TOP500 at ISC’26: We have a New Number 1 Supercomputer

(chipsandcheese.com)

90 points | by rbanffy 9 hours ago

16 comments

  • brianolson 8 hours ago
    > Why aren’t these AI companies submitting to the TOP500 to show off their computing prowess?

    my knowledge is 10+ years out of date, but once upon a time if they'd chosen to, Google could have had _several_ entries in the top 10 of the TOP500 list

    It's just poker, they didn't want to tip their hand

    • davidmr 7 hours ago
      I’ve worked on several systems that had enough flop/s to make it in the top 5-10, but for which we never submitted benchmarks. Sometimes their backend network layout technically would make them several smaller clusters for an HPL run, sometimes it’s because the cluster is too heterogeneous to get a good benchmark result, and sometimes it’s because the employer wants to keep a low profile.

      Most of the time, it just that it’s a hassle. It takes a while to prep and tune a big hero run for benchmarking, and if you spend a billion dollars on a cluster, it’s making you a lot more than that. Taking it down for a day or two stops the money printers.

    • ziofill 8 hours ago
      Also, would those 550k Blackwell have good FP64 performance? How would one even compare them?
      • NitpickLawyer 1 hour ago
        Yeah, that's true. GPU flops are really impressive in fp16, and more recently fp8 and fp4. When the 40xx GPUs came out, Tim Dettmers had a really cool blog looking into the numbers, and for fp8 one 4090 GPU had enough flops to match the best supercomputer somewhere in the 2000s, with an 8x 4090 build being top for a few years more. It's insane. But it gets nowhere close on fp64, where most of the physics simulations and other usual supercomputer tasks shine.
    • JumpCrisscross 3 hours ago
      Is there international value to these designations? As in, would it be worth it for the U.S. to pay a bonus to anyone who qualifies into the TOP500, to offset the cost of the run?
    • ls612 6 hours ago
      Why would the scientific computing people want to tip their hand? It’s an open secret that the main point of these mammoth FP64 compute machines is to simulate nuclear weapons detonations to comply with the CTBT you’d think that crowd would really not be fans of broadcasting their capabilities.
      • kube-system 4 hours ago
        In adversarial scenarios, there are varying strategies in communicating one's capabilities, just as one might do in a poker game.

        Sometimes you want to show off what you can do to dissuade others from fucking with you. Sometimes you want to undersell your capabilities to hide your true ability. Sometimes you want others to think you are underselling your capabilities when you are actually at a disadvantage.

      • mrlongroots 1 hour ago
        It is partly this and partly a funding vehicle for American next-gen computing. It is not that hard to estimate FP64 ballpark from a whole bunch of public statistics. And it takes a looot more than raw FLOPs to get a simulation working. And presumably a looot more to translate it into practice. And the openness makes it easier to talk to different vendors and not get in the way of them having all the H1Bs it takes to get these things to work.

        Plus one think I like to say is that if a bullet is flying towards you, you could know everything about the chemistry of the gunpowder and the composition of the alloy without it affecting what happens next.

      • dopa42365 3 hours ago
        What for? You only need to match the performance that existed in the 1950s. In the Soviet Union. Everything else is a lack of knowledge rather than computing power.

        Also you should read the second sentence of the CTBT Wikipedia article to find out why it's not even in force (spoiler: US hasn't ratified it).

      • Onavo 3 hours ago
        At some point, you will get diminishing returns no? I don't think compute is the bottleneck right now for mechanical engineering if you don't count AI.
    • iberator 8 hours ago
      Cloud computing is not a supercomputer. Different architecture, bandwitch, interconnectivity and latencies.
      • dgacmu 8 hours ago
        That's not nearly as true when you look at AI training clusters. They're basically supercomputers but without an FP64 focus.

        (These are the systems to which GP was referring at Google.)

        • cynicalkane 7 hours ago
          Even before AI training clusters became important, Google has had an outstanding custom fabric (there's papers about it) together with the ability to tune NICs for their own cases, and "their own cases" meant nearly everything engineered within Google. Ethernet hardware has had low kernel latency and DMA for a long time; it's the rest of the stack that hurts. But as far back as the early 2010s (if not further back, that goes beyond my knowledge horizon), you could just make it not hurt, if you had the software engineers to do it.
          • mrlongroots 1 hour ago
            All that would not help you with an AI training cluster interconnect. See Amin Vahdat's keynote at HotInterconnects 2025. Everyone is building a fabric for this stuff from scratch (Google/Falcon, Amazon/EFA, Azure/MANA, Cornelis/CN5000, and obviously Mellanox).
        • jeffbee 8 hours ago
          I thought TPUs couldn't reasonably run LINPACK at all because TPUs do not acknowledge that FP64 exists.

