Caporegime
- Joined
- 21 Jun 2006
- Posts
- 38,367
Thanks. I may get one for an itx build I'm thinking of possibly doing.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
AMD what you doing to fight off Alderlake?
- Thread starter subbytna
- Start date
Thanks. I may get one for an itx build I'm thinking of possibly doing.
Soldato
I think it's worth separating the ARM nomenclature from what Intel is doing. Intel is doing performance and efficiency cores. They don't necessarily have to be "big" or "little" cores, although by virtue of what an "efficiency" core is designed to do, it will be smaller than a "performance" core as we've seen with Alder Lake. Are the leaks for AMD's hybrid approach for Zen 5 even considered a "performance" and "efficiency" hybrid? It almost feels like "focussed" and "wide" design, with a Zen 5 chiplet with fewer cores doing the fast, focussed work, and then a Zen 4 Dense chiplet with a lot more cores aimed at highly threaded "wide" workloads. And it seems that's what Bergamo is, a purely "wide" design cramming 128 Zen 4 Dense cores into the package.
This is what Intel will have to contend with. I have no doubt that Meteor Lake is going to be the true return for Intel, but I still don't see it as a design concept to be anywhere near as flexible and scalable as AMD's chiplet approach. Unless, of course, Intel do their own chiplet/tile implementation with an interconnect that actually works.
Soldato
Oh grow up, that's not what humbug said and you know it.
Has anyone done a proper analysis of ADL's performance cores vs Zen3 cores?
Did a quick calculation of ADL using this Reddit post:
https://www.reddit.com/r/intel/comments/qhbbow/10nm_esf_intel_7_alder_lake_die_shot/:
and some calculation (based on the 24MP version, but that didn't have all the annotations).
So at least the E-cores at @ 1.92mm² are nearly a quarter of the size of the P-cores @ 7.36mm².
How that compares to Zen 3 (and for that matter M1 Max) is a more interesting question.
That is, who's design is the most efficient at performance per transistors?
Did a quick calculation of ADL using this Reddit post:
https://www.reddit.com/r/intel/comments/qhbbow/10nm_esf_intel_7_alder_lake_die_shot/:
and some calculation (based on the 24MP version, but that didn't have all the annotations).
So at least the E-cores at @ 1.92mm² are nearly a quarter of the size of the P-cores @ 7.36mm².
How that compares to Zen 3 (and for that matter M1 Max) is a more interesting question.
That is, who's design is the most efficient at performance per transistors?
Soldato
- Joined
- 31 May 2009
- Posts
- 21,456
It'll be the ARM M1 for apple, as its on a much smaller process I'd have thought.
Intel need to get past its nonsense nodes and designs and hope AMD aren’t too far ahead.
It’s Nvidia Joxen should be concern about. Intel win by default.
Has anyone done a proper analysis of ADL's performance cores vs Zen3 cores?
Did a quick calculation of ADL using this Reddit post:
https://www.reddit.com/r/intel/comments/qhbbow/10nm_esf_intel_7_alder_lake_die_shot/:
and some calculation (based on the 24MP version, but that didn't have all the annotations).
So at least the E-cores at @ 1.92mm² are nearly a quarter of the size of the P-cores @ 7.36mm².
How that compares to Zen 3 (and for that matter M1 Max) is a more interesting question.
That is, who's design is the most efficient at performance per transistors?
Interesting question, not gone into the detail you have but a quick google/look around gives a total die size for a Zen3 chiplet of ~80mm2 and looking at the annotated die shot from here:
https://www.reddit.com/r/Amd/comments/jqjg8e/quick_zen3_die_shot_annotations_die_shot_from/
The actual cores themselves are less than half of the die area, so if my brain is working that's <5mm2 per core. So between the roughly in the middle of the Intel P and E cores, but performance is closer to the P cores.
Although it's quite likely you could design a core the same size as the E cores but give better performance if you wanted to, I'd imagine the Intel engineers made deliberate decisions for efficiency that would hurt performance, even just layout of the same transistor/logic blocks can affect that.
Has anyone done a proper analysis of ADL's performance cores vs Zen3 cores?
Did a quick calculation of ADL using this Reddit post:
https://www.reddit.com/r/intel/comments/qhbbow/10nm_esf_intel_7_alder_lake_die_shot/:
and some calculation (based on the 24MP version, but that didn't have all the annotations).
So at least the E-cores at @ 1.92mm² are nearly a quarter of the size of the P-cores @ 7.36mm².
How that compares to Zen 3 (and for that matter M1 Max) is a more interesting question.
That is, who's design is the most efficient at performance per transistors?
Intel’s designs are way too big to win the density fight, hence the big-little hack. But 10nm is a pretty solid node for low power. The T parts and 6+6+2 Alder lake/Raptor chips could be half decent.
With just how little you know, I'm fine with you having an opinion like that
I know a little less today. Would you like to know a little less too?
Think it's great to see actual good products from both camps but cost is out of proportion as a package, the cheapest ram kit cost more then some of the cpus skus right now. The motherboard all have a high starting price.
Just need a whole cheaper package and it be great.
Just need a whole cheaper package and it be great.
It is. Intel have managed to produce a chip that improves over the last gen.
We could talk about how it’s taken 5 years, the timing of the launch and point out the weaknesses. But Intel deserves credit for taking what was essentially two piles of garbage and whipping them into shape.
Caporegime
Has anyone done a proper analysis of ADL's performance cores vs Zen3 cores?
Did a quick calculation of ADL using this Reddit post:
https://www.reddit.com/r/intel/comments/qhbbow/10nm_esf_intel_7_alder_lake_die_shot/:
and some calculation (based on the 24MP version, but that didn't have all the annotations).
So at least the E-cores at @ 1.92mm² are nearly a quarter of the size of the P-cores @ 7.36mm².
How that compares to Zen 3 (and for that matter M1 Max) is a more interesting question.
That is, who's design is the most efficient at performance per transistors?
Same sort of thing with Zen 3, these are not exacting but i think thy are close enough. https://www.reddit.com/r/intel/comments/jtnvry/rough_comparison_of_structure_sizes_between_amd/
Zen 3 cores without the L2: 3.24mm.
With L2 they are 4.04mm
Soldato
- Joined
- 6 Jan 2012
- Posts
- 5,505
AMD prices were lower a couple of months ago, what's going on?
Caporegime
I don't know but they still seem to be selling well, they are still all over the sales charts.
Hardware Unboxed said retailers told him ADL sales are steady but no where near as high as they expected, one said they sold more Ryzen 5000 CPU's in hours than ADL sold in days.
The reason for that is probably a combination of things and not least of them the apparent marketing team up with Microsoft and Windows 11, like MS was expecting a higher Windows 11 adoption through a marketing team up with Intel, what its actually done is hamper Intel's sales.
I love it when stuff like that backfires.
Soldato
- Joined
- 6 Jan 2012
- Posts
- 5,505
all very confusing ... has your AMD not exploded at 5ghz yet ?
Caporegime
No, its fine.
I think ADL would do much better if it was not so tied to Windows 11 in the marketing and with it peoples minds.
This all day. I suspect also that a lot of the 5000 series purchases are for people upgrading existing AM4 CPU's. Platform cost is probably also a factorNo, its fine.
I think ADL would do much better if it was not so tied to Windows 11 in the marketing and with it peoples minds.
No, its fine.
I think ADL would do much better if it was not so tied to Windows 11 in the marketing and with it peoples minds.
it's Microsoft's fault too, they've actually stated they will be working to fix W11 issues for Alder Lake first before they even touch W10 - another case of Microsoft taking advantage of the situation to push W11 adoption