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。

Japan Captures TOP500 Crown with Arm-Powered Supercomputer

315 点作者 l31g将近 5 年前

32 条评论

pininja将近 5 年前
The link is working for me, here are the details on the winner.<p>“The new top system, Fugaku, turned in a High Performance Linpack (HPL) result of 415.5 petaflops, besting the now second-place Summit system by a factor of 2.8x. Fugaku, is powered by Fujitsu’s 48-core A64FX SoC, becoming the first number one system on the list to be powered by ARM processors. In single or further reduced precision, which are often used in machine learning and AI applications, Fugaku’s peak performance is over 1,000 petaflops (1 exaflops). The new system is installed at RIKEN Center for Computational Science (R-CCS) in Kobe, Japan.
评论 #23605199 未加载
评论 #23605609 未加载
Animats将近 5 年前
The end of the US semiconductor industry is now in sight.<p>The only US owned state of the art fabs in the US belong to Intel. Intel survives because they have a high margin on x86 CPUs. Today, TMSC announced 5nm, and the top supercomputer is ARM-based.<p>Apple seems to be going ARM. Chromebooks are ARM. Microsoft now offers Windows on ARM, on the Surface Pro X. Mobile never used x86. x86 is on the way out. What&#x27;s left for Intel?<p>(Micron is still a major force in DRAM, amazingly.)
评论 #23604931 未加载
评论 #23610373 未加载
评论 #23607615 未加载
评论 #23605999 未加载
评论 #23605582 未加载
评论 #23610137 未加载
gnufx将近 5 年前
Perhaps it&#x27;s worth pointing out some context. Given the remarkable predecessor, K Computer, this was only a matter of time. (I heard a great early talk on K, and I wish I knew the speaker for credit who was obviously working quite hard in English, but flawless, ending with basically we did it all ourselves largely de nuovo.) It seems that given the current circumstances, they haven&#x27;t kept to schedule -- it was supposed to be operating <i>next</i> year.<p>There&#x27;s a lot non-mainstream in this, like K, but partly influenced by K experience. Unusually, it&#x27;s all apparently specifically designed for the job, from the processor to the operating system (only partly GNU&#x2F;Linux). Notably, despite the innovation, it should still run anything that can reasonably be built for aarch64 straight off and use the whole node, even if it doesn&#x27;t run particularly fast; contrast GPU-based systems. (With something like simde, you may even be able to run typical x86-specific code.) However, the amount of memory&#x2F;core is surprising -- even less than Blue Gene Q -- and I wonder how that works out for large-scale materials science work for which it&#x27;s obviously prepared. Also note Fujitsu&#x27;s consideration of reliability, though the oft-quoted theory of failure rates in exascale-ish machines was obviously wrong, otherwise as the Livermore CTO said, he&#x27;d be out of a job.<p>The bad news for anyone potentially operating a similar system in a university, for instance, is that the typical nightmare proprietary software is apparently becoming available for it...
评论 #23610765 未加载
评论 #23609996 未加载
评论 #23609840 未加载
ViralBShah将近 5 年前
Given that today&#x27;s HPC architectures are mostly power constrained, and a majority of the FLOPS often come from GPUs (for their flop&#x2F;watt ratios), this direction is not surprising.<p>ARM has been making major strides in the high performance area. The new AWS Graviton processors are pretty nice from what I have heard. And then there&#x27;s the ARM in Mac. Yup and Julia will run on all of these!<p>While I say all of this, I should also point out that the top500 benchmark pretty much is not representative of most real-life workloads, and is largely based on your ability to solve the largest dense linear solve you possibly can - something almost no real application does.<p>(The website is down, so I haven&#x27;t been able to look at the specs of the actual machine).
评论 #23601312 未加载
评论 #23601334 未加载
评论 #23601537 未加载
评论 #23605902 未加载
评论 #23604743 未加载
评论 #23601444 未加载
评论 #23602138 未加载
dman将近 5 年前
Really wish Fujitsu sold a developer kit with an A64fx chip - its the only shipping ARM chip with SVE that I know of and I would love to get my hands on one to play with.
