Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Arm's Cortex X925: Reaching Desktop Performance

chipsandcheese.com

Post reply on HN
Print view View original

41–50 of 169 posts

Re: Arm's Cortex X925: Reaching Desktop Performance

#41

Post by Supersaiyan_IV »

Another good read is about ARM's SVE2 extensions: https://gist.github.com/zingaburga/805669eb891c820bd220418ee...

It has some interesting conclusions, such as that it covers certain AVX512 gaps:

"AVX512 plugs many of the holes that SSE had, whilst SVE2 adds more complex operations (such as histogramming and bit permutation), and even introduces new ‘gaps’ (such as 32/64-bit element only COMPACT, no general vector byte left-shift, non-universal predication etc)."

And also that rusty x86 developers might face skill issues:

"Depending on your application, writing code for SVE2 can bring about new challenges. In particular, tailoring fixed-width problems and swizzling data around vectors may become much more difficult when the length is unknown."

Re: Arm's Cortex X925: Reaching Desktop Performance

#42

Post by ezst »

flembat wrote: post #16
ezst wrote: post #10

Earlier quoted context omitted.

That would only matter (to me, at least) if those Apple chips were propping up an open platform that suits my needs. As things stand today, procuring an M chip represents a commitment to the Apple software ecosystem, which Apple made abundantly clear doesn't optimize for user needs. Those marginally faster CPU cycles happen on a time scale that anyway can't offset the wasted time fighting MacOS and re-building decade…

When purchasing any ARM based computer a key question for me, is how many of those can I purchase for the cost of a Mac mini, and how many Mac mini can I purchase for the cost of that, and does that have working drivers...

And the answer there may absolutely be "none", which equates to doing away with ARM, which is totally fine. I don't have a horse in the x86 vs ARM race, especially since it's pretty clear that performance per watt stands within a narrow margin across arches on recent nodes.

Re: Arm's Cortex X925: Reaching Desktop Performance

#43

Post by galangalalgol »

silon42 wrote: post #19
jordiburgos wrote: post #17

Earlier quoted context omitted.

Only for the hand-written assemply parts of the source code. The rest will be handled by the compilers.

Not even close. Except maybe in Rust /s

For rustaceans missing that /s, if you just use Relaxed ordering everywhere and you aren't sure why, but hey tests pass on x86, then yeah on arm it may have a problem. On x86 it effectively is SeqCst even if you specify Relaxed.

Re: Arm's Cortex X925: Reaching Desktop Performance

#44

Post by rick_dalton »

amelius wrote: post #34
pdpi wrote: post #14

Earlier quoted context omitted.

Sure. Insofar as Apple Silicon beats these things, "I'll take less powerful hardware if it means I'm not stuck with the Apple ecosystem" is a perfectly reasonable tradeoff to make. Two things, though. First, I don't like making blind tradeoffs. If what I need (for whatever reason) is a really beefy ARM CPU, I'd like to know what the "Apple-less tax" costs me (if anything!) Second, the status quo is that Apple Silicon…

Let's say my company makes systems for in-flight entertainment, with content from my company. I am looking for a CPU. I don't want to confront my users with "Please enter your Apple ID" or any other unexpected messages that I have no control over. Is Apple M series an option for me?

The X925 core is used in chips like the gb10 for the nvidia dgx spark. So it is relevant to compare to apple silicon performance imo. The mac studio is pretty much a competitor to it.

Re: Arm's Cortex X925: Reaching Desktop Performance

#45

Post by bluGill »

pdpi wrote: post #14
ezst wrote: post #10

Earlier quoted context omitted.

That would only matter (to me, at least) if those Apple chips were propping up an open platform that suits my needs. As things stand today, procuring an M chip represents a commitment to the Apple software ecosystem, which Apple made abundantly clear doesn't optimize for user needs. Those marginally faster CPU cycles happen on a time scale that anyway can't offset the wasted time fighting MacOS and re-building decade…

Sure. Insofar as Apple Silicon beats these things, "I'll take less powerful hardware if it means I'm not stuck with the Apple ecosystem" is a perfectly reasonable tradeoff to make. Two things, though. First, I don't like making blind tradeoffs. If what I need (for whatever reason) is a really beefy ARM CPU, I'd like to know what the "Apple-less tax" costs me (if anything!) Second, the status quo is that Apple Silicon…

Why do you need ARM? There is nothing magic, most CPUs are an internal instruction set with a decoder on top. bad as x86 is, decoding is not the issue. they can make lower power use x86 if they want. They can also make mips or riskv chips that are good.