          I know Google wants to compare their stuff to El Capitan or whatever but the comparison does not seem valid to me.

      • wmf 8 hours ago
        Historically there have been a bunch of clusters on the Top 500 that weren't used for HPC. The tell is that they used Ethernet (this was before RoCE). It's less efficient but you can still get an OK Linpack score.
  • adrian_b 52 minutes ago
  • jandrewrogers 8 hours ago
    TOP500 hasn't been a particularly useful measure of practical computing power in modern systems for many years because what it measures isn't a significant bottleneck in most real systems. It has become a measure of how much money someone is willing to spend for bragging rights. (HPCG is better in that it is a bit more bandwidth focused but still pretty narrow.)

    Most companies with huge systems don't participate.

    • adrian_b 34 minutes ago
      This seems like a "sour grapes" comment.

      The new Chinese supercomputer beats all US supercomputers also in HPCG, not only in Linpack.

      What is remarkable is that this was done despite the US attempts of sabotaging HPC in China by "sanctions".

      This uses custom CPUs designed in China, which implement an Armv9-A ISA with SME (scalable matrix extension) and which use fast HBM memory. These CPUs are fast enough that they do not need any GPUs for exceeding the throughput of the American supercomputers, which use GPUs. This is like in the Japanese Fugaku, which was the first to implement the Armv8-A ISA with SVE, but which now is rather old.

      Like in all CPU-based supercomputers, for this new Chinese supercomputer it is much easier to reach a higher percentage of the theoretical maximum throughput, when solving any problem. So for most practical problems it will be faster than a GPU-based supercomputer that would have the same theoretical maximum throughput.

      So this is a much more interesting supercomputer than those built by just buying some HPC racks from HPE (Cray). Because China was forbidden to buy the American equipment, they had to innovate and design their own. Eventually they made something better than what they could not buy.

    • bee_rider 7 hours ago
      I wonder if there would have been an opportunity to generate some finer-grained benchmarks with something like BiCGStab+ILU (or maybe CG+incomplete cholesky). Instead of CG+Gauss Seidel. The pitch being, you might have made different memory vs compute trade-offs with designing your cluster, but you should be able to select a fill-in factor for the preconditioner to suit it.
  • flopsamjetsam 8 hours ago
    > We think it is highly likely that these LX2 chiplets are etched using SMIC 7 nanometer processes at the N+3 refinement, and we base that on the fact that the chip only runs at 1.55 GHz. That is nowhere near the 3 GHz that SMIC can push with that process, but it is probably lower to get the memory and core speeds more balanced. [1]

    Based on the ARMv9.2.

    [1] https://www.nextplatform.com/hpc/2026/06/25/a-deep-dive-on-c...

  • chrisss395 5 hours ago
    I haven't kept up with the latest on supercomputing power, but I recall some years ago there being strong evidence that China had a couple of un-announced supercomputers that would have topped the charts. It makes me wonder what is publicly disclosed vs. actual.
  • mlmonkey 2 hours ago
    You can see the list yourself at https://top500.org/

    Deep link: https://top500.org/lists/top500/list/2026/06/

  • Retr0id 7 hours ago
    Interesting to see PAC mentioned on the slide, I'd have assumed security features would be a waste of transistors on something so compute-optimized - but maybe they want to isolate workloads from each other?
    • fragmede 1 hour ago
      Yes exactly. it means you can sell isolated work units. With cloud being, well, cloud, if you can do that, you then have a market, with bidders able to bid higher for more capacity right now, vs a lower bid for eventually.
  • 2OEH8eoCRo0 9 hours ago
    Extremely impressive accomplishment considering they did this with Chinese interconnects and Chinese chips. This is a wake up call.
    • jandrewrogers 8 hours ago
      TOP500 can be done with inexpensive silicon. It is more about a willingness to aggregate enough hardware in one place. As a benchmark, it tells you almost nothing about computing power or scalability for other applications because it doesn't exercise the bottlenecks most high-scale applications have.
      • NitpickLawyer 1 hour ago
        > TOP500 can be done with inexpensive silicon.

        Didn't the DoD at one point build a 1k+ PS3 cluster based on their multi-core chip and had a mini supercomputer CotS?

        I remember Sony not liking that people were buying them for other things rather than gaming (iirc they were losing money on hardware at the time) so they bricked linux support soon after.

        • packetslave 47 minutes ago
          > Didn't the DoD at one point build a 1k+ PS3 cluster based on their multi-core chip and had a mini supercomputer CotS?

          The Air Force did (Condor) and it hit #33 on the 2010 Top500.

    • echelon 8 hours ago
      We're too busy regulating the tech, not granting access to US engineers and companies, arguing against power and data centers, stopping skilled immigration.

