3.7 GHz Zen 4（`-O3`）: [6.57](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9R14_O3_001.txt) [6.41](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9R14_O3_002.txt) - AMD EPYC 9T24 @ 3.7 GHz Zen 4（`-O3`）: [6.94](https://jia.je/benchmark/data-bookworm/int2017_rate1/AMD_EPYC_9T24_O3_001.txt) - AWS Graviton 3 @ 2.6 GHz Neoverse V1（`-O3`）: [5.43](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_3_O3_001.txt) - AWS Graviton 3E @ 2.6 GHz Neoverse V1（`-O3`）: [5.53](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_3E_O3_001.txt) - AWS Graviton 4 @ 2.8 GHz Neoverse V2（`-O3`）: [7.00](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3_001.txt) [6.85](https://jia.je/benchmark/data-bookworm/int2017_rate1/AWS_Graviton_4_O3_002.txt) - Ampere Altra @ 3.0 GHz Neoverse N1（`-O3`）: [4.41](https://jia.je/benchmark/data-bookworm/int2017_rate1/Ampere_Altra_O3_001.txt) - Hygon C86 7390（`-O3`）: [2.97](https://jia.je/benchmark/data-bookworm/int2017_rate1/Hygon_C86_7390_O3_001.txt) - IBM POWER8NVL @ 4.0 GHz POWER8（`-O3`）: [3.54](https://jia.je/benchmark/data-bookworm/int2017_rate1/IBM_POWER8NVL_O3_001.txt) - Intel Xeon 6981E Crestmont（`-O3`）: [4.48](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_6981E_O3_001.txt) - Intel Xeon 6982P-C @ 3.6 GHz Redwood Cove（`-O3`）: [6.68](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_6982P-C_O3_001.txt) - Intel Xeon D-2146NT @ 2.9 GHz Skylake（`-O3`）: [3.96](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_D-2146NT_O3_001.txt) - Intel Xeon E5-2603 v4 @ 1.7 GHz Broadwell（`-O3`）: [2.48](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_E5-2603_v4_O3_001.txt) - Intel Xeon E5-2680 v3 @ 3.3 GHz Haswell（`-O3`）: [4.01](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_E5-2680_v3_O3_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3`）: [4.35](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_E5-2680_v4_O3_001.txt) - Intel Xeon E5-4610 v2 @ 2.7 GHz Ivy Bridge EP（`-O3`）: [3.06](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_E5-4610_v2_O3_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3`）: [5.66](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_Platinum_8358P_O3_001.txt) - Intel Xeon Platinum 8576C Raptor Cove（`-O3`）: [5.72](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_Platinum_8576C_O3_001.txt) - Intel Xeon Platinum 8581C @ 3.4 GHz Raptor Cove（`-O3`）: [6.52](https://jia.je/benchmark/data-bookworm/int2017_rate1/Intel_Xeon_Platinum_8581C_O3_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3`）: [3.10](https://jia.je/benchmark/data-bookworm/int2017_rate1/Kunpeng_920_O3_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3`）: [5.53](https://jia.je/benchmark/data-bookworm/int2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3_001.txt) - T-Head Yitian 710 @ 3.0 GHz Neoverse N2（`-O3`）: [5.79](https://jia.je/benchmark/data-bookworm/int2017_rate1/T-Head_Yitian_710_O3_001.txt) #### HarmonyOS 桌面平台（LTO）： - Huawei Kirin X90 E-Core @ 2.0 GHz（`-O3 -flto`）: [4.28](https://jia.je/benchmark/data-harmonyos/int2017_rate1/Huawei_Kirin_X90_E-Core_O3-flto_001.txt) - Huawei Kirin X90 P-Core @ 2.3 GHz（`-O3 -flto`）: [4.87](https://jia.je/benchmark/data-harmonyos/int2017_rate1/Huawei_Kirin_X90_P-Core_O3-flto_001.txt) 手机平台（LTO）： - Huawei Kirin 9010 E-Core Full @ 2.2 GHz（`-O3 -flto`）: [3.21](https://jia.je/benchmark/data-harmonyos/int2017_rate1/Huawei_Kirin_9010_E-Core_Full_O3-flto_001.txt) - Huawei Kirin 9010 P-Core Best @ 2.3 GHz（`-O3 -flto`）: [4.18](https://jia.je/benchmark/data-harmonyos/int2017_rate1/Huawei_Kirin_9010_P-Core_Best_O3-flto_001.txt) - Huawei Kirin 9010 P-Core Full @ 2.3 GHz（`-O3 -flto`）: [3.96](https://jia.je/benchmark/data-harmonyos/int2017_rate1/Huawei_Kirin_9010_P-Core_Full_O3-flto_001.txt) #### 备注 1. SPEC INT 2017 Rate-1 结果受 `-flto`（分数 +4%，主要优化 mcf/deepsjeng）和 `-ljemalloc`（分数 +4-10%，主要优化 omnetpp/xalancbmk）影响很明显。`-Ofast` 和 `-O3` 区别很小，`-march=native` 影响很小。 1. 在部分处理器上，Linux 不能保证程序被调度到性能最高的核心上，例如： 1. Qualcomm X1E80100 上，负载不一定会调度到有 Boost 的核上，因此需要手动绑核。没有 Boost 的核心会跑在 3.4 GHz，Boost 的核心最高可以达到 4.0 GHz，对应 14% 的性能提升。具体地讲，它有三个 Cluster，0-3 是没有 Boost 的 Cluster，4-7 和 8-11 每个 Cluster 中可以有一个核心 Boost 到 4.0 GHz，也就是说，最多有两个核达到 4.0 GHz，这两个核需要分别位于 4-7 和 8-11 两个 Cluster 当中。如果一个 Cluster 有两个或者以上的核有负载，那么他们都只有 3.4 GHz。 1. AMD Ryzen 9 9950X 不同核能够达到的最大频率不同，目前 Linux（6.11）的调度算法不一定可以保证跑到最大频率 5.75 GHz 上，可能会飘到频率低一些（5.45 GHz 左右）的核心上，损失 4% 的性能，因此需要绑核心，详见 [Linux 大小核的调度算法探究](https://jia.je/software/2024/10/23/linux-core-scheduling/index.md) 以及 [谈谈 Linux 与 ITMT 调度器与多簇处理器](https://blog.hjc.im/thoughts-on-linux-preferred-cores-and-multi-ccx.html)。这个问题已经有 Patch 进行修复。 1. 对于服务器 CPU，默认设置可能没有打开 C6 State，此时单核不一定能 Boost 到宣称的最高频率，需要进 BIOS 打开 C6 State，使得空闲的核心进入低功耗模式，才能发挥出最高的 Boost 频率。 1. 对于除了苹果以外的 ARM64 核心，内核的 branch-misses 计数器考虑了 speculative 而不只是 retired，因此数字会偏高，此时要用 r22 计数替代。 1. Google Cloud 只有部分机型（如 C4 和 C4A）支持 PMU，并且需要手动开启（参考 [Enable the PMU in VMs](https://cloud.google.com/compute/docs/enable-pmu-in-vms)）： ```shell $ gcloud compute instances export VM_NAME \ --destination=YAML_FILE \ --zone=ZONE $ vim YAML_FILE # append the following lines advancedMachineFeatures: performanceMonitoringUnit: STANDARD $ gcloud compute instances update-from-file VM_NAME \ --most-disruptive-allowed-action=RESTART \ --source=YAML_FILE \ --zone=ZONE ``` 6. 华为云的 kc2 实例的 PMU 默认不开启，需要通过工单申请，步骤如下： 1. 创建一个私有镜像（可以先用公共镜像起一个虚拟机，再从虚拟机创建私有镜像） 1. 