从英特尔十一代酷睿处理器i7-1185G7测试 谈十一代酷睿性能提升之根本 ZT
<p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">英特尔十一代酷睿处理器上市已经有一段时间了，基于其打造的新品也已经全面上市。不过对于用户而言，最常见的十一代酷睿产品主要是搭载了英特尔十一代酷睿i5-1135G7和i7-1165G7两款处理器。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">但其实关注十一代酷睿的朋友已经发现了，酷睿i7-1165G7并非该系列最高端的型号，因为在其之上还有一颗酷睿i7-1185G7处理器。它与i7-1165G7处理器的主要区别就在于主频和睿频加速能力。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753207" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><span style="font-size: 0px; line-height: 0; height: 0px; display: block; clear: both;"></span><img id="content-first-img" src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZTuIJCwxAAG0a8zSI7IAAB0hwGR_8AAAbSD519.jpg" width="640" height="258" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 0px; border-style: initial; max-width: 100%;"></a></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">那么十一代酷睿i7-1185G7处理器的实际性能表现如何呢？近期我们终于拿到了搭载这颗处理器的<a href="https://nb.zol.com.cn/" class="hui14_line" style="text-decoration-line: none; color: rgb(51, 51, 51); padding: 0px 0px 2px; border-bottom: 1px dashed rgb(51, 51, 51);">笔记本</a>产品，接下来我们就来看看十一代酷睿家族目前的这颗顶级处理器，究竟有怎样的性能表现？</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">首先我们先来看看这颗处理器的基本信息，根据AIDA 64的CPUID检索来看，这颗处理器为原生4核8线程设计，主频为3GHz，睿频加速可达4.8GHz，三级缓存高达12MB。此外这颗处理器支持12-28W的动态TDP调整，基于深度优化的Willow Cove微架构设计，并且采用了更为先进的10nm SuperFin<a href="https://xiazai.zol.com.cn/detail/44/432854.shtml" class="hui14_line" style="text-decoration-line: none; color: rgb(51, 51, 51); padding: 0px 0px 2px; border-bottom: 1px dashed rgb(51, 51, 51);">晶体管</a>工艺。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/04/08/ChMkKV_S1QSICTHzAAlbo8xBsh4AAGkjgI_xTwACVu7419.jpg" width="640" height="746" alt="联想YOGA Pro 14c 2021评测 " style="vertical-align: middle; border: 0px; max-width: 100%;"><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">其实从最基本的参数来看，如果比较熟悉近两代酷睿处理器的话就会知道，十一代酷睿处理器主要是解决了十代Ice Lake，也就是上一代10nm处理器在主频和睿频加速方面的不足。那么在大幅提升主频和睿频加速能力之后，这颗十一代酷睿家族中的顶级处理器的实际性能表现如何呢？</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">先以CINEBENCH R15/R20测试标准为例。在CR15标准下，其单核得分217cb，多核得分887cb；CR20标准下，其单核得分574cb，多核得分2055cb。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><img src="https://article-fd.zol-img.com.cn/t_s312x2000/g6/M00/08/0D/ChMkKV_Z0w2IE1RBAAN-pG_WOqAAAG1wwOQvT0AA368876.jpg" width="312" height="705" alt="皮革A面 极致体验 联想YOGA Pro 14c 2021评测 " style="vertical-align: middle; border: 0px; max-width: 100%;"> <img src="https://article-fd.zol-img.com.cn/t_s312x2000/g6/M00/08/0D/ChMkKl_Z0w2IbJrxAAQpOZlbyDEAAG1wwOUO_kABClR297.jpg" width="312" height="705" alt="皮革A面 极致体验 联想YOGA Pro 14c 2021评测 " style="vertical-align: middle; border: 0px; max-width: 100%;"></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">可以看到，如果对比上一代酷睿i7-1065G7处理器的话，十一代酷睿处理器在单核和多核性能方面都有了非常明显的提升。其中，多核性能提升约30%，单核性能提升约31%。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/02/08/ChMkKV_pqhSIFFcdAABdQPEMvPgAAHcQQLj0CAAAF1Y371.jpg" width="640" height="358" alt="十一代酷睿 " style="vertical-align: middle; border: 1px solid rgb(229, 229, 229); max-width: 100%; cursor: pointer;"><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">在Geekbench 5测试标准中，英特尔酷睿i7-1185G7处理器单核得分1511，多核得分5212。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/04/08/ChMkKV_S1QiIK476AAVoljseOGgAAGkjgJk3vgABWiu099.jpg" width="640" height="462" alt="联想YOGA Pro 14c 2021评测 " style="vertical-align: middle; border: 0px; max-width: 100%;"></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">对比Geekbench 5数据库中的多个酷睿i7-1065G7处理器的测试结果，也可以看到十一代酷睿处理器在大幅提升主频和睿频加速能力之后，无论是单核还是多核性能，都有了非常巨大的提升。