TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Modern CPUs have a backstage cast

212 pointsby hlandaualmost 2 years ago

13 comments

chasilalmost 2 years ago
&quot;...this is interesting is because POWER9 is basically the first time the public got a real view of how sophisticated the backstage cast actually is of a modern server CPU.&quot;<p>Not quite correct; the OpenSPARC T1 and T2 were publicly released and available by 2008.<p><a href="https:&#x2F;&#x2F;www.oracle.com&#x2F;servers&#x2F;technologies&#x2F;opensparc.html" rel="nofollow">https:&#x2F;&#x2F;www.oracle.com&#x2F;servers&#x2F;technologies&#x2F;opensparc.html</a><p>&quot;Large parts of this process are handled by vendor-supplied mystery firmware blobs, which may as well be boxes with “???” written in them.<p>The maintainers of the me_cleaner script likely have the clearest view of what is known.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;corna&#x2F;me_cleaner">https:&#x2F;&#x2F;github.com&#x2F;corna&#x2F;me_cleaner</a>
评论 #36128941 未加载
kccqzyalmost 2 years ago
&gt; It&#x27;s responsible for initialising the chip and getting it out of bed enough to the point where at least one of the main cores can run using cache-as-RAM mode<p>The somewhat surprising but true implication is that on boot, the CPU is initialized before the RAM is initialized. So there is a window of time during boot when the main core on the CPU is running instructions that cannot access the RAM. Even on register-starved x86 it is possible to write code without using RAM, but it certainly seems more convenient to me to treat the cache as RAM.<p>Documentation for a special compiler that compiles to code that doesn&#x27;t use RAM: <a href="https:&#x2F;&#x2F;github.com&#x2F;wt&#x2F;coreboot&#x2F;blob&#x2F;master&#x2F;util&#x2F;romcc&#x2F;romcc.1">https:&#x2F;&#x2F;github.com&#x2F;wt&#x2F;coreboot&#x2F;blob&#x2F;master&#x2F;util&#x2F;romcc&#x2F;romcc....</a>
评论 #36129292 未加载
评论 #36128882 未加载
dist-epochalmost 2 years ago
A modern motherboard can update it&#x27;s BIOS from a USB stick WITHOUT a CPU or memory installed.<p>Think about that. The motherboard &quot;knows&quot; how to read a FAT file system from a USB mass storage device, verify it&#x27;s digital signature and flash it with no main CPU or memory.
评论 #36128883 未加载
pabs3almost 2 years ago
This reminds me of the &quot;It&#x27;s Time for Operating Systems to Rediscover Hardware&quot; talk by Timothy Roscoe:<p><a href="https:&#x2F;&#x2F;www.usenix.org&#x2F;conference&#x2F;osdi21&#x2F;presentation&#x2F;fri-keynote" rel="nofollow">https:&#x2F;&#x2F;www.usenix.org&#x2F;conference&#x2F;osdi21&#x2F;presentation&#x2F;fri-ke...</a>
JdeBPalmost 2 years ago
As the author of <a href="https:&#x2F;&#x2F;superuser.com&#x2F;a&#x2F;347115&#x2F;38062" rel="nofollow">https:&#x2F;&#x2F;superuser.com&#x2F;a&#x2F;347115&#x2F;38062</a> and <a href="https:&#x2F;&#x2F;superuser.com&#x2F;a&#x2F;345333&#x2F;38062" rel="nofollow">https:&#x2F;&#x2F;superuser.com&#x2F;a&#x2F;345333&#x2F;38062</a>, you have my sympathy about the &quot;pack of lies&quot; involving real mode and several wrong combinations of selector and offset.
评论 #36130927 未加载
评论 #36130523 未加载
评论 #36134393 未加载
wmfalmost 2 years ago
Much of the openness of Power7&#x2F;8&#x2F;9 was <i>encouraged</i> by Google who wanted to have control over all the firmware, even the secret firmware. I think Google is also auditing PSP&#x2F;ME source code but the public only sees the audit results.
buildbotalmost 2 years ago
“Turtles all the way down” Modern CPUs are so complex you need simpler ones to abstract it! Very cool breakdown of how power9 does this.
travisgriggsalmost 2 years ago
I miss these kinds of articles on the net. Is anyone else reminded of the CPU Praxis articles that were part of ARS Technica&#x27;s early rise to popularity? I really miss those. This article, is of course, much shorter, but still, I miss that sort of content on the internet.
shrubblealmost 2 years ago
A while ago I bought some older AMD 8350 systems, which apparently are the last without a PSP, the platform security processor.<p>I did this as a sort of &#x27;just in case&#x27; setup, was planning to put OpenSolaris on it and run things under Zones or LX zones and to run it as a backup server. Fast enough to get some work done and possibly more secure if the PSP is ever used&#x2F;broken maliciously...
评论 #36131380 未加载
评论 #36142613 未加载
MichaelZuoalmost 2 years ago
&gt; The Self-Boot Engine (SBE) (quantity: 1) is a core which is responsible for booting the entire system. It&#x27;s responsible for initialising the chip and getting it out of bed enough to the point where at least one of the main cores can run using cache-as-RAM mode; it does little after that point. It has some SRAM to do its work in and uses a slightly custom variant of the 32-bit Power ISA, extended to support 64-bit loads and stores using adjacent GPR pairs. This core design is known as a PPE. It&#x27;s the first thing that runs on the CPU die.<p>What I’m curious about is how does the ‘Self Boot Engine’ initialize itself in the first few miliseconds?<p>Maybe, a motherboard chip does the actual work, but that just invites the same question, at some point something must be initializing itself, how?
评论 #36134282 未加载
评论 #36134280 未加载
评论 #36135318 未加载
giuliomagnificoalmost 2 years ago
I understood nothing (as a sysadmin) but this looks like a very interesting article for who can understand it.
评论 #36128637 未加载
Simplicitasalmost 2 years ago
Doesn’t mention the special core reserved for the NSA and other national security agencies :-)
评论 #36142637 未加载
tyingqalmost 2 years ago
I was surprised when I read how the Raspberry PI&#x27;s GPU handles the booting.<p><a href="https:&#x2F;&#x2F;forums.raspberrypi.com&#x2F;viewtopic.php?t=266130" rel="nofollow">https:&#x2F;&#x2F;forums.raspberrypi.com&#x2F;viewtopic.php?t=266130</a>