• We've upgraded our forums. Please post any issues/requests in this thread.

Intel to crush Titan Supercomputer with Tianhe 2

Joined
Apr 30, 2012
Messages
2,419 (1.18/day)
Likes
1,333
#1
Incase you havent been hearing the buzz leading up to ISC. Chinese are building Tianhe 2 Supercomputer

The reported performance (we have four highly credible sources confirming) is between 53-55 peak and between 27-29 LINPACK sustained performance. This is actually better than we were led to believe yesterday when it felt a little wrong to make the "50 petaflop" claim, despite our best sources telling us it was so
This is a MIC-only system from what we understand—no GPU acceleration thrown in for added fire, which was what gave the Tianhe-1A its chart-topping power.
Source: HPCWire
 
Joined
Dec 14, 2009
Messages
6,591 (2.25/day)
Likes
5,825
Location
Glasgow - home of formal profanity
System Name New Ho'Ryzen
Processor Ryzen 1700X @ 3.82Ghz
Motherboard Asus Crosshair VI Hero
Cooling TR Le Grand Macho & custom GPU loop
Memory 16Gb G.Skill 3200 RGB
Video Card(s) GTX1080ti (Heatkiller WB) @ 2Ghz core/1.5(12)Ghz mem
Storage Samsumg 960 Pro m2. 512Gb
Display(s) Dell Ultrasharp 27" (2560x1440)
Case Lian Li PC-V33WX
Audio Device(s) On Board
Power Supply Seasonic Prime TItanium 850
Software W10
Benchmark Scores Look, it's a Ryzen on air........ What's the point?
#2
Intels Xeon Phi in action. CD at S|A said this a long way back that Phi (having the ancestry of Larabee) was actually going to really hurt Nvidia, not so much as it is better but it uses a far more common instruction set that many more people can readily use (than say CUDA).
 
Joined
Apr 30, 2012
Messages
2,419 (1.18/day)
Likes
1,333
#3
Last year Beacon running Intel with Xeon Phi got #1 on Green500

The chinese just decided to do it bigger. Tianhe 2 will be x2 faster then Titan. It will run faster at sustained then Titan does at peak. :eek:

As for the instruction set. During the Titan presentation it was mentioned that same thing that it was capable of doing OpenCL/OpenMP which people prefer since people who will be using it wont always be on CUDA and they wont have to spend extra time debug CUDA and it runs just as fast.