• 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 Intel current and future Lakes & Rapids thread

Page 699 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I don't buy that. If they had them working other than any 'accelerators' they would already be selling them minus those that feature, and could start selling the ones with the accelerators later.
I think a big feature with SPR is their software licensing feature to unlock accelerators for later use. They don't want to admit defeat by launching a range of SKUs without those accelerators and licensing options. That's supposed to be their big ticket to make them "great again" in the enterprise world.
 
  • Like
Reactions: ftt
Certain AI/HPC Apps that combine AMX/AVX512 that benefit from High Bandwidth memory will see a boost in performance in the Magnitude of Order when compared to standard settings. Thats where Intel will claim victory over any new AMD offerings. But on general task workload apps, Sapphire Rapids is 2 years late.
 
  • Like
Reactions: ftt
Royal core aims for double single core performance of golden cove and will have four way hyperthreading.. so royal core cinebench prediction 4000 single core performance.. just guessing
During that Keller's talk at Berkeley, he sounded like Intel was truly aiming for a something different. Was it a simple PR or is Intel really going to experiment? Who knows. There are always wild R&D projects with a long way to the market.

Ocean Cove is gone. Royal core on the radar? Who knows
 
Intel's management is faced with many big issues from process problems they've battled for years, a resurgent AMD, and the COVID PC buying spike ending, and this is what they deem important to spend their time on?

I guess from the school of "we have nothing positive to market, so let's distract people with pointless name changes!"
 
Intel Sapphire Rapids 300mm Si Wafer.

Total Complete Dies: 148

With 50 killer defects per wafer they could be getting 98 fully functional dies. Out of that how many 60C/120T Xeon Platinum 8490 are they getting?

1663371324042.png


This was a very good yield calculator but it's been on "Maintenance" for a few months now.
 
Intel's management is faced with many big issues from process problems they've battled for years, a resurgent AMD, and the COVID PC buying spike ending, and this is what they deem important to spend their time on?

Marketing's gotta do something. Plus I think the idea was that the Pentium and Celeron product names were confusing people, so they figured they could up sales somehow if they just stuck to a series of numbers and letters and did away with the Pentium/Celeron titles.
 
@nicalandia
If find this calculator to be pretty decent: http://cloud.mooreelite.com/tools/die-yield-calculator/index.html
Without knowing Intel's defect density we cannot calculate the percentage of good dies. But instead we can do a comparison:
  • Let us assume a defect density of 0,1 defects/cm²
  • For the Alder Lake 8+8 die with a size of 215mm² (10.5 * 20.5) you will get 207 good dies out of 256 in total - around 81% yield
  • To get 148 complete dies from 300mm with a squared die it needs to be 19mmx19mm = 361mm² (there is margin for error because of scribe lines)
  • Everything else being equal you get 104 good dies and a yield of 70.5%
 
Didn't notice it at first but the Intel PR makes it sound like it's only changing for laptops. So that there would continue to be Pentiums and Celerons for desktops.
 
Back
Top