PC Perspective Podcast #442 – 03/23/17
Join us for Topre and CORSAIR Keyboards, ARM DynamIQ, Optane Launch, EVGA 4K gaming laptop, and more!
You can subscribe to us through iTunes and you can still access it directly through the RSS page HERE.
The URL for the podcast is: https://pcper.com/podcast – Share with your friends!
- iTunes – Subscribe to the podcast directly through the iTunes Store (audio only)
- Google Play – Subscribe to our audio podcast directly through Google Play!
- RSS – Subscribe through your regular RSS reader (audio only)
- MP3 – Direct download link to the MP3 file
Hosts: Ryan Shrout, Jeremy Hellstrom, Allyn Malventano, Ken Addison
Program length: 1:35:25
Podcast topics of discussion:
Week in Review:
News items of interest:
Hardware/Software Picks of the Week
If I was a Bengals fan, I too
If I was a Bengals fan, I too would have that scowl look in the screen preview of the video.
Lesson#1 For all those who
Lesson#1 For all those who think that any defective Zeppelin(Dual CCX unit die) based Zeppelin dies are being used to make any 16/24/32 core server or 16(maybe) core consumer variants.
Remember that all the various Zen/Naples/other server/workstation and Ryzen 7(5 and 3 also) consumer variants are using the same single Zeppelin monolithic die(2 CCX units/8 core die) in a scalable fashion from 1 Zeppelin Die(Ryzen 7, 5, 3, no MCM used) up to 4(Naples) Zeppelin dies on an MCM module(Server variants).
So any Multi-Zeppelin die MCM(Multi Chip Module) processor variants 16, 24, 32 cores SKUs are made up of multiple Zeppelin dies that have been pre-tested as non defective before they are added to any MCM module and used to make any 16, 24, 32 core variants on a MCM module. There are no defective die harvested multi-Zeppelin die procesor on an MCM based SKUs currently.
If any of the modular Zeppelin dies are found to have any defects those dies are used(Most likely this early in Zen’s production) to make the consumer(Ryzen 5 and Ryzen 3) variants. So any 16, 24, 32 core MCM based processors will be using 2(16 CPU core), 3(24 CPU core), or 4(CPU 32 core) Zeppelin dies based SKUs using pre-tested and fully working Zeppelin dies that are then placed in increasing numbers ON an MCM to make any 16, 24, 32 core processor variants.
Currently AMD is only using its Zeppelin(2 CCX unit) scalable die to make its entire line of Zen/Ryzen(7, 5, 3 made with 1 Zeppelin die NOT on an MCM) up to its Zen/Naples(4 Zeppelin dies on an MCM) line of SKUs. There are no Zen monolithic dies with more than 8 CPU cores(2 CCX units) and all of AMD’s 16 core(Rumored), 24 core(Rumored), 32 core(Naples) are/will be made up of multiple Zeppelin dies wired together on an MCM module, except for the Ryzen 7, 5, 3(single Zeppelin die variants NOT being on a MCM) with the only parts currently using bad/defective Zeppelin Dies being the Ryzen 5 and 3 consumer variants.
Could AMD at some future time use any defective Zeppelin dies to make say a 12 core 2(Zeppelin dies) variant on an MCM module. Well that depends of how much demand there will be for the Ryzen(5, 3) series consumer parts and how much stocks of any defective Zeppelin Dies that AMD may have available now and in the future to use for any 12 core(using defective Zeppelin) dies on an MCM/modular variants.
All of AMD’s server Zen/Naples/other variants and consumer Ryzen(7, 5, 3) variants will be made from the same Zeppelin(2 CCX Unit/8 CPU core) monolithic die in increasing Zeppelin die numbers of 1 for (Ryzen 7, 5, 3 NO MCM); 2 for (rumored 16 core on MCM), 3(rumored 24 core ON MCM), 4(32 core Naples on MCM).
AMD has gone all scalable with its Zeppelin dual CCX unit die but some people still think that AMD is making some large monolithic die SKUs for its Naples(32 core), and any 24, 16 core rumored variants also. This includes some the technical press speculating about “Defective” 32 core dies(Does not exist for Zen) to make a rumored 24 core part(Does not exist for Zen). The largest Zen/Ryzen die is only 8 cores with any larger Zen/Naples and other server variants and Zen/Ryzen(possible maybe 16 or 12 core/binned) parts made up of multiple Zeppelin dies on an MCM module.
slight issue with this
slight issue with this podcast @ 50:30 there is a slight massive error in understanding the latency charts its in “us” not “ms” they said that a DC SSD had access times of nearly 3 seconds when in fact its 3ms 99% (about 10ms for 1%) for the DC P3700X
Optane SSD DC P4800X is somthing like 0.05ms 99%
i do hope they respond to this in the next pod cast