Intel-Powered Aurora Supercomputer Breaks The Exascale Barrier

[qarium]

???
Rather RISC CPU like MIPS, SunSPARC, DEC Alpha, PowerPC lost with Intel Pentium.
And AMD lost GPU battle with Nvidia.
You can look at it this way: this is Intel's first approach to high-performance GPUs. The GPU war for Intel is just beginning. The results will be known in 10-15 years.
For the time being, the winning company is Nvida, which has about 96% share of the high-performance GPU market and has profits from its DataCenter division 20 times greater than AMD's - reflecting market capitalization.

well all these architectures "MIPS, SunSPARC, DEC Alpha," are death...
MIPS only continue to exist because of chinese Longson cpus and they even renamed the architecture.
PowerPC only continue to exist in a very niche market.

"Intel Pentium" so hysterically we can say intel won big in this battle...

intel is only in trouble because AMD build better x86 systems than intel.

and on GPU side newer benchmarks show an intel ARC 770 can beat a 7600XT in raytracing but the lead is very small. but then people just buy a radeon 7800XT and well intel lost the sale. it is that simple.

Nvidia "has profits from its DataCenter division 20 times greater than AMD's"

well yes... but who cares ?

[qarium]

Of course. But it's AMD fanboys write:
When Intel still has higher sales in the laptop (80% of the market), desktop (76%) and server (also 76%) segments (- the data comes from the Mercury report).
And just started to enter the GPU market. And it has practically the same position in Data Center GPU as AMD, which has been in this market from the very beginning .
The only loser is AMD in Data Center GPU, because it has practically zero share.
Sure. E.g. DEC Alpha was even better than Intel Pentium. You can say that over its grave.

"Of course. But it's AMD fanboys write:
"intel clearly lost the chip war.... the last time someone lost the chip war was in 1980 the UDSSR lost the chip war.""

i am not a AMD fanboy i also have a Intel ARC A380 around... but i can not take them seriously in the GPU market.
that intel won the chip war in the 1980 er years against the UDSSR is fact... they did not only won it against the UDSSR they won against any competitiot in the west to.

"The only loser is AMD in Data Center GPU, because it has practically zero share."

this is true but it is historical ... its historical because what exactly stops you to buy AMD and just use zluda ?

zluda even performs better than ROCm/HIP native..

[qarium]

• inaccurate weather forecast
• mass surveillance of innocent citizens
• cracking encryption
• less atomic bomb tests in your neighborhood

Plus they help predict the exact date of human extinction.

yes ... you are right ;-)

[coder]

What's the benefit of these super computers? Show off?

According to the Frontier web page, research applications include:

• Fusion energy
• Material science
• Low-temperature, reactivity-controlled compression ignition engine
• Microstructure properties of additively-manufactured parts
• Cosmological simulations
• Optimization of scientific code for GPUs
• Computational modeling of biological systems
• Particle physics
• Electronic properties of complex molecules
• Reconstruction of biological structures scanned using a Linac Coherent Light Source
• Gravitational waves generated by supernovas
• Biomass fuels

There's a lot more on the ALCF page (where Aurora is located), but I'm too lazy to summarize it for you. Best to browse it yourself.

https://www.alcf.anl.gov/science

[coder]

• ...
• mass surveillance of innocent citizens
• cracking encryption

Plus they help predict the exact date of human extinction.

Neither of those.

Sounds like someone is jealous!

[coder]

AMD could easily compete in the GPU market if they wanted to, they don't want to.

In fact, they very much are. Compare the specs of MI300 vs. H200 (note: you can't buy Blackwell, yet).

Competing effectively in the GPU powered HPC and AI markets would be in direct conflict with their obvious strategy of competing in those markets with more CPU cores.

This point just shows you don't know what the hell you're talking about, because CPU performance is no match for GPUs on this front!

If it were true that CPUs could even come close to matching GPUs, there's no way Intel would've killed Xeon Phi and made such a risky and expensive bet on Ponte Vecchio!

[Anux]

Sounds like someone is jealous!

Jealous? Of what? You know that this wasn't a serious reply? Anyone can google the applications of HPC clusters in a second.