评论 #23602385 未加载
评论 #23602728 未加载
评论 #23607760 未加载
评论 #23603133 未加载
calaphos将近 5 年前
Even more impressive than the linpack result (2.8x faster than the runner up) is the HPCG result at 4.6x the result of summit in second place.<p>That benchmark consists of more sparse matrixes which are a lot more realistic depiction of hpc workloads. Seems to scale a lot better with irregular access patterns than basically Nvidia GPUs on the other systems.
ksec将近 5 年前
Well the link is dead. But I am guessing it is from Fujitsu A64FX, a 512 bit SIMD extension for ARM.<p>Edit: Turns out I was right. May be this link is better.<p><a href="https:&#x2F;&#x2F;www.anandtech.com&#x2F;show&#x2F;15869&#x2F;new-1-supercomputer-fujitsus-fugaku-and-a64fx-take-arm-to-the-top-with-415-petaflops" rel="nofollow">https:&#x2F;&#x2F;www.anandtech.com&#x2F;show&#x2F;15869&#x2F;new-1-supercomputer-fuj...</a>
评论 #23604796 未加载
Symmetry将近 5 年前
For those interested in more details they did a presentation at Hot Chips. The slides are here:<p><a href="https:&#x2F;&#x2F;www.hotchips.org&#x2F;hc30&#x2F;2conf&#x2F;2.13_Fujitsu_HC30.Fujitsu.Yoshida.rev1.2.pdf" rel="nofollow">https:&#x2F;&#x2F;www.hotchips.org&#x2F;hc30&#x2F;2conf&#x2F;2.13_Fujitsu_HC30.Fujits...</a>
akamoonknight将近 5 年前
Is there information on how the Fugaku machines are connected together? The highest performing Power9 ones seem to use InfiniBand, but is that still true with the ARM devices?<p>Edit: seems to be a Fujitsu designed interconnect [0]. Wonder how much of the overall performance is dependent on the difference in communication.<p><a href="https:&#x2F;&#x2F;www.fujitsu.com&#x2F;global&#x2F;documents&#x2F;solutions&#x2F;business-technology&#x2F;tc&#x2F;catalog&#x2F;08514929.pdf" rel="nofollow">https:&#x2F;&#x2F;www.fujitsu.com&#x2F;global&#x2F;documents&#x2F;solutions&#x2F;business-...</a>
评论 #23603712 未加载
评论 #23603872 未加载
gok将近 5 年前
It&#x27;s kind of interesting that in terms of perf&#x2F;watt, it&#x27;s actually slightly less efficient than Summit, which is over 2 years old. Also interesting that they went with a homogenous design (all Arm cores) instead of a heterogenous CPU+GPU setup.
sadfev将近 5 年前
Can any of the HPC experts shed some light on how these ARM chips are better than their predecessors. I toured a small cluster in LANL, where the ARM chips ran the hottest and their cooling was the loudest.
评论 #23601350 未加载
评论 #23603359 未加载
praveen9920将近 5 年前
Pushing the boundaries is the best way to advance technology<p>Car manufacturers need ridiculously expensive race cars to push the technology to get advancements in everyday car technology. Similarly, Top500 is one way to push the technology for not just computationally but also things like better power, heat and network management in processors and computers in general.<p>With ever doubling server farms, Heat management of these systems will become the major contributor for environmental pollution than all vehicle exhaust. Rather than spending on renewable sources of energy for these farms, it makes sense to optimize energy consumed per processor. Hoping to see advancements in this area<p>In my opinion, next arm&#x2F;intel will be the company who does energy efficient processors
评论 #23611021 未加载
评论 #23605520 未加载
评论 #23602197 未加载
flyGuyOnTheSly将近 5 年前
There&#x27;s something amusing about a blog post announcing the fastest computer in the world being unable to serve up web requests in under 10 seconds.<p>(When I wrote this comment, I was seeing 500 status code errors when trying to load the page)
评论 #23609725 未加载
d_tr将近 5 年前
For what it&#x27;s worth, Cray is also offering supercomputers with A64FX chips.
29athrowaway将近 5 年前
Some more information about Fugaku, aka Post-K, here:<p>- Slides on Fugaku: <a href="https:&#x2F;&#x2F;www.fujitsu.com&#x2F;global&#x2F;Images&#x2F;supercomputer-fugaku.pdf" rel="nofollow">https:&#x2F;&#x2F;www.fujitsu.com&#x2F;global&#x2F;Images&#x2F;supercomputer-fugaku.p...</a><p>- SVE 512 instructions for armv8: <a href="https:&#x2F;&#x2F;www.fujitsu.com&#x2F;global&#x2F;Images&#x2F;armv8-a-scalable-vector-extension-for-post-k.pdf" rel="nofollow">https:&#x2F;&#x2F;www.fujitsu.com&#x2F;global&#x2F;Images&#x2F;armv8-a-scalable-vecto...</a>
mxcrossb将近 5 年前
It’s sort of funny to watch the ISC event and see news of a machine with an exaop of AI performance, while the zoom presentation still can’t properly crop out the background.
hangonhn将近 5 年前
Anyone know the reason for the dominance of Power processors in the top 10 other than it&#x27;s from IBM and they get a lot of contracts for HPC?
