诚然,高端游戏通常需要高端硬件。但情况并非总是如此,尤其是笔记本电脑。在紧凑的机箱中,性能较强的组件往往会受到散热限制,有时甚至比技术较弱的同类产品更受影响。
CPU 就是一个很好的例子。在过去几年中,我们看到笔记本电脑的所谓较弱处理器比顶级芯片保持了更高更稳定的性能,而顶级芯片则由于热量限制而严重节流。
目前,我们已经测试了联想Legion Pro 15 16IAX10 的两种配置:第一种是配备英特尔酷睿 Ultra 9 处理器的高端型号;第二种是配备英特尔酷睿 Ultra 9 处理器的低端型号。 型号配备了英特尔酷睿 Ultra 9 275HX 和英伟达GeForce RTX 5070 Ti。以及最近推出的 版本配备了酷睿 Ultra 7 255HX 和 RTX 5070(非 Ti(非 Ti)的版本。 
| 规格 | Intel Core Ultra 7 255HX | Intel Core Ultra 9 275HX | 
|---|---|---|
| 架构/系列 | Arrow Lake (Series 2) | Arrow Lake (Series 2) | 
| 核心 | 20 (8P + 12E) | 24 (8P + 16E) | 
| 线程 | 20 | 24 | 
| 最大。Turbo P-Cores | 5.20 GHz | 5.40 GHz | 
| 最大E-Cores | 4.50 GHz | 4.60 GHz | 
| 基频 P-Cores | 2.40 GHz | 2.70 GHz | 
| 基本频率 E-Cores | 1.80 GHz | 2.10 GHz | 
| L3 / 智能高速缓存 | 30 MB | 36 MB | 
| 二级缓存 | 36 MB | 40 MB | 
| 设计功率(TDP,基准) | 55 W | 55 W | 
| 最大涡轮功率 | 160 W | 160 W | 
| 最小功率 | 45 W | 45 W | 
| 工艺 | 3 纳米 (TSMC) | 3 纳米 (TSMC) | 
» Notebookcheck多媒体笔记本电脑Top 10排名
» Notebookcheck游戏笔记本电脑Top 10排名
» Notebookcheck低价办公/商务笔记本电脑Top 10排名
» Notebookcheck高端办公/商务笔记本电脑Top 10排名
» Notebookcheck工作站笔记本电脑Top 10排名
» Notebookcheck亚笔记本电脑Top 10排名
» Notebookcheck超级本产品Top 10排名
» Notebookcheck变形本产品Top 10排名
» Notebookcheck平板电脑Top 10排名
» Notebookcheck智能手机Top 10排名
» Notebookcheck评测过最出色的笔记本电脑屏幕
» Notebookcheck售价500欧元以下笔记本电脑Top 10排名
» Notebookcheck售价300欧元以下笔记本电脑Top 10排名
虽然这两个版本的 GPU 性能差距明显,但 CPU 方面的差距却出奇地小--尽管 Ultra 9 提供了 24 个内核(而 Ultra 9 只有 20 个)和稍高的提升时钟。在所有 CPU 基准测试中,总体性能差距约为 7% 至 8%。
不过,根据工作负载的不同,差距会有所扩大。在单核基准测试中,Ultra 9 的领先优势还在误差范围之内,但在多核测试中,它的领先优势就会扩大到 10%-13%。
Cinebench R23: Multi Core | Single Core
Cinebench R20: CPU (Multi Core) | CPU (Single Core)
Cinebench R15: CPU Multi 64Bit | CPU Single 64Bit
Blender: v2.79 BMW27 CPU
7-Zip 18.03: 7z b 4 | 7z b 4 -mmt1
Geekbench 6.5: Multi-Core | Single-Core
Geekbench 5.5: Multi-Core | Single-Core
HWBOT x265 Benchmark v2.2: 4k Preset
LibreOffice : 20 Documents To PDF
R Benchmark 2.5: Overall mean
| CPU Performance rating | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX | |
| Average of class Gaming | |
| Cinebench R23 / Multi Core | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (27129 - 34005, n=4) | |
| Average of class Gaming (5668 - 40970, n=145, last 2 years) | |
| Cinebench R23 / Single Core | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (2111 - 2167, n=4) | |
| Average of class Gaming (1136 - 2267, n=145, last 2 years) | |
| Cinebench R20 / CPU (Multi Core) | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (10626 - 13056, n=4) | |
| Average of class Gaming (2179 - 16108, n=144, last 2 years) | |
| Cinebench R20 / CPU (Single Core) | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (822 - 835, n=4) | |
