• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion ARM Cortex/Neoverse IP + SoCs (no custom cores) Discussion

Page 33 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
FlameTail said:
ARM surpasses the market cap of Intel

https://twitter.com/x/status/1796544544286412997

The symbolism is deep.
Click to expand...
Take a long, hard look at the insanity of Tesla's market practices and compare it to their even more insane market cap.

The stock market really isn't a viable barometer for this stuff.

IMHO it's a foundation for the biggest and longest financial con game in the history of modern civilisation.
 
soresu said:
Take a long, hard look at the insanity of Tesla's market practices and compare it to their even more insane market cap.

The stock market really isn't a viable barometer for this stuff.

IMHO it's a foundation for the biggest and longest financial con game in the history of modern civilisation.
Click to expand...
Rather look at the balance sheet and financials
 
Is Cortex X925 a clean sheet/ground up new design?
Screenshot_20240530_194108_Gallery.jpg
With the 2024 IP, the X core's codename is no longer related to the A7xx core.
 
Based on the Software Optimization Guide for X925.

Instructions are first fetched and then decoded into internal Macro-OPerations (MOPs). From there,
the MOPs proceed through register renaming and dispatch stages. A MOP can be split into two
Micro-OPerations (µOPs) further down the pipeline after the decode stage. Once dispatched, µOPs
wait for their operands and issue out-of-order to one of 23 issue pipelines. Each issue pipeline can
accept one µOP per cycle.
Click to expand...

The dispatch stage can process up to 10 MOPs per cycle and dispatch up to 20 µOPs per cycle
Click to expand...
 
That is certainly an ambitious design! Maybe the plan is to sort of "grow into it" over a few iterations, so they have a list of bottlenecks on X5 and they cross a few off the list with X6, and so forth.

Alternatively it is too ambitious, and some of the bottlenecks simply can't be fixed because they are inherent to a design so wide. Guess we'll find out.
 
soresu said:
Yeah ironically it sounds like A725 has more interesting features than X925 despite being on the lower performance tier.
Click to expand...
Yeah, A725 looks pretty good after A710/715 and kinda A720 were mid.
Especially A710, that was less efficient than A78 which is rofl.

Now somebody please taze ARM into making a proper OoO LITTLE.
Very long overdue.
 
soresu said:
A7xx is effectively becoming the Little core based on ARM's own PR.
Click to expand...
On N4 an A715 is about 2x less efficient than an N3 Apple E core when clocked down and at the same perf (platform power).

Is that bad? I mean sure. But AMD and Intel can’t do better than that at all for the area and energy, so it’s Apple’s E Cores at home. Good news for MediaTek tbh.

TBD on Oryon’s in 8 Gen 4.IMG_1941.jpeg
 
adroc_thurston said:
They had A65 once but that died with no apparent successor.
Click to expand...
I heard it was co produced by some different Chandler design team that had never been mentioned before.

For sure it was extremely odd to have gone in that direction only to change again for E2/A510.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top