The fiftieth TOP500 list has been released. Although there has been little change at the top of the list, China now dominates the list in terms of the number of systems, rising to 202 from 160 in June, with the U.S. falling to 143 systems from 169. However, this seems to be the result of Chinese vendors pushing more commercial systems to get on the list:
An examination of the new systems China is adding to the list indicates concerted efforts by Chinese vendors Inspur, Lenovo, Sugon and more recently Huawei to benchmark loosely coupled Web/cloud systems that strain the definition of HPC. To wit, 68 out of the 96 systems that China introduced onto the latest list utilize 10G networking and none are deployed at research sites. The benchmarking of Internet and telecom systems for Top500 glory is not new. You can see similar fingerprints on the list (current and historical) from HPE and IBM, but China has doubled down. For comparison's sake, the US put 19 new systems on the list and eight of those rely on 10G networking. [...] Snell provided additional perspective: "What we're seeing is a concerted effort to list systems in China, particularly from China-based system vendors. The submission rules allow for what is essentially benchmarking by proxy. If Linpack is run and verified on one system, the result can be assumed for other systems of the same (or greater) configuration, so it's possible to put together concerted efforts to list more systems, whether out of a desire to show apparent market share, or simply for national pride."
Sunway TaihuLight continues to lead the list at just over 93 petaflops. The Gyoukou supercomputer has jumped from #69 (~1.677 petaflops) in the June list to #4 (~19.136 petaflops). Due to its use of PEZY "manycore" processors, Gyoukou is now the supercomputer with the highest number of cores in the list's history (19,860,000). The Trinity supercomputer has been upgraded with Xeon Phi processors, more than tripling the core count and bringing performance to ~14.137 petaflops (#7) from ~8.1 petaflops (#10). Each of the top 10 supercomputers now has a measured LINPACK performance of at least 10 petaflops.
The #100 system has an Rmax of 1.283 petaflops, up from 1.193 petaflops in June. The #500 system has an Rmax of 548.7 teraflops, up from 432.2 teraflops in June. 181 systems have a performance of at least 1 petaflops, up from 138 systems. The combined peformance of the top 500 systems is 845 petaflops, up from 749 petaflops.
Things are a little more interesting on the Green500 list. The Shoubu system B has an efficiency of 17.009 gigaflops per Watt, up from TSUBAME3.0's 14.11 GFLOPS/W at the #1 position in June (TSUBAME3.0 quadrupled its performance while its efficiency dipped to 13.704 GFLOPS/W (#6) on the new list). The top 4 systems all exceed 15 GFLOPS/W. #5 on the Green500 list is Gyoukou, which is #4 on the TOP500. Piz Daint is hanging in there at #10 on the Green500 list and #3 on the TOP500.
All of the new top 3 systems on the Green500 list (and Gyoukou at #5) use the PEZY-SC2 manycore processor. The SC2 has 2,048 cores and 8 threads per core, and has a single-precision peak performance of about 8.192 TFLOPS. Each SC2 also includes six MIPS management cores, making it possible to eliminate the need for an Intel Xeon host processor, although that has not been done in any of the new systems.
At 17 GFLOPS/W, it would take about 58.8 megawatts to power a 1 exaflops supercomputer. 20-25 MW is the preferred power level for initial exascale systems, although we may see a 40 MW system.
Previously: New List of TOP500 Supercomputers [Updated]
Related Stories
A new list was published on top500.org. It might be noteworthy that the NSA, Google, Amazon, Microsoft etc. are not submitting information to this list. Currently, the top two places are occupied by China, with a comfortable 400% head-start in peak-performance and 370% Rmax performance to the 3rd place (Switzerland). US appears on rank 4, Japan on rank 7, and Germany is not in the top ten at all.
All operating systems in the top-10 are Linux and derivates. It seems obvious that, since it is highly optimized hardware, only operating systems are viable which can be fine-tune (so, either open source or with vendor-support for such customizations). Still I would have thought that, since a lot of effort needs to be invested anyway, maybe other systems (BSD?) could be equally suited to the task.
The founder, President, and CEO of PEZY Computing, Motoaki Saito, has been arrested for allegedly defrauding the Japanese government:
The head of Japanese supercomputing firm PEZY Computing was arrested Tuesday on suspicion of defrauding a government institution of 431 million yen (~$3.8 million). According to reports in the Japanese press, PEZY founder, president and CEO Motoaki Saito and another PEZY employee, Daisuke Suzuki, are charged with profiting from padded claims they submitted to the New Energy and Industrial Technology Development Organization (NEDO).