| Average of class Gaming (439 - 870, n=144, last 2 years) | |
| Cinebench R15 / CPU Multi 64Bit | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (4114 - 5190, n=4) | |
| Average of class Gaming (1537 - 6271, n=144, last 2 years) | |
| Cinebench R15 / CPU Single 64Bit | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (322 - 328, n=4) | |
| Average of class Gaming (188.8 - 343, n=143, last 2 years) | |
| Blender / v2.79 BMW27 CPU | |
| Average of class Gaming (80 - 517, n=140, last 2 years) | |
| Average Intel Core Ultra 7 255HX (104 - 122, n=4) | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Lenovo Legion 5 Pro 16IAX10H | |
| 7-Zip 18.03 / 7z b 4 | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (87806 - 104055, n=4) | |
| Average of class Gaming (23795 - 148086, n=142, last 2 years) | |
| 7-Zip 18.03 / 7z b 4 -mmt1 | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (6407 - 6576, n=4) | |
| Average of class Gaming (4199 - 7508, n=142, last 2 years) | |
| Geekbench 6.5 / Multi-Core | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (16396 - 19623, n=5) | |
| Average of class Gaming (5340 - 22104, n=142, last 2 years) | |
| Geekbench 6.5 / Single-Core | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (2917 - 3028, n=5) | |
| Average of class Gaming (1490 - 3253, n=142, last 2 years) | |
| Geekbench 5.5 / Multi-Core | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (18263 - 21014, n=4) | |
| Average of class Gaming (4557 - 27010, n=141, last 2 years) | |
| Geekbench 5.5 / Single-Core | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (2112 - 2211, n=4) | |
| Average of class Gaming (986 - 2423, n=141, last 2 years) | |
| HWBOT x265 Benchmark v2.2 / 4k Preset | |
| Lenovo Legion 5 Pro 16IAX10H | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (30.4 - 36.9, n=4) | |
| Average of class Gaming (7.7 - 44.3, n=142, last 2 years) | |
| LibreOffice / 20 Documents To PDF | |
| Average of class Gaming (19 - 88.8, n=141, last 2 years) | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (35.4 - 45.9, n=4) | |
| Lenovo Legion 5 Pro 16IAX10H | |
| R Benchmark 2.5 / Overall mean | |
| Average of class Gaming (0.3551 - 0.759, n=143, last 2 years) | |
| Lenovo Legion Pro 5 16IAX10, RTX 5070 | |
| Average Intel Core Ultra 7 255HX (0.421 - 0.4287, n=4) | |
| Lenovo Legion 5 Pro 16IAX10H | |
* ... smaller is better
有趣的是,在 Cinebench R15 耐用性测试中,两款处理器的性能曲线几乎完全相同。这意味着Legion Pro 中性能更强的芯片不会比低端机型更早或更猛烈地节流。
Cinebench R15 多重测试
简而言之,在大多数实际应用场景中,尤其是在游戏场景中,CPU 时钟速度通常会比峰值低几百 MHz,GPU 成为决定性因素,因此两者之间的差异可以忽略不计。除非你特别需要最高的多核性能,否则选择 CPU 稍微小一点的笔记本电脑是一种明智的省钱方法,同时又不会牺牲太多的游戏性能。