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=21004465" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/02/08/ChMkKl_pqquIdHUIAAIodBaDtcIAAHcQgDJ2-QAAiiM620.jpg" width="640" height="517" alt="十一代酷睿 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/04/08/ChMkKV_S1QiIfo31AAJVhFE3QBUAAGkjgJs9CwAAlWc963.jpg" width="640" height="275" alt="联想YOGA Pro 14c 2021评测 " style="vertical-align: middle; border: 0px; max-width: 100%;"><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">此外在渲染能力方面，我们以V-Ray <a href="https://xiazai.zol.com.cn/detail/43/428734.shtml" class="hui14_line" style="text-decoration-line: none; color: rgb(51, 51, 51); padding: 0px 0px 2px; border-bottom: 1px dashed rgb(51, 51, 51);">Benchmark</a>渲染测试为参考标准，可以看到，这颗处理器采样数量达到6056ksamples，说明其在应对高负载渲染任务时能够给用户带来比较不错的效率，而且低功耗处理器能够达到6000以上的采样率实属不易。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">那么为什么十一代酷睿处理器相对十代酷睿处理器有如此巨大的提升呢？我们先来看看十一代酷睿处理器的新特性，它们包括：</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);"><strong>·全新设计的Willow Cove微架构<br>·全新的10nm SuperFin晶体管技术<br></strong><strong>·全新的Intel Iris Xe锐炬核显<br></strong><strong>·全新的AI功能<br></strong><strong>·全新的媒体和显示引擎<br></strong><strong>·全新的硬件强化安全特性<br></strong><strong>·整合了Thunderbolt 4<br></strong><strong>·整合了全新的PCIe 4.0接口</strong></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753208" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZTyIVjmaAAIM1enFbKsAAB0hwGWNQ8AAgzt640.jpg" width="640" height="225" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">而在这些新特性中，帮助十一代酷睿处理器在性能上实现突破的主要有三点，其一是10nm SuperFin晶体管技术，其二是Willow Cove微架构，其三是引入了AI加速引擎。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);"><strong>·SuperFin晶体管技术助力性能释放</strong></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">全新的10nm SuperFin晶体管技术可以说是十一代酷睿处理器性能突破的关键所在，它从最底层帮助十一代酷睿构筑根基。<br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753224" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZgOIQ4ciAAGiY6j0AhsAAB0iQD_g_gAAaJ7877.jpg" width="640" height="243" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 0px; border-style: initial; max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753223" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZgKIYSQ7AAFZxBj4EDQAAB0iQD49qMAAVnc434.jpg" width="640" height="279" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 0px; border-style: initial; max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">简单而言，SuperFin技术从底层晶体管设计上实现了进一步优化，不仅重新设计了晶体管，而且重新设计了金属堆栈。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">首先，英特尔通过添加全新的高性能晶体管，以及改善的栅极工艺来提升<a href="https://driver.zol.com.cn/" class="hui14_line" style="text-decoration-line: none; color: rgb(51, 51, 51); padding: 0px 0px 2px; border-bottom: 1px dashed rgb(51, 51, 51);">驱动</a>电流，使电荷具有更高的移动性，并降低了源漏电阻，实现了更低的电容。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">英特尔在高频敏感IP中使用这一全新的晶体管技术，如处理器内核，高速总线和内存子系统，同时还在非高频率关键IP中，如Type-C和PCIe中，使用现有高阀值电压晶体管，从而使其变得更加高效。这些技术使得晶体管的运行速度得到提升，同时降低泄漏，进而降低这些晶体管的运行电压。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">其次，在改善金属堆栈方面，英特尔的工程师大幅改善了中低层电阻，并大量使用导通孔。同时，在晶体管顶层还增加了2层额外的高性能层，以使其达到更高的峰值频率。此外，通过提升MIM电容器能力，Tiger Lake处理器可以胜任更高负载的任务，提供快速而稳定的供电响应。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">因此，Tiger Lake处理器能够释放出比前代产品更强的CPU和GPU性能，根本原因就在于全新的SuperFin晶体管技术带来了卓越的底层优化。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);"><strong>·更加全能的架构单元</strong></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">底层晶体管技术突破之外，更加高效的微架构设计可以让晶体管工艺得到更为充分的发挥。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">上一代Ice Lake处理器采用了Sunny Cove微架构，而全新的Tiger Lake则采用了在Sunny Cove微架构上进一步优化的Willow Cove微架构，这使得Tiger Lake处理器在4核8线程基础之上，最高做到了4.8GHz单核睿频，同时其整体功耗能够控制在7-15W和12-28W之间，给OEM留下了极大的可控空间。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">我们都知道，Ice Lake的软肋在于主频和睿频频率都比较低，这直接影响了Ice Lake CPU部分的性能释放，而Tiger Lake通过优化微架构来提升睿频能力，并使之看齐标压处理器，可见英特尔对Ice Lake的短板认知还是非常明确的。Tiger Lake一举弥补了这一不足。