创建工单申请，申请给私有镜像启用 PMU 1. 用私有镜像创建新的虚拟机，在这个虚拟机内就可以使用 PMU： ```shell root@kc2 ~# dmesg | grep PMU [ 1.196145] hw perfevents: enabled with armv8_pmuv3_0 PMU driver, 9 counters available ``` 7. Kirin 9010 因为散热问题，单独跑测试，和顺着跑一遍测试，结果差距比较大。因此提供了两组数据：Best（每一项单独跑，取最短时间，散热影响比较小）和 Full（按照顺序跑一次，散热影响比较大）。 ### 分支预测器比较 #### Debian Trixie x86 平台的分支预测准确率（Average）由高到低（`-O3`）： 1. Zen 5(AMD 9R45): MPKI=4.52 1. Zen 5c(AMD 9T95): MPKI=4.52 1. Zen 3(AMD 5700X): MPKI=4.75 1. Zen 2(AMD 7742): MPKI=4.80 1. Redwood Cove(Intel 6975P-C): MPKI=4.81 1. Golden Cove(Intel 12900KS P-Core)/Raptor Cove(Intel 13700K P-Core/Intel 14900K P-Core): MPKI=4.92 1. Willow Cove(Intel i5-1135G7): MPKI=4.96 1. Gracemont(Intel 12900KS E-Core/Intel 13700K E-Core/Intel 14900K E-Core): MPKI=5.19 1. Cascade Lake(Intel 10980XE): MPKI=5.51 1. Zen 1(AMD 7551): MPKI=5.90 1. Broadwell(Intel E5-2680 v4): MPKI=6.05 ARM64 平台的分支预测准确率（Average）由高到低（`-O3`）： 1. Neoverse V3(AWS Graviton 5): MPKI=4.33 1. Neoverse V2(Google Axion C4A): MPKI=4.59 1. Avalanche(Apple M2 P-Core): MPKI=4.62 1. Firestorm(Apple M1 P-Core): MPKI=4.89 1. Icestorm(Apple M1 E-Core): MPKI=5.47 LoongArch64 平台的分支预测准确率（Average）由高到低（`-O3`）： 1. LA664(3A6000/3C6000): MPKI=5.01 1. LA464(3C5000): MPKI=8.39 #### Debian Bookworm x86 平台的分支预测准确率（Average）由高到低（`-O3`）： 1. Zen 5(AMD 9950X/AMD 9755/AMD 9K85): MPKI=4.48 Mispred=2.52% 1. Zen 5c(AMD 9K65): MPKI=4.51 Mispred=2.54% 1. Zen 4(AMD 9T24/9R14): MPKI=4.57 Mispred=2.57% 1. Zen 4c(AMD 9754): MPKI=4.66 Mispred=2.63% 1. Zen 4(AMD 7500F): MPKI=4.68 Mispred=2.64% 1. Zen 3(AMD 5700X): MPKI=4.68 Mispred=2.64% 1. Zen 2(AMD 7742): MPKI=4.77 Mispred=2.69% 1. Redwood Cove(Intel 6982P-C): MPKI=4.77 Mispred=2.71% 1. Sunny Cove(Intel 8358P)/Golden Cove(Intel 12900KS P-Core)/Raptor Cove(Intel 14900K P-Core/Intel 8581C): MPKI=4.86 Mispred=2.75% 1. Gracemont(Intel 12900KS P-Core/Intel 14900K P-Core): MPKI=5.15 Mispred=2.92% 1. Skylake(Intel D-2146NT)/Cascade Lake(Intel 10980XE): MPKI=5.50 Mispred=3.13% 1. Zen 1(AMD 7551): MPKI=5.82 Mispred=3.31% 1. Haswell(Intel E5-2680 v3)/Broadwell(Intel E5-2680 v4): MPKI=5.98 Mispred=3.34% ARM64 平台的分支预测准确率（Average）由高到低（`-O3`）： 1. Neoverse V2(AWS Graviton 4): MPKI=4.50 Mispred=2.47% 1. Oryon(Qualcomm X1E80100): MPKI=4.71 Mispred=2.58% 1. Neoverse N2(Aliyun Yitian 710): MPKI=4.80 Mispred=2.64% 1. Firestorm(Apple M1 P-Core): MPKI=4.81 Mispred=2.62% 1. Neoverse V1(AWS Graviton 3/AWS Graviton 3E)/Cortex X1C(Qualcomm 8cx Gen3): MPKI=4.91 Mispred=2.69% 1. HuaweiCloud kc2: MPKI=5.17 Mispred=2.85% 1. Neoverse N1(Ampere Altra)/Cortex A78C(Qualcomm 8cx Gen3 E-Core): MPKI=5.21 Mispred=2.87% 1. Icestorm(Apple M1 E-Core): MPKI=5.41 Mispred=2.99% 1. TSV110(Hisilicon Kunpeng 920): MPKI=6.54 Mispred=3.58% ### 网上的数据 [SPEC CPU 2017 by David Huang](https://blog.hjc.im/spec-cpu-2017): - Apple M4 Pro: 13.7 - AMD Ryzen 9950X Zen 5: 12.6 - Apple M3 Pro: 11.8 - Intel Core 13900K Raptor Cove: 11.5 - Intel Core Ultra 7 265K Arrow Lake Lion Cove+Skymont: 11.1 - AMD AI Max+ 395 Zen 5: 10.6 - Apple M2 Pro: 10.3 - Apple M2: 9.95 - AMD HX 370 Strix Point Zen 5: 9.64 - Intel Core Ultra 258V Lunar Lake Lion Cove+Skymont: 9.46 - Apple M1 Max Firestorm: 9.2 - AMD Ryzen 5950X Zen 3: 9.15 - Kunpeng 920 TSV120: 6.00 - Loongson 3A6000 LA664: 4.29 - Phytium D3000 FTC862: 4.24 - Loongson 3A5000 LA464: 3.04 [高通 X Elite Oryon 微架构评测：走走停停 by JamesAslan](https://zhuanlan.zhihu.com/p/704707254): - AMD Ryzen 7700X Zen 4: 10.35 - Intel Core 13700K Raptor Cove: 9.81 - Intel Core 12700K Golden Cove: 9.13 - AMD Ryzen 5950X Zen 3: 8.45 - Apple M2 Avalanche+Blizzard: 8.40 - Qualcomm X1E80100 Oryon: 8.19 - Apple M1 Firestorm+Icestorm: 7.40 - Qualcomm 8 Gen 2 Cortex-X3: 6.58 [Running SPEC CPU2017 on Chinese CPUs, and More](https://old.chipsandcheese.com/2024/10/18/running-spec-cpu2017-on-chinese-cpus-and-more/) - AMD Ryzen 9 7950X3D Non-VCache: 10.5 - Intel Core Ultra 7 258V Lion Cove: 9.37 - Intel Core Ultra 7 115H Redwood Cove: 7.6 - Intel Core Ultra 7 258V Skymont: 5.92 - Intel Core Ultra 7 115H Crestmont: 5.88 - Intel Core i5-6600K Skylake: 5.65 - Loongson 3A6000 LA664: 4.27 - Mediatek Genio 1200 Cortex A78: 3.8 - AMD FX-8150: 3.5 - Intel Core Ultra 7 115H Low Power Crestmont: 3.32 - Loongson 3A5000: 2.93 - Intel Celeron J4125 Goldmont Plus: 2.43 - Zhaoxin KaiXian KX-6640MA: 2.07 - Amlogic S922X Cortex A73: 1.77 - Mediatek Genio 1200 Cortex A55: 1.19 [Running SPEC CPU2017 at Chips and Cheese?](https://old.chipsandcheese.com/2024/09/19/running-spec-cpu2017-at-chips-and-cheese/) - AMD Ryzen 9 9950X: 11.9 - AMD Ryzen 9 7950X3D Non-VCache: 10.5 - AMD Ryzen 9 7950X3D VCache: 10.5 - Intel Core Ultra 7 115H Redwood Cove: 7.58 - Intel Core Ultra 7 115H Crestmont: 5.34 - AMD Ryzen 9 3950X 3.5GHz: 5.28 - Ampere Altra: 3.98 - AmpereOne: 3.94 - AMD FX-8159: 3.46 [The AMD Ryzen 9 9950X and Ryzen 9 9900X Review: Flagship Zen 5 Soars - and Stalls](https://www.anandtech.com/show/21524/the-amd-ryzen-9-9950x-and-ryzen-9-9900x-review/5) - AMD Ryzen 9 9950X Zen 5: 10.95 - Intel Core i9-14900K Raptor Cove: 10.94 - AMD Ryzen 9 7950X Zen 4: 9.88 [Snapdragon X Elite Qualcomm Oryon CPU Design and Architecture Hot Chips 2024](https://www.servethehome.com/snapdragon-x-elite-qualcomm-oryon-cpu-design-and-architecture-hot-chips-2024-arm/) - Qualcomm X Elite Oryon on Linux: 10.64 - Qualcomm X Elite Oryon on Windows: 9.70 [ARM Cortex X1 微架构评测（上）：向山进发](https://zhuanlan.zhihu.com/p/619033328) - Zen 3 @ 4.95 GHz: 8.4 - Firestorm @ 3.0 GHz: 7.4 - Cortex X1 @ 3.0 GHz: 5.7 - Cortex A78 @ 2.4 GHz: 3.9 [极客湾•麒麟 9010，测评汇总：IPC 性能，巨幅提升！