TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Apple “pays to make Galaxy S23 seem slower”, True or false? Key CEO reacts

12 点作者 jerryjerryjerry大约 2 年前

2 条评论

kramerger大约 2 年前
Someone on reddit did a &quot;normal usage&quot; test and the Samsung was a tiny bit faster than the iphone. But every benchmark I&#x27;ve seen shows that the iPhone is significantly faster.<p>This just shows that these cross-plattform benchmarks are meaningless. Just getting the CPU to switch to the fastest core and run at max speed for a few minutes is a science in itself. And don&#x27;t forget that the software stacks are completely different, we are comparing apples to oranges.<p>But seriously, will anyone from either camp switch due to benchmarks? I don&#x27;t think anyone outside MacWorld and a few bloggers even care
评论 #34959765 未加载
2OEH8eoCRo0大约 2 年前
Do they just keep adding new tests that iPhones have hardware acceleration for? Linus Torvalds complained about Geekbench awhile back because many of the integer workloads were offloaded to the crypto unit.<p><a href="https:&#x2F;&#x2F;www.realworldtech.com&#x2F;forum&#x2F;?threadid=136526&amp;curpostid=136666" rel="nofollow">https:&#x2F;&#x2F;www.realworldtech.com&#x2F;forum&#x2F;?threadid=136526&amp;curpost...</a><p>&gt; On ARM64, that SHA1 performance is hardware-assisted. I don&#x27;t know if SHA2 is too, but Aarch64 does apparently do SHA256 in the crypto unit, so it might be fully or partially so. And on both ARM and x86, the AES numbers are similarly just about the crypto unit. So basically a quarter to a third of the &quot;integer&quot; workloads are just utter BS. They are not comparable across architectures due to the crypto units, and even within one architecture the numbers just don&#x27;t mean much of anything.
评论 #34961606 未加载