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753210" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZX6IVjU9AAPFKdCFqYYAAB0hwJiEaUAA8VB253.jpg" width="640" height="259" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">为了提升整体性能，Tiger Lake的总线从基本结构上实现了带宽提升，有效降低延迟。并且采用了全新的内存控制器，支持LP4/x-4266和DDR4-3200规格的内存，最大容量分别可支持32GB和64GB，进一步消除内存瓶颈。</p> <strong style="color: rgb(51, 51, 51); font-family: "Microsoft Yahei"; font-size: 16px;">·三大AI技术加持使Tiger Lake更全能</strong><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">无论是CPU还是GPU，还是XPU，通过人工智能算法进行加速、帮助提升计算效率已经是行业的大趋势。而近年来对AI极为重视的英特尔，自然会将AI技术释放到自家的处理器当中，因此其为Tiger Lake带来了包括Intel DL Boost：VNNI，Intel DL Boost：DP4a，Intel Gaussian&Neural Accelerator在内的三大AI<a href="https://xiazai.zol.com.cn/netaccelerate_soft_index/netaccelerate_page_1.html" class="hui14_line" style="text-decoration-line: none; color: rgb(51, 51, 51); padding: 0px 0px 2px; border-bottom: 1px dashed rgb(51, 51, 51);">加速器</a>。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753227" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZiaIYqh-AAChK1heaXMAAB0iQG1MkIAAKFD465.jpg" width="640" height="130" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">很多人对于AI在PC上应用的意义并没有概念。其实简单来说，AI技术能够帮助PC更加智能的调控性能、延长续航，也能够在实际应用中帮助用户简单、快速的完成相关任务。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753228" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZieIITAwAAJHKeLQm9YAAB0iQG3NlcAAkdB928.jpg" width="640" height="295" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753229" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZieICJIoAAJKln0rJeMAAB0iQG5fZgAAkqu928.jpg" width="640" height="305" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">首先来看本次最为引人关注的Intel DL Boost：DP4a</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">Intel DL Boost：DP4a是Xe核显的标志性指令之一。它使用32位元累积来计算，一个4位浅层矢量点积，从而加快了8位元整数推断。这种强大的4元素矢量点积扩展，经过优化之后加速了用于人工智能的推断，并且在深度学习中也有所应用。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753230" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZimII4exAAICql1I40QAAB0iQHBymEAAgLC295.jpg" width="640" height="349" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">Intel DL Boost：VNNI并非首次出现在酷睿平台之上。英特尔在Tiger Lake上继续沿用了这一技术。VNNI为处理器提供了卓越的人工智能性能或矢量神经网络指令，可以加速基于卷积神经网络的算法。它把此前的三个独立指令，也就是一个矢量相乘和两个矢量相加，整合成一个矢量点积指令，从而提供比竞品高1.7倍的性能。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753231" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZiqICdH0AAHKWagn8LwAAB0iQHFJ-cAAcpx398.jpg" width="640" height="339" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">Intel Gaussian&Neural Accelerator，即GNA2.0，是英特尔打造的低功耗人工智能加速器。它专门针对工作流负载进行了优化，比如在听写、翻译或是动态降噪方面，能够帮助用户提升体验。如视频会议过程中的背景消噪，就是其典型应用场景。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51); text-align: center;"><a href="https://nb.zol.com.cn/pic/7596084.html#id=20753232" rel="nofollow" style="text-decoration-line: none; color: rgb(10, 79, 181);"><img src="https://article-fd.zol-img.com.cn/t_s640x2000/g6/M00/0D/03/ChMkKV9QZiuIEsvAAALpFYQZftkAAB0iQHIjWYAAukt462.jpg" width="640" height="356" alt="英特尔11代酷睿深度解析 " style="vertical-align: middle; border-width: 1px; border-style: solid; border-color: rgb(229, 229, 229); max-width: 100%;"></a><br></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">此外，GNA2.0最大的特性在于它是一个独立的IP区块，因此能够在CPU或GPU忙于执行其它工作负载时而正常工作，并且能够以更省电、更低功耗的状态，实现每毫瓦每秒10亿次运算。而其峰值运算能力可达到每秒380亿次！</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">通过把正确的人工智能工作流工作负载交付给GNA处理，就可以进一步缓解CPU的计算负荷，这是GNA技术最大的价值所在。</p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);"><strong>·结语</strong></p><p style="margin-top: 20px; font-variant-numeric: normal; font-variant-east-asian: normal; font-stretch: normal; font-size: 16px; line-height: 30px; font-family: "Microsoft Yahei"; word-break: normal; color: rgb(51, 51, 51);">目前，英特尔十一代酷睿家族只推出了低功耗平台，而在2021年1月，英特尔或将推出桌面级十一代酷睿处理器，同时如果没有意外的话，会在3、4月份推出标压移动级处理器。而十一代酷睿桌面级和标压移动级处理器将为我们带来怎样的性能表现呢？让我们拭目以待！</p><br><br> 页:
[1]