CPU 能效全频段领先，麒麟 9000S！](https://www.bilibili.com/opus/925502754370093090) - Huawei Kirin 9010: 4.54 - HUawei Kirin 9000s: 4.06 ### 多版本 GCC 和 LLVM 性能比较 在 Intel i9-14900K @ 5.7 GHz 上用 -O3 测试几种编译器组合的 SPEC INT 2017 Rate-1 性能： | Benchmark | GCC 15 | GCC 14 | GCC 13 | GCC 12 | GCC 11 | LLVM 19 | LLVM 18 | LLVM 17 | LLVM 20 | LLVM 20 w/ -fwrapv | | --------------- | -------- | -------- | ------ | -------- | -------- | -------- | ------- | ------- | ------- | ------------------ | | 500.perlbench_r | 12.0 | 11.8 | 12.0 | **12.3** | **12.3** | 10.9 | 10.9 | 10.8 | 10.8 | 10.9 | | 502.gcc_r | **14.0** | 13.7 | 13.7 | 13.6 | 13.5 | 13.5 | 13.5 | 13.5 | 13.5 | 13.5 | | 505.mcf_r | 9.34 | **9.48** | 9.19 | 9.32 | 9.38 | 8.32 | 8.40 | 8.76 | 8.27 | 8.26 | | 520.omnetpp_r | **9.39** | 9.07 | 8.87 | 9.17 | 9.17 | 8.78 | 8.80 | 8.77 | 8.74 | 8.73 | | 523.xalancbmk_r | 8.91 | 8.91 | 8.95 | 8.85 | **9.11** | 8.88 | 8.86 | 8.85 | 8.86 | 8.83 | | 525.x264_r | **23.7** | 19.7 | 18.6 | 18.5 | 19.4 | 19.5 | 18.8 | 18.5 | 19.9 | 20.0 | | 531.deepsjeng_r | **7.43** | 7.36 | 7.36 | 7.24 | 6.95 | 7.18 | 7.27 | 7.22 | 7.17 | 7.29 | | 541.leela_r | 7.20 | 7.06 | 7.13 | 7.16 | 7.00 | **7.45** | 7.39 | 7.36 | 7.41 | 7.19 | | 548.exchange2_r | **32.5** | 29.9 | 28.8 | 28.2 | 16.2 | 14.4 | 14.4 | 12.9 | 10.9 | 14.4 | | 557.xz_r | 5.69 | 5.62 | 5.59 | 5.62 | 5.55 | **5.71** | 5.69 | 5.70 | 5.69 | 5.66 | | geomean | **11.2** | 10.8 | 10.7 | 10.7 | 10.1 | 9.80 | 9.78 | 9.68 | 9.52 | 9.78 | 完整数据： - [GCC 11.3.0](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_GCC_11.txt) - [GCC 12.2.0](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_GCC_12.txt) - [GCC 13.3.0](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_GCC_13.txt) - [GCC 14.2.0](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_GCC_14.txt) - [GCC 15.1.0](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_GCC_15.txt) - [LLVM 19.1.4](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_LLVM_19.txt) - [LLVM 18.1.8](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_LLVM_18.txt) - [LLVM 17.0.6](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_LLVM_17.txt) - [LLVM 20.1.5](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_LLVM_20.txt) LLVM 20 的 548.exchange2_r 性能下降可以通过添加 `-fwrapv` 选项来解决，见 [548.exchange2_r of SPEC CPU 2017 has 30% performance regression between LLVM 18/19 and LLVM 20 on amd64](https://github.com/llvm/llvm-project/issues/139274)： - [LLVM 20.1.5 with -fwrapv](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Intel_i9-14900K_O3_LLVM_20_fwrapv.txt) 注：GCC 指 GCC + GFortran，LLVM 指 Clang + Flang-new ### LA664 不同编译器版本和编译选项下的测试结果 鉴于网上针对 LA664 的 SPEC INT 2017 Rate-1 性能测试有一些争议： - [龙芯 3A6000、华为鲲鹏 920B 与 Intel 各代 CPU GCC14 Spec 2017 性能比对评测](https://zhuanlan.zhihu.com/p/711617301) - [开源软件环境下龙芯 3A6000 的性能](https://zhuanlan.zhihu.com/p/7264671348) - [是什么原因导致 guee 测试 3C6000 同编译参数同编译器下会有两份差异较大的 Spec2017 测试报告？](https://www.zhihu.com/question/9063557412) 小结一下上面的文章里的结果： - [3A6000 GCC 15.0.0 -O3 -march=native -flto by guee: 5.11, 2.04/GHz]() - [3A6000 GCC 14.0.1 -Ofast -march=native -flto -ljemalloc by Matterhorn: 4.73, 1.89/GHz](https://gitee.com/matter2024/CPU/blob/master/Spec2017/Ofast%2Bflto%2Bnative%2Bjemalloc/3A6000%E6%96%B0%E4%B8%96%E7%95%8C/CPU2017.008.intrate.txt) - [3A6000 GCC 14.0.1 -Ofast -march=native -flto by Matterhorn: 4.69, 1.88/GHz](https://gitee.com/matter2024/CPU/blob/master/Spec2017/Ofast%2Bflto%2Bnative/3A6000%E6%96%B0%E4%B8%96%E7%95%8C/CPU2017.003.intrate.txt) - [3A6000 GCC 14.0.1 -O3 -msimd=lasx by Matterhorn: 4.50, 1.80/GHz](https://gitee.com/matter2024/CPU/blob/master/Spec2017/O3/3A6000%E6%96%B0%E4%B8%96%E7%95%8C/CPU2017.004.intrate.txt) - [3A6000 GCC 14.0.1 -O3 by Matterhorn: 4.17, 1.67/GHz](https://gitee.com/matter2024/CPU/blob/master/Spec2017/O3/3A6000%E6%96%B0%E4%B8%96%E7%95%8C/CPU2017.010.intrate.txt) 可见主要的分歧是在 GCC 版本和编译选项上。 下面贴出本人测试的结果： - [3C6000 GCC 15.1.0 -O3 -msimd=lasx -flto -ljemalloc: 4.92, 2.24/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3-msimd%3Dlasx-flto-ljemalloc.txt) - [3C6000 GCC 15.1.0 -O3 -flto -ljemalloc: 4.90, 2.23/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3-flto-ljemalloc.txt) - [3A6000 GCC 14.2.0 -O3 -flto -ljemalloc: 4.86, 1.94/GHz](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3-flto-ljemalloc_001.txt) - [3C6000 GCC 15.1.0 -O3 -march=native -flto -ljemalloc: 4.82, 2.20/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3-march%3Dnative-flto-ljemalloc.txt) - [3C6000 GCC 15.1.0 -O3 -flto: 4.67, 2.12/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3-flto.txt) - [3A6000 GCC 14.2.0 -O3 -flto: 4.56, 1.82/GHz](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3-flto_001.txt) - [3C6000 GCC 14.2.0 -O3 -flto -ljemalloc: 4.54, 2.06/GHz](https://jia.je/benchmark/spec-cpu-2017-rate/data-trixie/int2017_rate1/Loongson_3C6000_O3-flto-ljemalloc_001.txt) - [3C6000 GCC 15.1.0 -O3: 4.49, 2.04/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3.txt) - [3C6000 GCC 15.1.0 -O3 -march=la464: 4.49, 2.04/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3-march%3Dla464.txt) - [3C6000 GCC 15.1.0 -O3 -march=la664: 4.40, 2.04/GHz](https://jia.je/benchmark/data-bookworm/others/SPEC_INT_2017_Loongson_3C6000_O3_GCC_15_O3-march%3Dla664.txt) - [3C6000 GCC 14.2.0 -O3 -flto: 4.39, 2.00/GHz](https://jia.je/benchmark/spec-cpu-2017-rate/data-trixie/int2017_rate1/Loongson_3C6000_O3-flto_001.txt) - [3A6000 GCC 14.2.0 -O3: 4.35, 1.74/GHz](https://jia.je/benchmark/data-trixie/int2017_rate1/Loongson_3A6000_O3_001.txt) - [3C6000 GCC 14.2.0 -O3: 4.19, 1.90/GHz](https://jia.je/benchmark/spec-cpu-2017-rate/data-trixie/int2017_rate1/Loongson_3C6000_O3_001.txt) 注：3A6000 频率是 2.5 GHz，3C6000 频率是 2.2 GHz。 