      This is absolutely going to bite us in the face in five to ten years.

      • 2OEH8eoCRo0 8 hours ago
        Separate issue that has nothing to do with US manufacturing or HPC. I think our retreat from science funding and offshoring advanced manufacturing is a bigger issue.
  • b33f 6 hours ago
    Why are they not using GPUs? is it use cases that don't suit GPUs or because of the limitations they are imposing on themselves to use SMIC domestic chips?
    • wmf 4 hours ago
      I suspect Chinese GPUs (e.g. Biren) are not mature.
    • amelius 5 hours ago
      GPUs are for graphics (the G in GPU). These systems are used for more general computations.
      • galaxy_quest 4 hours ago
        I’m not sure if I’m missing a joke, but that’s why we have general purpose computing on graphics processing units (GPGPU) which is why 8/10 of the top 10 machines have GPUs.
      • antonvs 2 hours ago
        GPUs were for graphics. Now, they're mainly used for machine learning training and inference. The big tech companies are spending eye-watering amounts for GPUs - hundreds of billions of dollars a year each. That's the reason that Nvidia's market cap is at $4.66 trillion.
  • ziofill 8 hours ago
    > Two cores are disabled per cluster.

    I’m sure there is a good reason for this, which is..?

    • jandrewrogers 8 hours ago
      It is likely that those cores are dedicated to unrelated management, monitoring, and administrative tasks. This is common and many workloads are throttled on bandwidth anyway. For the purposes of the benchmark, those cores are not participating in the workload.
    • brianolson 7 hours ago
      Yield. Some fraction of cores had a speck of dust or something, but at 38/40 good cores per chip they got economical yield
      • tjhei 7 hours ago
        And then even if some nodes had 40/40 "good" cores, it would make load balancing a lot more complicated if core counts vary. Easier to turn them off at the hardware level.
      • dist-epoch 7 hours ago
        Couldn't some chips have 40 good cores, while others have only 36? Do they all need to be exactly 38?
  • rippeltippel 9 hours ago
  • lokimedes 9 hours ago
    Would the AI “GW-scale” clusters be able to run the Top500 benchmarks meaningfully? And what might be the outcome?
    • adrian_b 13 minutes ago
      No.

      The AI oriented GPUs or TPUs have either weak FP64 throughput or they may not support FP64 at all.

      They can compete neither with CPUs nor with GPUs that have good FP64 support, like the AMD CDNA datacenter GPUs, which occupy all the top places among American supercomputers.

      NVIDIA has stopped improving the FP64 throughput even in their "datacenter" GPUs, abandoning this nowadays smaller market to AMD.

      The AMD CDNA GPUs can be used for both HPC and AI, so only an AI cluster based on them could have dual use, but most who want AI choose NVIDIA.

    • mrlongroots 1 hour ago
      > And what might be the outcome?

      DoE compute budgets are ~10B USD across labs. AI training is a trillion-dollar workload. Different league.

    • wmf 7 hours ago
      Yes, they should score well on Linpack as long as they use Ozaki emulation.
      • adrian_b 17 minutes ago
        No, that is too slow.

        Most claims about the cost of emulating FP64 on GPUs are wrong, because they assume that only the significand of floating-point numbers must be extended.

        In reality it is even more important to extend the exponent, because with the exponent of FP32 overflows would be much too frequent in scientific/technical computations to accomplish anything.

        The minimum FP64 emulation on FP32-capable GPUs requires 3 numbers per emulated FP64, which may be 3 FP32 numbers, or the exponent may be an Int32, if that works better on the target GPU. An emulated FP64 operation is likely to be at least 20 times slower than a FP32 operation.

        That is much faster than the 1:64 ratio provided in hardware by an NVIDIA GPU, but even on the fastest FP32 GPUs it is too slow to compete with CPUs, in a professional setting.

        FP64 emulation on a GPU can be useful only in a home computer, which may have a rather weak CPU and increasing the FP64 throughput using the GPU can be done at no additional cost, so it can be worthwhile.

  • dgellow 8 hours ago
    Just glad to see Hamburg mentioned :) Hope you all didn’t suffer too much through the current heatwave
  • techsystems 9 hours ago
    Is it the first to reach 2 exaflops?
  • amelius 7 hours ago
    How many tokens/s? :)
  • TacticalCoder 1 hour ago
    And you know it needs a well-conceived OS to be able to run it: it's got to be... Microsoft Windows right?

    The OS powering 0% of the 500 supercomputers of the Top 500. But this time, it has to be Windows, right? Amirite?

    Ah, no, just kidding: it's "Kylin OS". It used to be a BSD derivative and now it's just based on Linux.

    I know, I know: "It's a heavily modified Linux". Whatever, it's not Windows and that makes me very happy.