AMD EPYC 9965 "Turin Dense" Delivers Better Performance/Power Efficiency vs. AmpereOne 192-Core ARM CPU

[oleid]

The AmpereOne CPUs aren't optimized for high performance. They are good for running web servers, let's say. Not for LULESH, HPCG, GROMACS, OpenFOAM...

They are not optimized for performance, true. But it would seem Epic beats them in efficiency as well.

[filbo]

Michael
How can the joules per run be lower while the average watts are higher for Epic?

He's measuring how much energy it takes to complete a particular unit of work. If one CPU takes 100s to build a complete Linux kernel with all modules, running at 300W; and another takes 200s at 200W -- the first one uses 75% as much energy. These aren't things you do continuously; rather, sporadically, as needed.

[oleid]

He's measuring how much energy it takes to complete a particular unit of work. If one CPU takes 100s to build a complete Linux kernel with all modules, running at 300W; and another takes 200s at 200W -- the first one uses 75% as much energy. These aren't things you do continuously; rather, sporadically, as needed.

That's why I was talking about the average power consumption. But you're right, I somehow assumed that the time to complete the task would be similar. Don't post before the first coffee 😅

[Anux]

There is often much chatter online that likes to imply ARM is inherently more power efficient or similar.

Yes chatter there is but always failing to present reasons and logic.

I'm thinking if the 100W idle problem has any influence on the tests here as they are mostly full load scenarios. Anyway if your server is idling you have already done something wrong.

[Weasel]

Ampere was founded in 2017 and today they're producing ARM processors which aren't far away from catching up to AMD's flagship CPU

Things are slowly taking shape with ARM performance in Linux continuously improving and more distros enhancing support for it

Really can't wait to buy a ARM powered workstation in the future

Found the sore loser fanboy shill.

[Weasel]

That's why I was talking about the average power consumption. But you're right, I somehow assumed that the time to complete the task would be similar. Don't post before the first coffee 😅

Imagine starting with the assumption ARM can compete at performance (how fast it finishes tasks), lmfao.

It's heresy to do this even while drunk.

[oleid]

Imagine starting with the assumption ARM can compete at performance (how fast it finishes tasks), lmfao.

It's heresy to do this even while drunk.

Ah, the auxiliary trolls are active again.

[Kjell]

Found the sore loser fanboy shill.

Found the troll

[Weasel]

Found the troll

Except everything I said was pure fact.

[Weasel]

Ah, the auxiliary trolls are active again.

And yet, had you been under my mentality, you would have never made such an embarrassing mistake.

Keep thinking I'm the troll and not you.