Intel Announces "Biggest Brand Update" For Core CPUs

[Anux]

i know its naive to even think about this but why do these companies do not name the process node 100 MTr/mm² - 208 MTr/mm²
instead of these fake 4nm numbers ?

Because that isn't much better either. A node might reach 100 MTr/mm² but only at certain frequencies and temperatures.
Also different structures (cache, IO, compute) need different amount of space.

See this article, it shows how you can reach higher density by sacrificing max frequency.

[Sonadow]

Add this naming scheme to Intel dropping Pentium and Celeron for "Intel Processor" and now we have a branding clusterfuck.

Can't wait for Apple to sue them for branding their processors as Ultra too ...

Is it really a clusterfuck or it's just you having a bone to pick with Intel.

Intel Processor -> Low-end
Intel Core -> Mainstream
Intel Core Ultra -> HEDT

Looks a heck more simple than AMD's scheme of Ryzen X - ABCD where A is allowed to be higher than X. At least Intel's numbering scheme of iX - XABC made more sense where people knew an i3 -3XXX is definitely inferior to an i5-5XXX as compared to people scratching their heads wondering if a Ryzen 3 7320U was superior to a Ryzen 5 5500U

[Ronshere]

Users report a distinct improvement in core performance after dropping the letter "i".

[Vermilion]

Is it really a clusterfuck or it's just you having a bone to pick with Intel.

It's not picking a bone when you call out their marketing BS. And btw AMD's new mobile naming scheme is also marketing BS aimed to confuse the consumer.

Tech savvy people will always read specifications and look for benchmarks and reviews before buying something, no matter how dumb the naming scheme is. But imagine giving advice to a friend or a relative. Before it used to be "Look for an i5 or i7, depending on your needs", now it's "Buy an Intel Processor, but it has to be an Intel Core Processor, preferably 5 or 7, and stay clear of Intel Core Ultra Processor if you're on a budget."

Looks a heck more simple than AMD's scheme of Ryzen X - ABCD where A is allowed to be higher than X. At least Intel's numbering scheme of iX - XABC made more sense where people knew an i3 -3XXX is definitely inferior to an i5-5XXX as compared to people scratching their heads wondering if a Ryzen 3 7320U was superior to a Ryzen 5 5500U

What are you on about? Both Intel and AMD sell Core i3‑10105 as well as Ryzen 3 7320U respectively, where the A is allowed to be higher than X, because they're unrelated. One is the tier, the other is the generation. And Intel's new naming scheme creates more tiers.

[qarium]

Because that isn't much better either. A node might reach 100 MTr/mm² but only at certain frequencies and temperatures.
Also different structures (cache, IO, compute) need different amount of space.

See this article, it shows how you can reach higher density by sacrificing max frequency.

you know that ryzen 7990X3D 8990X3D will have one zen4 die with 8 cores and 3D cache ontop and one die with zen4C with 16cores...

this will give you total of 24 cores and the games will run fast on the max-clock-speed zen4 cores with 3D cache and multicore applications will run fast on the zen4c die....

"it shows how you can reach higher density by sacrificing max frequency."

for the customers they will not have to deal with this tradeoff because future AMD cpus have one zen4 die with 3D cache and one die with zen4c...

by this move the customer will have zero loss and the scheduler will move the workload on the right die and core.

[Anux]

you know that ryzen 7990X3D 8990X3D will have one zen4 die with 8 cores and 3D cache ontop and one die with zen4C with 16cores...

I never heard of those, where did you get this from?

... by this move the customer will have zero loss and the scheduler will move the workload on the right die and core.

Not sure what that has to do with the naming of process nodes?

[qarium]

I never heard of those, where did you get this from?
Not sure what that has to do with the naming of process nodes?

there is no source just logical thinking. AMD has their new zen4C 16core chiplet die. but this alone would not be successfull because of the lag of of IPC and the lag of clock speed and so one.

but in a asymetric design like the 7950X3D it could be a success. then you have one chiplet of zen4 with 8cores and 3D cache and the second chiplet with zen4c... then you have total of 24 cores and AMD would beat intel in the multicore benchmarks

yes the process nodes the interesting part is that ZEN4c is also 5nm the same as zen4 is 5nm

this means AMD could beat intel with chipdesign alone without newer process node.

right now a intel 13900K or 13900KS has the same performance as a 7950X3D in multicore workloads

but a zen4+zen4c combi of a 24core 8990X3D would beat intel very hard..

[creative]

qarium

Ever since Intel released the i7 12700K then the i5 13600K, I pretty much thought "Well there goes AMD's accessible multithreaded dominance." That being said, I still am in no hurry to trade out my R7 5800X, it would be just pointless for what I do.