评论 #23603576 未加载
评论 #23604686 未加载
fomine3将近 5 年前
Green500 #1 is MN-3 by PFN that&#x27;s also from Japan and they also use original chip!
评论 #23671916 未加载
cwaffles将近 5 年前
Backup link if you&#x27;re getting http 503: <a href="http:&#x2F;&#x2F;archive.is&#x2F;JSvCi" rel="nofollow">http:&#x2F;&#x2F;archive.is&#x2F;JSvCi</a>
IncRnd将近 5 年前
<a href="http:&#x2F;&#x2F;archive.is&#x2F;PNY80" rel="nofollow">http:&#x2F;&#x2F;archive.is&#x2F;PNY80</a>
cinntaile将近 5 年前
This kind of feels like a publicity stunt for Arm. Arm is owned by the Japanese company SoftBank and now Japan captures the supercomputer crown. I don&#x27;t want to take away from the achievement and it&#x27;s certainly possible that this is just a coincidence, maybe someone with more knowledge on the subject can comment on this?
评论 #23608758 未加载
Aaronstotle将近 5 年前
CoreTeks on youtube has a great video that explains the brilliance of this chip.
评论 #23623537 未加载
评论 #23604321 未加载
ausbah将近 5 年前
is x86 likely to ever go away, or anytime soon? asking as a non-systems guy
评论 #23609617 未加载
ClarkMills将近 5 年前
Ah but does it run Linux? [probably but couldn&#x27;t see in the link] Looks like another nail in Intel&#x27;s coffin...
mortenjorck将近 5 年前
I cannot think of any plausible way in which Apple could have influenced the date of this announcement, but the timing, given what is expected to be announced later today, is uncanny.
评论 #23603615 未加载
评论 #23602752 未加载
评论 #23604220 未加载
评论 #23604441 未加载
评论 #23602528 未加载
评论 #23603926 未加载
评论 #23602351 未加载
davidhyde将近 5 年前
The site of the organisation responsible for assessing the fastest computers in the world succumbs to the hacker news hug of death.
评论 #23601765 未加载
评论 #23601477 未加载
评论 #23603590 未加载
评论 #23601944 未加载
mtgx将近 5 年前
Also a king in efficiency:<p><a href="https:&#x2F;&#x2F;www.nextplatform.com&#x2F;2019&#x2F;11&#x2F;22&#x2F;arm-supercomputer-captures-the-energy-efficiency-crown&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.nextplatform.com&#x2F;2019&#x2F;11&#x2F;22&#x2F;arm-supercomputer-ca...</a>
评论 #23601954 未加载
UI_at_80x24将近 5 年前
I can&#x27;t help but think that the top minds from Cyrix aren&#x27;t feeling both smug at the vindication and dismayed that they were just a little too ahead of the curve.<p>The writing was on the wall that RISC would win, but the x86 juggernaut appeared unbeatable.
评论 #23601628 未加载
评论 #23603196 未加载
SomeoneFromCA将近 5 年前
I suggest building a supercomputer out of Ivy Bridge Celerons. They are dirty cheap, like $2 in bulk, yet quite performant.
评论 #23601748 未加载
评论 #23603032 未加载
xhkkffbf将近 5 年前
Not to be dismissive, but can&#x27;t anyone &quot;build&quot; the biggest supercomputer by reserving enough instances at AWS or GCP? I&#x27;m sure that AWS or GCP would like to encourage this competition, but it seems a bit, well, boring.
评论 #23601432 未加载
评论 #23601439 未加载
评论 #23601391 未加载
评论 #23601589 未加载
评论 #23601585 未加载
评论 #23603419 未加载
fizixer将近 5 年前
As I have said countless times in the past:<p>- Moore&#x27;s law is dead at the level of the transistor<p>- Architecture, HPC updates will keep coming for many years into the future<p>- AGI has already escaped Moore&#x27;s law (i.e., development of a fully functional AGI will not be constrained by lack of Moore&#x27;s law progress). And that&#x27;s what really matters.<p>- Related note on AGI: it has escaped the data problem as well (as in we have the right kind of sensors: mainly cameras, microphones, and so on). That is, according to the categorization of AGI challenges in terms of hardware, data, algorithms, the only missing piece is the right set of algorithms.
评论 #23602903 未加载
KKKKkkkk1将近 5 年前
We&#x27;re living in an age in which AWS dwarfs all the machines on the TOP500 taken together. The TOP500 is a vestige of the cold war that needs to be retired. Similarly to how the US and the USSR used to compare their numbers of nuclear warheads, it is comparing a reserve of capacity that is probably going to be retired having brought marginal benefits at best, all in order to goad taxpayers into a futile competition.
评论 #23601564 未加载
评论 #23602718 未加载
评论 #23602954 未加载
评论 #23602642 未加载