• 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.

Raptor Lake Leaks + Intel 4 developments

Suspect we'll have further leaks for Raptor Lake very soon, if I were Intel I'd want to be releasing it at the same time as the Zen 3 refresh, to steal thunder from that launch. Excited to see the reviews :)
 
Grim5 said:
50% faster than ryzen 5000 in some scenarios, and we don't even have super fast DDR5 yet, can only get better for Intel
Click to expand...

I think we'd need to wait to see the performance of AMD's V-cache CPUs to see if these can compete with Alder Lake on price / performance...

AMD can still get good performance out of DDR4 for another year, I think mostly because DDR5 isn't running in 'gear 1' mode yet, with Alder Lake CPUs.
 
Last edited:
Grim5 said:
Alder lake so good Intel doesn't need to launch anything till Zen 4
Click to expand...

If Intel can ready a 7nm CPU generation for release at the end of 2022, I think they will (that isn't Meteor Lake). Products based on the initial 7nm fab tech. were delayed by a year, but were originally planned for 2021 (there were defects which caused yield issues). It could be that they still don't know if 7nm yields will be sufficient to release a new series at the end of the year.

If they can't release the next gen by the end of the year, it would probably make more sense to delay the launch until Q1 2023, rather than rush out something inferior, especially as it's very likely Zen 4 desktop CPUs will release by this point.
 
Last edited:
Nothing on Intel 7nm (Intel 4) is coming until 2023, Intel have already stated as much. The interesting side of 7nm is the + variant slated for H2 '23, also now known as 'Intel 3' this is indicated to have a 14% PPW over 'Intel 4' just 6-9 months later. But again to quote Intel directly, they state "Intel is reaffirming that it intends to return to leadership performance in semiconductors in 2025" which means it needs to be besting the TSMC 3nm/2nm process by that time and in volume. If anyone can do it Intel can, they have the resources and the knowledge, they just needed refocusing to catch up.
 
Last edited:
g67575 said:
Here's what Intel's previous CEO said back in 2020:
https://uk.pcmag.com/news-analysis/127899/intel-sorry-but-our-7nm-chips-will-be-delayed-to-2022-2023
Click to expand...

So why would you quote the previous CEO, who was known to stretch the truth all of the time, look at how often he said 10nm was 'on track' you could, I mean it is possible, just quote the current one???

Pat Gelsinger

“I've had the opportunity to personally examine progress on Intel's 7nm technology over the last week. Based on initial reviews, I am pleased with the progress made on the health and recovery of the 7nm program. I am confident that the majority of our 2023 products will be manufactured internally,”

And you could then just look up the IDM 2.0 strategy, and what was said by Intel on their conference call for investors, stating that 2023 was the release window for 7nm. Or you could just find more ways of being wrong.
 
tbf, that article does say late 2022/early 2023 :cry:

Here's what the last CEO said “The primary driver is the yield of Intel's 7nm process, which based on recent data, is now trending approximately twelve months behind the company's internal target.”

I'm not sure it would be in Intel's interests to discuss 7nm CPUs details yet, at least until the entire Alder Lake series has been released for at least a few months.
 
Last edited:
Intel discuss the upcoming process nodes continuously, to advise steak holders and partners where they are, and how they are progressing. Hence the fabrication from Bob Swan about 10nm leading to the now current CEO having to be much more transparent on the real progress with 7nm, which of you believed Bob Swan it has already been out nearly 18 months.

7nm is 2023, end of discussion really, until Intel say it is either delayed or expedited.
 
Well, I'm just pointing out that they've already released their first 7nm chip this year (apparently Intel 4), called the Loihi 2. But what we don't know yet is the yields of Intel 4 with other chips.
 
A test chip at 31mm2, with 1/3 the density of transistors required to make it useable at desktop level, and they've stated it's pre-production. Call me back at the end of 2022 when they've actually got it working properly.
 
What are downsides of using hybrid memory controller? Zen 4 as competitor of RPL will have ddr5 only controller, i guess that means it will support higher speeds, function better.
 
Gondar said:
What are downsides of using hybrid memory controller? Zen 4 as competitor of RPL will have ddr5 only controller, i guess that means it will support higher speeds, function better.
Click to expand...

Mostly die space saved. How that die space is used is anyone's guess.

Remeber that Zen4 will feature iGPU - so this will take up a large % of the die, compared to Zen3 which has no iGPU. Or if they added it as a 'chipset' on the package, it would take the place of a potential CCX etc.
 
There seems to be some kind of confirmation bias going on, e.g. lots of people believing that Raptor Lake has been confirmed + will have a really awesome spec etc.
 
Raptor Lake is the tock to Alder Lake's Tick, process node improvement and better PPW.

As for the AMD comment.

Remeber that Zen4 will feature iGPU - so this will take up a large % of the die
Click to expand...

Totally forgetting that currently AM4 uses a huge I/O die that is on 12nm, ~125mm² and about 2.1 billion transistors, compare that to a 7nm CCD which is ~75mm², yet has 3.8 billion transistors, shrinking the I/O die to a mature 7nm/6nm process and adding an integrated GPU will be reasonably easy on the AM5 (LGA) package, as much like Intel it will be very slightly larger socket area. The 5700G APU is 10.8 billion at ~180mm² on a single monolithic 7nm die, yet contains almost the same I/O as a standard Zen3 based part, and a full compliment of 8c/16t, but half the L3 cache amount. So I doubt they'll struggle to fit a comparable GPU into the same space plus the I/O using the node shrink to free up the space. The I/O also counts for a huge amount of the power draw with Zen2/3 based parts, so power budget should be better with the smaller node.
 
You must log in or register to reply here.
Back
Top Bottom