结论：性能受编译器版本和编译选项影响很大，如果对不上，那么性能的差距可能会影响和其他处理器比较的结论。在上面的例子里，这些编译器版本和编译选项带来的优化： 1. `-flto`：约 5% 提升，`4.39/4.19=1.05`, `4.56/4.35=1.05` 1. `-march=native`（仅 GCC 14）或 `-msimd=lasx`: 约 8% 提升，`4.50/4.17=1.08` 1. GCC 15.1.0 vs GCC 14.2.0: 约 7% 提升，`4.49/4.19=1.07` 1. `-ljemalloc`: 约 3-7% 提升，`4.90/4.63=1.06`, `4.86/4.56=1.07`, `4.54/4.39=1.03` ## SPEC FP 2017 Rate-1 下面贴出自己测的数据（SPECfp2017，Estimated，rate，base，1 copy），不保证满足 SPEC 的要求，仅供参考。总运行时间基本和分数成反比，乘积按 1e5 估算。 ### 原始数据 #### Debian Trixie 桌面平台（`-march=native`）： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3 -march=native`）: [11.7](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_Ryzen_7_5700X_O3-march%3Dnative_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3 -march=native`）: [3.93](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M1_E-Core_O3-march%3Dnative_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3 -march=native`）: [12.2](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M1_P-Core_O3-march%3Dnative_001.txt) - Apple M2 E-Core @ 2.4 GHz Blizzard（`-O3 -march=native`）: [5.13](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M2_E-Core_O3-march%3Dnative_001.txt) - Apple M2 P-Core @ 3.5 GHz Avalanche（`-O3 -march=native`）: [13.3](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M2_P-Core_O3-march%3Dnative_001.txt) - Huawei Kirin X90 VM P-Core @ 2.3 GHz（`-O3 -march=native`）: [8.32](https://jia.je/benchmark/data-trixie/fp2017_rate1/Huawei_Kirin_X90_VM_P-Core_O3-march%3Dnative_001.txt) - Intel Core i5-1135G7 @ 4.2 GHz Willow Cove（`-O3 -march=native`）: [9.93](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i5-1135G7_O3-march%3Dnative_001.txt) - Intel Core i7-13700K E-Core @ 4.2 GHz Gracemont（`-O3 -march=native`）: [7.22](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i7-13700K_E-Core_O3-march%3Dnative_001.txt) - Intel Core i7-13700K P-Core @ 5.4 GHz Raptor Cove（`-O3 -march=native`）: [15.0](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i7-13700K_P-Core_O3-march%3Dnative_001.txt) - Intel Core i9-10980XE @ 4.7 GHz (AVX-512 @ 4.0 GHz) Cascade Lake（`-O3 -march=native`）: [7.85](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-10980XE_O3-march%3Dnative_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3 -march=native`）: [7.23](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-12900KS_E-Core_O3-march%3Dnative_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3 -march=native`）: [15.4](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-12900KS_P-Core_O3-march%3Dnative_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3 -march=native`）: [7.70](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-14900K_E-Core_O3-march%3Dnative_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3 -march=native`）: [18.0](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-14900K_P-Core_O3-march%3Dnative_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3 -march=native`）: [11.9](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_w9-3595X_O3-march%3Dnative_001.txt) - Loongson 3A6000 @ 2.5 GHz LA664（`-O3 -march=native`）: [5.73](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3A6000_O3-march%3Dnative_001.txt) 桌面平台： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3`）: [10.9](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_Ryzen_7_5700X_O3_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3`）: [3.93](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M1_E-Core_O3_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3`）: [12.2](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M1_P-Core_O3_001.txt) - Apple M2 E-Core @ 2.4 GHz Blizzard（`-O3`）: [5.13](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M2_E-Core_O3_001.txt) - Apple M2 P-Core @ 3.5 GHz Avalanche（`-O3`）: [13.3](https://jia.je/benchmark/data-trixie/fp2017_rate1/Apple_M2_P-Core_O3_001.txt) - Huawei Kirin X90 VM P-Core @ 2.3 GHz（`-O3`）: [8.28](https://jia.je/benchmark/data-trixie/fp2017_rate1/Huawei_Kirin_X90_VM_P-Core_O3_001.txt) - Intel Core i5-1135G7 @ 4.2 GHz Willow Cove（`-O3`）: [9.04](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i5-1135G7_O3_001.txt) - Intel Core i7-13700K E-Core @ 4.2 GHz Gracemont（`-O3`）: [6.93](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i7-13700K_E-Core_O3_001.txt) - Intel Core i7-13700K P-Core @ 5.4 GHz Raptor Cove（`-O3`）: [14.0](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i7-13700K_P-Core_O3_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3`）: [7.24](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-10980XE_O3_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3`）: [6.97](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-12900KS_E-Core_O3_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3`）: [14.4](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-12900KS_P-Core_O3_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3`）: [7.42](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-14900K_E-Core_O3_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3`）: [16.8](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Core_i9-14900K_P-Core_O3_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3`）: [11.0](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_w9-3595X_O3_001.txt) - Loongson 3A6000 @ 2.5 GHz LA664（`-O3`）: [5.56](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3A6000_O3_001.txt) 服务器平台（`-march=native`）： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3 -march=native`）: [4.42](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_7551_O3-march%3Dnative_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3 -march=native`）: [7.96](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_7742_O3-march%3Dnative_001.txt) - AMD EPYC 9R45 @ 4.5 GHz Zen 5（`-O3 -march=native`）: [16.2](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_9R45_O3-march%3Dnative_001.txt) - AMD EPYC 9T95 @ 3.7 GHz Zen 5c（`-O3 -march=native`）: [13.9](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_9T95_O3-march%3Dnative_001.txt) - AWS Graviton 5 @ 3.3 GHz Neoverse V3（`-O3 -march=native`）: [12.6](https://jia.je/benchmark/data-trixie/fp2017_rate1/AWS_Graviton_5_O3-march%3Dnative_001.txt) - Google Axion C4A @ Neoverse V2（`-O3 -march=native`）: [10.8](https://jia.je/benchmark/data-trixie/fp2017_rate1/Google_Axion_C4A_O3-march%3Dnative_001.txt) - Google Axion N4A @ Neoverse N3（`-O3 -march=native`）: [8.94](https://jia.je/benchmark/data-trixie/fp2017_rate1/Google_Axion_N4A_O3-march%3Dnative_001.txt) - Intel Xeon 6975P-C @ 3.9 GHz Redwood Cove（`-O3 -march=native`）: [11.0](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_6975P-C_O3-march%3Dnative_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3 -march=native`）: [5.58](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_E5-2680_v4_O3-march%3Dnative_001.txt) - Intel Xeon Gold 6430 @ 2.6 GHz Golden Cove（`-O3 -march=native`）: [7.64](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_Gold_6430_O3-march%3Dnative_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3 -march=native`）: [8.03](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_Platinum_8358P_O3-march%3Dnative_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3 -march=native`）: [3.19](https://jia.je/benchmark/data-trixie/fp2017_rate1/Kunpeng_920_O3-march%3Dnative_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3 -march=native`）: [7.63](https://jia.je/benchmark/data-trixie/fp2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3-march%3Dnative_001.txt) - Loongson 3C5000 @ 2.2 GHz LA464（`-O3 -march=native`）: [3.09](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3C5000_O3-march%3Dnative_001.txt) - Loongson 3C6000S @ 2.2 GHz LA664（`-O3 -march=native`）: [4.93](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3C6000S_O3-march%3Dnative_001.txt) 服务器平台： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3`）: [4.15](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_7551_O3_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3`）: [7.37](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_7742_O3_001.txt) - AMD EPYC 9R45 @ 4.5 GHz Zen 5（`-O3`）: [14.5](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_9R45_O3_001.txt) - AMD EPYC 9T95 @ 3.7 GHz Zen 5c（`-O3`）: [12.5](https://jia.je/benchmark/data-trixie/fp2017_rate1/AMD_EPYC_9T95_O3_001.txt) - AWS Graviton 5 @ 3.3 GHz Neoverse V3（`-O3`）: [12.4](https://jia.je/benchmark/data-trixie/fp2017_rate1/AWS_Graviton_5_O3_001.txt) - Google Axion C4A @ Neoverse V2（`-O3`）: [10.8](https://jia.je/benchmark/data-trixie/fp2017_rate1/Google_Axion_C4A_O3_001.txt) - Google Axion N4A @ Neoverse N3（`-O3`）: [9.18](https://jia.je/benchmark/data-trixie/fp2017_rate1/Google_Axion_N4A_O3_001.txt) - IBM POWER8 @ 3.2 GHz POWER8（`-O3`）: [3.47](https://jia.je/benchmark/data-trixie/fp2017_rate1/IBM_POWER8_O3_001.txt) - IBM POWER9 3.2 GHz @ 3.2 GHz POWER9（`-O3`）: [3.84](https://jia.je/benchmark/data-trixie/fp2017_rate1/IBM_POWER9_3.2_GHz_O3_001.txt) - IBM POWER9 3.8 GHz @ 3.8 GHz POWER9（`-O3`）: [4.75](https://jia.je/benchmark/data-trixie/fp2017_rate1/IBM_POWER9_3.8_GHz_O3_001.txt) - Intel Xeon 6975P-C @ 3.9 GHz Redwood Cove（`-O3`）: [10.3](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_6975P-C_O3_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3`）: [5.49](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_E5-2680_v4_O3_001.txt) - Intel Xeon Gold 6430 @ 2.6 GHz Golden Cove（`-O3`）: [7.01](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_Gold_6430_O3_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3`）: [7.24](https://jia.je/benchmark/data-trixie/fp2017_rate1/Intel_Xeon_Platinum_8358P_O3_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3`）: [3.17](https://jia.je/benchmark/data-trixie/fp2017_rate1/Kunpeng_920_O3_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3`）: [8.24](https://jia.je/benchmark/data-trixie/fp2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3_001.txt) - Loongson 3C5000 @ 2.2 GHz LA464（`-O3`）: [3.00](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3C5000_O3_001.txt) - Loongson 3C6000S @ 2.2 GHz LA664（`-O3`）: [4.75](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3C6000S_O3_001.txt) [4.77](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3C6000S_O3_002.txt) [4.75](https://jia.je/benchmark/data-trixie/fp2017_rate1/Loongson_3C6000S_O3_003.txt) #### Debian Bookworm 桌面平台（`-march=native`）： - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3 -march=native`）: [11.4](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_Ryzen_7_5700X_O3-march%3Dnative_001.txt) - AMD Ryzen 9 9950X @ 5.7 GHz Zen 5（`-O3 -march=native`）: [17.