On the 21st Green500 list, the top three most efficient supercomputers as well as the #5 most efficient supercomputer all use PEZY-SC2 "manycore" chips.
Previously: PEZY's Next Many-Core Chip Will Include a MIPS 64-Bit CPU
TOP500 Analysis Shows "Nothing Wrong with Moore's Law" and the November 2015 Green500 List
Shoubu Continues to Lead June 2016 Green500 List, World's Fastest Supercomputer Comes in at #3
The U.S. leads the June 2018 TOP500 list with a 122.3 petaflops system:
The TOP500 celebrates its 25th anniversary with a major shakeup at the top of the list. For the first time since November 2012, the US claims the most powerful supercomputer in the world, leading a significant turnover in which four of the five top systems were either new or substantially upgraded.
Summit, an IBM-built supercomputer now running at the Department of Energy's (DOE) Oak Ridge National Laboratory (ORNL), captured the number one spot with a performance of 122.3 petaflops on High Performance Linpack (HPL), the benchmark used to rank the TOP500 list. Summit has 4,356 nodes, each one equipped with two 22-core Power9 CPUs, and six NVIDIA Tesla V100 GPUs. The nodes are linked together with a Mellanox dual-rail EDR InfiniBand network.
[...] Sierra, a new system at the DOE's Lawrence Livermore National Laboratory took the number three spot, delivering 71.6 petaflops on HPL. Built by IBM, Sierra's architecture is quite similar to that of Summit, with each of its 4,320 nodes powered by two Power9 CPUs plus four NVIDIA Tesla V100 GPUs and using the same Mellanox EDR InfiniBand as the system interconnect.
The #100 system has an Rmax of 1.703 petaflops, up from 1.283 petaflops in November. The #500 system has an Rmax of 715.6 teraflops, up from 548.7 teraflops in June.
273 systems have a performance of at least 1 petaflops, up from 181 systems. The combined performance of the top 500 systems is 1.22 exaflops, up from 845 petaflops.
On the Green500 list, Shoubu system B's efficiency has been adjusted to 18.404 gigaflops per Watt from 17.009 GFLOPS/W. The Summit supercomputer, #1 on TOP500, debuts at #5 on the Green500 with 13.889 GFLOPS/W. Japan's AI Bridging Cloud Infrastructure (ABCI) supercomputer, #5 on TOP500 (19.88 petaflops Rmax), is #8 on the Green500 with 12.054 GFLOPS/W.
Previously: TOP500 List #50 and Green500 List #21: November 2017
(Score: 2) by pkrasimirov on Tuesday November 21 2017, @02:51PM (10 children)
https://www.top500.org/list/2017/11/ [top500.org]
National Supercomputing Center in Wuxi, China
Cores: 10,649,600
Power (kW): 15,371
Does that mean 15 MW of electric power??
(Score: 4, Informative) by takyon on Tuesday November 21 2017, @02:59PM (1 child)
Yes, ~15 megawatts. Should include cooling/etc. (didn't see anything in FAQ). This might be helpful:
https://www.top500.org/static/media/uploads/methodology-2.0rc1.pdf [top500.org]
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by pkrasimirov on Tuesday November 21 2017, @03:12PM
https://pics.me.me/hory-shet-25793140.png [pics.me.me]
(Score: 2) by DannyB on Tuesday November 21 2017, @03:09PM (7 children)
Nailed it in one Mr. Garibaldi.
15,371 kW = 15.371 MW = 15,371,000 W.
Now 15,371,000 W divided by
Cores 10,649,600
gives: 1.443340594951923076 with the 923076 repeating infinitely.
watts per core?
Can that even be write?
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 2) by pkrasimirov on Tuesday November 21 2017, @03:25PM (5 children)
I don't know, do you suggest that's too low consumption or too high? For too low, take a look at Japan's first:
Japan Agency for Marine-Earth Science and Technology
Gyoukou - ZettaScaler-2.2 HPC system, Xeon D-1571 16C 1.3GHz, Infiniband EDR, PEZY-SC2 700Mhz
ExaScaler
Cores: 19,860,000
Power (kW): 1,350
So that's 1,350,000 W / 19,860,000 cores ≈ 0,068 W/core
(Score: 2) by DannyB on Tuesday November 21 2017, @03:27PM (4 children)
I was thinking too low. But your second data point is helpful. Thanks.
