Home
Creators
Search Recent Missing Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Discussing Zen 4 vs Raptor Lake Segmentation with James Prior - Guest Telegrams (Patreon)

Published:
2022-10-01 01:37:14
Imported:
2026-06
Tags:

Content

Broken Silicon's next guest will be a former AMD Product Manager, James Prior.  This man was directly in charge of deciding how to segment AMD CPUs from the Zen 1 - Zen 2 generations, and so I cannot think of a better person to have on to discuss AMD vs Intel competitiveness and segmentation!

He was also the Head of Global Communications at SiFive, is presently the Head of Marketing at Icegiant LLC, and is also presently the head of Product Management at NXP Semiconductors.  Needless to say, he can probably talk about anything related to silicon and segmentation in general.

So don't hold back!  We will be heavily discussing Zen 4, Raptor Lake, future CPUs, Lovelace, RDNA 3, and the state of the market in general.  You have 36 hours (until Sunday Morning) to submit questions/thoughts below!  Be concise, use good grammar, and be respectful to be considered.

Previous Episode with James: https://youtu.be/c2Xc2KtnAvQ

James Prior LinkedIn: https://www.linkedin.com/in/jamesdprior/

Comments

Max Eliaser

Hi James and Tom! Seems like we have had a lot of launches lately that have not been in tune with the times or with customer sentiment. Do you agree with this assessment and how long does it usually take a big tech firm to correct for this type of miss?

Max Eliaser

Another question for James: My ears perked up when I read about your association with SiFive. I am genuinely thrilled that an open hardware CPU family is taking off where others have failed, (including players with a head start like MIPS and POWER,) and I think that can only be a good thing for the long-term health of the CPU industry. My question is about where this leaves ARM: It seems like ARM is disrupting x86 from below while Risc-V is disrupting ARM from below. Obviously the ARM ISA has tremendous penetration and it isn't going extinct anytime soon, but do you see ARM settling into a different niche anytime soon (i.e. less emphasis on Cortex and more emphasis on Neoverse) or even a MIPS-style open hardware play?

MJ

Hi Tom and James. I could be mistaken, but a long time ago I thought Intel would release special SKUs for power efficient processors for desktop. With the rising power consumption is there and opening for latest generation platform with power efficient SKUs? I feel this is mostly for OEM platforms but the power consumption is insane lately. Almost a side effect of good competition.

MelodicArmor

Hi James, I remember your prior guest appearance. (I know terrible joke. Have mercy.) With the recent crashing of crypto, inflation still crawling along and most demand satisfied for especially the DIY market, do you think that product launches over the next year to even two years will need to bring alot of features and lower prices for them to be appealing to customers? Also what are your thoughts on where the foldable devices market is going? [Referring to devices like the Samsung Flip and Fold, ASUS's Zenbook Foldable OLED etc.] Thanks for the analysis in advance.

10GHz Netburst

Hi guys, I've often wondered why companies never segmented GPUs into different power classes, for example a full Navi21 "underclocked" to 150w to be a more powerful and more efficient alternative for those wanting a premium GPU but having limited cooling capacity (for example in SFF builds). Intel particularly, has long had separate lower power SKUs for their CPUs (T-series). Could you discuss why and why not it would make sense to segment like that?

Max Eliaser

Alternate history scenario: What if Intel had never sold XScale?

QuickJumper

Hi Tom and James. I was thinking of what AMD will do in future with their CPU's based on how zen4 looks like. How does E-core compare in size and efficiency to one zen4 core? Because if they make it even more efficient and smaller they could just offer core that is 90-95% performance of P-Core while being very similar in efficiency of the E-core and for gaming they could just add V-cache. I wonder if this is what they're going for because it would make bigLITTLE strategy basically useless.

Cleansweep

Hey Tom and James, In this era of CPUs going to boosting algorithms that max out performance based on operating temperatures, is there a point where the performance gains don't outweigh the inevitable consumer and reviewer consternation? Or could broader PC DIY market conditions, like cooler pricing and availability, be a factor in determining how far CPU manufacturers might push things?

MeyerTechRants

Do we have any guesses as to why AMD seems so happy to get bent over the table in nT performance at the lower end? Even the 7700X loses handily to the 13600k in nT performance. AMD could easily at least mitigate this by shifting core counts; 8c r5, 12c r7, and 16c r9. This way they'd have a small loss in 5 tier, a small win in 7 tier, and competitive in 9 tier. Does AMD see this loss in lower segment performance as significant?