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_Ryzen_9_9950X_O3-march%3Dnative_001.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3 -march=native`）: [3.89](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Apple_M1_E-Core_O3-march%3Dnative_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3 -march=native`）: [11.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Apple_M1_P-Core_O3-march%3Dnative_001.txt) - Intel Core i9-10980XE @ 4.7 GHz (AVX-512 @ 4.0 GHz) Cascade Lake（`-O3 -march=native`）: [7.24](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-10980XE_O3-march%3Dnative_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3 -march=native`）: [16.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-14900K_P-Core_O3-march%3Dnative_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3 -march=native`）: [11.0](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_w9-3595X_O3-march%3Dnative_001.txt) - Qualcomm X1E80100 @ 4.0 GHz X Elite（`-O3 -march=native`）: [14.4](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Qualcomm_X1E80100_O3-march%3Dnative_001.txt) 桌面平台： - AMD Ryzen 5 7500F @ 5.0 GHz Zen 4（`-O3`）: [11.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_Ryzen_5_7500F_O3_001.txt) - AMD Ryzen 7 5700X @ 4.65 GHz Zen 3（`-O3`）: [9.91](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_Ryzen_7_5700X_O3_001.txt) - AMD Ryzen 9 9950X @ 5.7 GHz Zen 5（`-O3`）: [16.3](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_Ryzen_9_9950X_O3_001.txt) [16.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_Ryzen_9_9950X_O3_002.txt) - Apple M1 E-Core @ 2.1 GHz Icestorm（`-O3`）: [3.89](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Apple_M1_E-Core_O3_001.txt) - Apple M1 P-Core @ 3.2 GHz Firestorm（`-O3`）: [11.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Apple_M1_P-Core_O3_001.txt) - Intel Core i9-10980XE @ 4.7 GHz Cascade Lake（`-O3`）: [6.91](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-10980XE_O3_001.txt) - Intel Core i9-12900KS E-Core @ 4.1 GHz Gracemont（`-O3`）: [6.90](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-12900KS_E-Core_O3_001.txt) - Intel Core i9-12900KS P-Core @ 5.5 GHz Golden Cove（`-O3`）: [14.3](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-12900KS_P-Core_O3_001.txt) - Intel Core i9-14900K E-Core @ 4.4 GHz Gracemont（`-O3`）: [7.31](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-14900K_E-Core_O3_001.txt) - Intel Core i9-14900K P-Core @ 6.0 GHz Raptor Cove（`-O3`）: [16.1](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Core_i9-14900K_P-Core_O3_001.txt) - Intel Xeon w9-3595X @ 4.5 GHz Golden Cove（`-O3`）: [10.6](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_w9-3595X_O3_001.txt) - Qualcomm 8cx Gen3 E-Core @ 2.4 GHz Cortex-A78C（`-O3`）: [6.08](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Qualcomm_8cx_Gen3_E-Core_O3_001.txt) - Qualcomm 8cx Gen3 P-Core @ 3.0 GHz Cortex-X1C（`-O3`）: [8.07](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Qualcomm_8cx_Gen3_P-Core_O3_001.txt) - Qualcomm X1E80100 @ 4.0 GHz X Elite（`-O3`）: [14.4](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Qualcomm_X1E80100_O3_001.txt) 服务器平台（`-march=native`）： - AMD EPYC 9754 @ 3.1 GHz Zen 4c（`-O3 -march=native`）: [8.42](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9754_O3-march%3Dnative_001.txt) - AMD EPYC 9755 @ 4.1 GHz Zen 5（`-O3 -march=native`）: [14.4](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9755_O3-march%3Dnative_001.txt) - AMD EPYC 9K65 @ 3.7 GHz Zen 5c（`-O3 -march=native`）: [12.7](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9K65_O3-march%3Dnative_001.txt) - AMD EPYC 9K85 @ 4.1 GHz Zen 5（`-O3 -march=native`）: [14.2](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9K85_O3-march%3Dnative_001.txt) - AMD EPYC 9R14 @ 3.7 GHz Zen 4（`-O3 -march=native`）: [10.1](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9R14_O3-march%3Dnative_001.txt) - AMD EPYC 9T24 @ 3.7 GHz Zen 4（`-O3 -march=native`）: [10.1](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9T24_O3-march%3Dnative_001.txt) - AWS Graviton 3 @ 2.6 GHz Neoverse V1（`-O3 -march=native`）: [7.73](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_3_O3-march%3Dnative_001.txt) - AWS Graviton 4 @ 2.8 GHz Neoverse V2（`-O3 -march=native`）: [9.29](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_4_O3-march%3Dnative_001.txt) [9.35](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_4_O3-march%3Dnative_002.txt) - Intel Xeon 6981E Crestmont（`-O3 -march=native`）: [4.77](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_6981E_O3-march%3Dnative_001.txt) - Intel Xeon 6982P-C @ 3.6 GHz Redwood Cove（`-O3 -march=native`）: [9.50](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_6982P-C_O3-march%3Dnative_001.txt) - Intel Xeon D-2146NT @ 2.9 GHz Skylake（`-O3 -march=native`）: [5.48](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_D-2146NT_O3-march%3Dnative_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3 -march=native`）: [7.60](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_Platinum_8358P_O3-march%3Dnative_001.txt) - Intel Xeon Platinum 8581C @ 3.4 GHz Raptor Cove（`-O3 -march=native`）: [8.60](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_Platinum_8581C_O3-march%3Dnative_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3 -march=native`）: [3.17](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Kunpeng_920_O3-march%3Dnative_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3 -march=native`）: [8.01](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3-march%3Dnative_001.txt) 服务器平台： - AMD EPYC 7551 @ 2.5 GHz Zen 1（`-O3`）: [4.05](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_7551_O3_001.txt) - AMD EPYC 7742 @ 3.4 GHz Zen 2（`-O3`）: [7.12](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_7742_O3_001.txt) - AMD EPYC 7H12 @ 3.3 GHz Zen 2（`-O3`）: [6.61](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_7H12_O3_001.txt) - AMD EPYC 7K83 Zen 3（`-O3`）: [7.63](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_7K83_O3_001.txt) - AMD EPYC 9754 @ 3.1 GHz Zen 4c（`-O3`）: [7.64](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9754_O3_001.txt) - AMD EPYC 9755 @ 4.1 GHz Zen 5（`-O3`）: [13.2](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9755_O3_001.txt) - AMD EPYC 9K65 @ 3.7 GHz Zen 5c（`-O3`）: [11.7](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9K65_O3_001.txt) - AMD EPYC 9K85 @ 4.1 GHz Zen 5（`-O3`）: [13.0](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9K85_O3_001.txt) - AMD EPYC 9R14 @ 3.7 GHz Zen 4（`-O3`）: [9.26](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9R14_O3_001.txt) - AMD EPYC 9T24 @ 3.7 GHz Zen 4（`-O3`）: [9.23](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AMD_EPYC_9T24_O3_001.txt) - AWS Graviton 3 @ 2.6 GHz Neoverse V1（`-O3`）: [7.80](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_3_O3_001.txt) - AWS Graviton 3E @ 2.6 GHz Neoverse V1（`-O3`）: [8.10](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_3E_O3_001.txt) - AWS Graviton 4 @ 2.8 GHz Neoverse V2（`-O3`）: [9.36](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_4_O3_001.txt) [9.39](https://jia.je/benchmark/data-bookworm/fp2017_rate1/AWS_Graviton_4_O3_002.txt) - Ampere Altra @ 3.0 GHz Neoverse N1（`-O3`）: [5.26](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Ampere_Altra_O3_001.txt) - Hygon C86 7390（`-O3`）: [3.95](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Hygon_C86_7390_O3_001.txt) - IBM POWER8NVL @ 4.0 GHz POWER8（`-O3`）: [4.10](https://jia.je/benchmark/data-bookworm/fp2017_rate1/IBM_POWER8NVL_O3_001.txt) - Intel Xeon 6981E Crestmont（`-O3`）: [4.80](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_6981E_O3_001.txt) - Intel Xeon 6982P-C @ 3.6 GHz Redwood Cove（`-O3`）: [9.50](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_6982P-C_O3_001.txt) - Intel Xeon D-2146NT @ 2.9 GHz Skylake（`-O3`）: [5.00](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_D-2146NT_O3_001.txt) - Intel Xeon E5-2603 v4 @ 1.7 GHz Broadwell（`-O3`）: [3.14](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_E5-2603_v4_O3_001.txt) - Intel Xeon E5-2680 v3 @ 3.3 GHz Haswell（`-O3`）: [5.15](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_E5-2680_v3_O3_001.txt) - Intel Xeon E5-2680 v4 @ 3.3 GHz Broadwell（`-O3`）: [5.44](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_E5-2680_v4_O3_001.txt) - Intel Xeon E5-4610 v2 @ 2.7 GHz Ivy Bridge EP（`-O3`）: [3.74](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_E5-4610_v2_O3_001.txt) - Intel Xeon Platinum 8358P @ 3.4 GHz Sunny Cove（`-O3`）: [7.12](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_Platinum_8358P_O3_001.txt) - Intel Xeon Platinum 8576C Raptor Cove（`-O3`）: [8.14](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_Platinum_8576C_O3_001.txt) - Intel Xeon Platinum 8581C @ 3.4 GHz Raptor Cove（`-O3`）: [8.42](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Intel_Xeon_Platinum_8581C_O3_001.txt) - Kunpeng 920 @ 2.6 GHz TaiShan V110（`-O3`）: [3.13](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Kunpeng_920_O3_001.txt) - Kunpeng 920 HuaweiCloud kc2 @ 2.9 GHz（`-O3`）: [8.17](https://jia.je/benchmark/data-bookworm/fp2017_rate1/Kunpeng_920_HuaweiCloud_kc2_O3_001.txt) - T-Head Yitian 710 @ 3.0 GHz Neoverse N2（`-O3`）: [7.63](https://jia.je/benchmark/data-bookworm/fp2017_rate1/T-Head_Yitian_710_O3_001.txt) #### HarmonyOS 桌面平台（LTO）： - Huawei Kirin X90 E-Core @ 2.0 GHz（`-O3 -flto`）: [6.52](https://jia.je/benchmark/data-harmonyos/fp2017_rate1/Huawei_Kirin_X90_E-Core_O3-flto_001.txt) - Huawei Kirin X90 P-Core @ 2.3 GHz（`-O3 -flto`）: [7.42](https://jia.je/benchmark/data-harmonyos/fp2017_rate1/Huawei_Kirin_X90_P-Core_O3-flto_001.txt) 手机平台（LTO）： - Huawei Kirin 9010 E-Core Full @ 2.2 GHz（`-O3 -flto`）: [4.72](https://jia.je/benchmark/data-harmonyos/fp2017_rate1/Huawei_Kirin_9010_E-Core_Full_O3-flto_001.txt) - Huawei Kirin 9010 P-Core Best @ 2.3 GHz（`-O3 -flto`）: [6.22](https://jia.je/benchmark/data-harmonyos/fp2017_rate1/Huawei_Kirin_9010_P-Core_Best_O3-flto_001.txt) - Huawei Kirin 9010 P-Core Full @ 2.3 GHz（`-O3 -flto`）: [5.86](https://jia.je/benchmark/data-harmonyos/fp2017_rate1/Huawei_Kirin_9010_P-Core_Full_O3-flto_001.txt) #### 备注 1. SPEC FP 2017 Rate-1 结果在 AMD64 平台下受 `-march=native` 影响很明显，特别是有 AVX-512 的平台，因为不开 `-march=native` 时，默认情况下 SIMD 最多用到 SSE。ARM64 平台下 `-march=native` 没有什么影响，甚至有一定的劣化。 1. 