Once you have a second data point, you should:
1. draw a straight line through the two points
2. jump up and down
3. yell out "we now understand this phenonema completely!"
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 1, Informative) by Anonymous Coward on Tuesday November 21 2017, @03:42PM (1 child)
You mean in order to maximize the errors? Correctly it would be "this phenomenon".
(Score: 1) by DannyB on Tuesday November 21 2017, @04:23PM
Thanks
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 2) by crafoo on Tuesday November 21 2017, @05:13PM (1 child)
yep. apply linear regression. extrapolate to infinity. problem solved. next please!
(Score: 2) by frojack on Tuesday November 21 2017, @07:25PM
Maybe we could submit the computation to an HPC somewhere?
No, you are mistaken. I've always had this sig.
(Score: 4, Informative) by takyon on Tuesday November 21 2017, @03:49PM
From the link to PEZY SC2 in the summary:
WikiChip [wikichip.org] says 180 W TDP, 130 W average dissipation.
130 Watts / 1984 cores = only 0.0655 Watts per core. The supercomputer also uses Intel Xeons which use a lot more Watts per core.
From the link about PEZY SC2, you can see that they plan to quadruple the core count in the next generation chip, to 8,192 cores, with efficiency rising to 40 GFLOPS/W. So this manycore chip could enable a 1 exaflops supercomputer in 2019-2020.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by DannyB on Tuesday November 21 2017, @02:55PM
So how many of the Top500 are running Docker?
Kubernetes?
Up in the air: who will win: Docker Swarm or Kubernetes?
Docker vs Rocket?
Vi vs Edlin vs Emacs?
Now please excuse me while I go order a couple more boxes of polygons. Also a few additional dimensions because Amazon is having a special if you order eight or more right before Bleak Friday!.
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 0) by Anonymous Coward on Tuesday November 21 2017, @05:04PM (2 children)
And how much does this kind of a rig cost to setup and then in electricity bill? Inquiring minds and all that.
I guess in reality they're used in something way more lucrative and clandestine that good ol BTC.
(Score: 3, Informative) by takyon on Tuesday November 21 2017, @06:15PM (1 child)
The U.S.'s most powerful supercomputers are used by the Dept. of Energy to play SimNuke.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by bob_super on Tuesday November 21 2017, @07:37PM
Which is disappointingly classified, so to keep Us Entertained, the White House is considering a return to Good Ol' atmospheric tests in the Far East.
(Score: 3, Informative) by LoRdTAW on Tuesday November 21 2017, @08:36PM (1 child)
That's a lot of juice to distribute is a small space.
I'm sure the utility company would have to run new sub transmission lines to whatever facility is using this much power. Even the higher primary medium voltages most utilities use for would require circuits of at least 670A @ 34.5kV. That much power would be difficult to distribute at low voltage throughout the building so small internal substations close to the racks reduces line losses and cost as more volts = less amps = less copper. So A big sub and switchgear for getting the utility power down to a reasonable distribution voltage, maybe 2400/4160 (4kV) and run that to smallish 250-500kVA distributions transformers feeding local panel boxes and right to a cluster of server racks. A 500kVA transformer can supply a ~1300A 120/208V three phase and would only require 70A @ 4kV from a single 33mm diameter cable. Whereas the 120V 1300A service requires at least four 23mm 500MCM cables *per phase*. Thats a lot of copper.
It also brings up a good question: Are there computer power supplies which operate at higher line voltages, e.g. 400V in Europe or 480V in NA? That would lower the install cost as 120V vs 480V is a 4x reduction in current and moving to three phase provides almost 7 times more power @ 480V vs 120V single phase. A single 20A 480V three phase circuit will provide 16.6kVA vs 2.4kVA for the same sized 120V single phase circuit or 7.2kVA for a 120/208V three phase circuit. I don't see why specialized data centers can't move to higher voltages for the power supplies. If you're worried about safety, don't so long as properly designed wiring devices are used. I did a bit of googling and only found an article from Cray for using 480V and another using the single phase 277V of a 480V wye.
(Score: 2) by takyon on Tuesday November 21 2017, @11:40PM
Honestly, if we believe this, 25 MW can be reached instead (see table above the following paragraph in the article):
https://fuse.wikichip.org/news/191/the-2048-core-pezy-sc2-sets-a-green500-record/ [wikichip.org]
PEZY-SC4 would enable a 17 MW exaflops supercomputer. And the power-hungry Intel Xeons can be ditched (not sure if they will).
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]