Re: Arm's Cortex X925: Reaching Desktop Performance

#46

Post by galangalalgol »

DeathArrow wrote: post #29
adrian_b wrote: post #25

Earlier quoted context omitted.

ARM Cortex-X925 achieves indeed a very good IPC, but it has competitive performance only in general-purpose applications that cannot benefit from using array operations (i.e. the vector instructions and registers). The results shown in the parent article for the integer tests of SPEC CPU2017 are probably representative for Cortex-X925 when running this kind of applications. While the parent article shows AMD Zen 5 ha…

Still, what percentage of software uses AVX512 for its core functionality, so vector performance matters in practice?

Auto vectorizing optimizers have gotten quite good. If you are using integers it often just happens whether you think about it or not. With floats unless you specify fast math you will need to use wide types to let it know you don't care about floating point addition order.

Re: Arm's Cortex X925: Reaching Desktop Performance

#47

Post by promiseofbeans »

atwrk wrote: post #21
pdpi wrote: post #2

Kind of weird to see an article about high-performance ARM cores without a single reference to Apple or how this hardware compares to M4 or M5 cores.

Those are of almost zero use for people wishing to run Linux etc. Yes, Asahi exists, and props to the developers, but I don't think I'm alone in being unwilling to buy hardware from a manufacturer who obviously is not interested in supporting open operating systems

I mean… Apple went out of their way to build a GUI OS picker that supports custom names and icons into their boot loader.

So they don’t actively help (or event make it easy by providing clear docs), but they do still do enough to enable really motivated people

Re: Arm's Cortex X925: Reaching Desktop Performance

#48

Post by pdpi »

bluGill wrote: post #45
pdpi wrote: post #14

Earlier quoted context omitted.

Sure. Insofar as Apple Silicon beats these things, "I'll take less powerful hardware if it means I'm not stuck with the Apple ecosystem" is a perfectly reasonable tradeoff to make. Two things, though. First, I don't like making blind tradeoffs. If what I need (for whatever reason) is a really beefy ARM CPU, I'd like to know what the "Apple-less tax" costs me (if anything!) Second, the status quo is that Apple Silicon…

Why do you need ARM? There is nothing magic, most CPUs are an internal instruction set with a decoder on top. bad as x86 is, decoding is not the issue. they can make lower power use x86 if they want. They can also make mips or riskv chips that are good.

There's nothing special about ARM, sure. Hence "for whatever reason". Still, ARM is a known quantity, and the leading alternative to x86 for desktop CPUs. The article is titled "reaching desktop performance".

We know how Apple's hardware performs on native workloads. We know how it performs emulating x86 workloads (and why). Surely "... and this is how this hardware measures up against the other guys trying to achieve the exact same thing" is a relevant comparison? I can't be the only person who reads "reaching desktop performance" and wonders "you mean comparable to the M1, or to the M3 Ultra?"

Re: Arm's Cortex X925: Reaching Desktop Performance

#49

Post by upcoming-sesame »

ezst wrote: post #10
pdpi wrote: post #2

Kind of weird to see an article about high-performance ARM cores without a single reference to Apple or how this hardware compares to M4 or M5 cores.

That would only matter (to me, at least) if those Apple chips were propping up an open platform that suits my needs. As things stand today, procuring an M chip represents a commitment to the Apple software ecosystem, which Apple made abundantly clear doesn't optimize for user needs. Those marginally faster CPU cycles happen on a time scale that anyway can't offset the wasted time fighting MacOS and re-building decade…

still matters as a benchmark imo

Re: Arm's Cortex X925: Reaching Desktop Performance

#50

Post by Zardoz84 »

dinglo wrote: post #6

If ARM starts dominating in desktop and laptop spaces with a quite different set of applications, might we start seeing more software bugs around race conditions? Caused by developers writing software with X86 in mind, with its differing constraints on memory ordering.

If it is programmed in assembly. This kind of nasty detail should be handled by the compilers.
Post reply on HN