部分内核版本（大约 6.7-6.11，在 6.12/6.11.7 中修复）会显著影响 503.bwaves_r 和 507.cactuBSSN_r 项目的性能，详见 [Intel Spots A 3888.9% Performance Improvement In The Linux Kernel From One Line Of Code](https://www.phoronix.com/news/Intel-Linux-3888.9-Performance)、[mm, mmap: limit THP alignment of anonymous mappings to PMD-aligned sizes](https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=d4148aeab412432bf928f311eca8a2ba52bb05df) 和 [kernel 6.10 THP causes abysmal performance drop](https://bugzilla.suse.com/show_bug.cgi?id=1229012)。 1. Qualcomm 8cx Gen3 在跑测试的时候，会因为过热降频，导致达不到最佳性能，三轮测试一轮比一轮慢。 1. 在华为云 kc2 实例上用 Debian Bookworm 带 `-march=native` 编译代码会报错，是 binutils 2.40 版本的问题；解决办法是手动安装一个 binutils 2.42： ```shell # Fix error building 511.povray_r: # /usr/bin/gcc -std=c99 -c -o image_validator/ImageValidator.o -DSPEC -DNDEBUG -Ifrontend -Ibase -I. -Ispec_qsort -DSPEC_AUTO_SUPPRESS_OPENMP -O3 -march=native -Wno-error=implicit-int -DSPEC_LP64 image_validator/ImageValidator.c # /tmp/cc0E80QY.s: Assembler messages: # /tmp/cc0E80QY.s:2340: Error: selected processor does not support `bcax v22.16b,v22.16b,v11.16b,v22.16b' # /tmp/cc0E80QY.s:2425: Error: selected processor does not support `bcax v8.16b,v8.16b,v16.16b,v8.16b' # /tmp/cc0E80QY.s:2502: Error: selected processor does not support `bcax v3.16b,v3.16b,v5.16b,v3.16b' apt update apt install -y texinfo wget https://mirrors.tuna.tsinghua.edu.cn/gnu/binutils/binutils-2.42.tar.xz tar xvf binutils-2.42.tar.xz cd binutils-2.42 mkdir build cd build/ ../configure make all -j4 make install -j4 ``` 5. Kirin 9010 因为散热问题，单独跑测试，和顺着跑一遍测试，结果差距比较大。因此提供了两组数据：Best（每一项单独跑，取最短时间，散热影响比较小）和 Full（按照顺序跑一次，散热影响比较大）。 ### 网上的数据 [高通 X Elite Oryon 微架构评测：走走停停 by JamesAslan](https://zhuanlan.zhihu.com/p/704707254): - Intel Core 13700K Raptor Cove: 14.56 - Qualcomm X1E80100 Oryon: 14.20 - AMD Ryzen 7700X Zen 4: 13.97 - Intel Core 12700K Golden Cove: 13.70 - Apple M2 Avalanche+Blizzard: 12.64 - AMD Ryzen 5950X Zen 3: 11.86 - Apple M1 Firestorm+Icestorm: 11.20 - Qualcomm 8 Gen 2 Cortex-X3: 9.91 [Running SPEC CPU2017 on Chinese CPUs, and More](https://old.chipsandcheese.com/2024/10/18/running-spec-cpu2017-on-chinese-cpus-and-more/) - AMD Ryzen 9 7950X3D Non-VCache: 15.4 - Intel Core Ultra 7 258V Lion Cove: 13.9 - Intel Core Ultra 7 115H Redwood Cove: 12 - Intel Core Ultra 7 258V Skymont: 7.94 - Intel Core i5-6600K Skylake: 7.92 - Intel Core Ultra 7 115H Crestmont: 6.86 - Loongson 3A6000 LA664: 5.49 - Mediatek Genio 1200 Cortex A78: 5.09 - Intel Core Ultra 7 115H Low Power Crestmont: 4.32 - AMD FX-8150: 3.63 - Loongson 3A5000 LA464: 3.38 - Intel Celeron J4125 Goldmont Plus: 2.45 - Amlogic S922X Cortex A73: 2.01 - Zhaoxin KaiXian KX-6640MA: 1.97 - Mediatek Genio 1200 Cortex A55: 1.01 [Running SPEC CPU2017 at Chips and Cheese?](https://old.chipsandcheese.com/2024/09/19/running-spec-cpu2017-at-chips-and-cheese/) - AMD Ryzen 9 9950X: 19.8 - AMD Ryzen 9 7950X3D Non-VCache: 15.3 - AMD Ryzen 9 7950X3D VCache: 12.7 - Intel Core Ultra 7 115H Redwood Cove: 11.2 - AMD Ryzen 9 3950X 3.5GHz: 7.26 - Intel Core Ultra 7 115H Crestmont: 5.83 - Ampere Altra: 5.62 - AmpereOne: 4.29 - AMD FX-8159: 3.47 [The AMD Ryzen 9 9950X and Ryzen 9 9900X Review: Flagship Zen 5 Soars - and Stalls](https://www.anandtech.com/show/21524/the-amd-ryzen-9-9950x-and-ryzen-9-9900x-review/5) - AMD Ryzen 9 9950X Zen 5: 17.72 - Intel Core i9-14900K Raptor Cove: 16.90 - AMD Ryzen 9 7950X Zen 4: 14.26 [Snapdragon X Elite Qualcomm Oryon CPU Design and Architecture Hot Chips 2024](https://www.servethehome.com/snapdragon-x-elite-qualcomm-oryon-cpu-design-and-architecture-hot-chips-2024-arm/) - Qualcomm X Elite Oryon on Linux: 17.77 - Qualcomm X Elite Oryon on Windows: 16.66 [ARM Cortex X1 微架构评测（上）：向山进发](https://zhuanlan.zhihu.com/p/619033328) - Zen 3 @ 4.95 GHz: 11.9 - Firestorm @ 3.0 GHz: 11.2 - Cortex X1 @ 3.0 GHz: 8.9 - Cortex A78 @ 2.4 GHz: 5.9 [极客湾•麒麟 9010，测评汇总：IPC 性能，巨幅提升！CPU 能效全频段领先，麒麟 9000S！](https://www.bilibili.com/opus/925502754370093090) - Huawei Kirin 9010: 7.77 - HUawei Kirin 9000s: 7.12 ### 多版本 GCC 和 LLVM 性能比较 在 Intel i9-14900K @ 5.7 GHz 上用 -O3 测试几种编译器组合的 SPEC FP 2017 Rate-1 性能： | Benchmark | GCC 12 | LLVM 20 | LLVM 19 | LLVM 18 | | --------------- | -------- | -------- | -------- | -------- | | 503.bwaves_r | **75.1** | 70.9 | 73.1 | 73.2 | | 507.cactuBSSN_r | **14.4** | 13.3 | 13.6 | 13.6 | | 508.namd_r | 9.24 | 10.5 | **10.6** | 10.5 | | 510.parest_r | **14.6** | **14.6** | **14.6** | 14.4 | | 511.povray_r | **14.6** | 13.7 | 13.7 | 13.7 | | 519.lbm_r | **12.1** | 11.2 | 11.2 | 11.2 | | 521.wrf_r | 13.5 | **14.0** | 13.3 | 13.3 | | 526.blender_r | **12.0** | 11.8 | 11.9 | 11.9 | | 527.cam4_r | **15.4** | 13.2 | 12.9 | 12.9 | | 538.imagick_r | 10.2 | 11.8 | 11.9 | **12.3** | | 544.nab_r | **12.6** | 8.47 | 8.22 | 8.22 | | 549.fotonik3d_r | **24.5** | **24.0** | 21.0 | 21.2 | | 554.roms_r | 14.1 | **14.3** | 13.7 | 13.7 | | geomean | **15.4** | 14.8 | 14.6 | 14.6 | 完整数据： - [GCC 12.2.0](https://jia.je/benchmark/data-bookworm/others/SPEC_FP_2017_Intel_i9-14900K_O3_GCC_12.txt) - [LLVM 18.1.8](https://jia.je/benchmark/data-bookworm/others/SPEC_FP_2017_Intel_i9-14900K_O3_LLVM_18.txt) - [LLVM 19.1.4](https://jia.je/benchmark/data-bookworm/others/SPEC_FP_2017_Intel_i9-14900K_O3_LLVM_19.txt) - [LLVM 20.1.5](https://jia.je/benchmark/data-bookworm/others/SPEC_FP_2017_Intel_i9-14900K_O3_LLVM_20.txt) 注：GCC 指 GCC + GFortran，LLVM 指 Clang + Flang-new