10GHz Netburst

How much of a say do engineering teams have in the marketing and segmentation of the products they design? How much of a say do you think they should have?

Benjamin Cannon

With HEDT and Workstation Threadripper being merged into one effective line up, will there be a new naming scheme to replace Ryzen HEDT or where will HEDT going?

rafa zaya

Hello! As we’ve seen features like ray tracing and DLSS add value for GPUs (at least for some) does James expect new features being important for CPUs? Neural engines or the like seem to be missing in this new generation of CPUs.

Amiablechief

Hey Tom and James, curious to know your thoughts about Intel's Software Defined silicon and what it might mean for the future of consumer CPU and GPU segmentation. Could this be a way to reduce e-waste, especially if yields improve on leading edge nodes?

Chris Rijk

In the space of a few months we have three GPU releases from three different companies and two CPU releases. Availability looks good. Prices are generally trending down. But... why does nobody seem happy despite all this?

Chris Rijk

If you're a gamer, then the 5800X3D was like getting Zen 4 early, at least in game benchmarks. There's a lot of anticipation for a 7800X3D or similar - it'll probably be like getting Zen 5 early. This line of products seems to have massive mindshare already despite only being several months old. Can Intel compete with this? Should they try?

Reginald Aryee

Binning Related: How has binning evolved over the years? How long does it take to bin a chip? Did Zen completely change the binning process for AMD? Why didn't AMD use 1st gen Zen to kill off everything lower than 6 core CPUs? (Or alternatively, how feasible would it have been?) It would have been a very interesting situation for Intel to find themselves in. RAM Support: What prevents AMD from increasing the official maximum supported RAM frequency beyond the JEDEC standard spec? (Zen 1 = 2666, Zen+ = 2933, Zen 2 = 3200, Zen 3 = 3200) Do you think AMD will use EXPO to increase their official maximum supported frequency beyond the JEDEC standard(s) in the future? OEMs: What is stopping AMD from requiring EOMs to use 2x DIMMs in every laptop in order to not gimp APU performance? This has been an issue for years. Does the AMD Advantage program solve this? Bonus Question: Is there a chance AMD will change how it defines TDP so that it can be more straight forward and useful? TDP (Watts) = (tCase°C - tAmbient°C)/(HSF θca) (Definitions taken from GamersNexus) tCase°C – maximum integrated heat spreader temperature tAmbient°C - maximum temperature at the HSF (heatsink fan) inlet to achieve rated performance HSF θca (°C/W) - minimum °C per Watt rating of the heatsink to achieve rated performance

Dig Wiggler

Hi Tom and James Will Amd ever get SMT4 working? I remember this being mentioned for years and years but it has yet to see the light of day, if it ever does in desktop/laptop markets. I think this could be something vital for Amd to combat Intel in segmenting their product stack instead of using little cores themselves. Say, a hypothetical 2 core / 8 thread Zen whatever chip for $100 (chose that price because Amd has nothing to compete with a 12100f i3 atm). Amd making a chiplet that small would also imply a more desktop specific architecture, since they haven’t made a dual core beside the Athlon 200GE. Maybe they never will do this since they have become a “premium” brand.

agent2

Hi Tom and James! Lovely to see you two back together in a podcast. My question relates to RDNA3 pricing and the Radeon Technologies Group in general: Why did AMD have such a HUGE gap in capabilities in executing on the CPU front and executing on the GPUs? As far as I know those are two separate divisions, so no employees are hopping in between in terms of resources. When AMD was releasing bangers like Zen+ and Zen 2, RTG was perceived as the second fiddle that was forever going to be starved of money and NVIDIA would remain the only player in the enthusiast graphics board market. Why so? ATI used to be a company that prided itself in easily beating NVIDIA more than 50% of the time when they were competing, but mid 2018 to mid 2019 it seemed like they were a ghost of their former selves. But when AMD came back with RDNA and RDNA2, why didn't they price it in a way that their cards would have been no-brainers to yank as much market share from NVIDIA? With more rumours about RDNA3 pricing, what would your estimate be for AMD to truly make RDNA3 a memorable gen? Thanks for reading and keep on rocking!!

Capelio

When will 24 and 32 core Ryzen land? Given where AMD’s core density is going (128 core Genoa & Bergamo), it feels like the time may soon be right to raise the bar against Intel.

Capelio

James: rumor has it you were at AMD during K12’s development. Jim Keller is on record calling its cancellation a huge mistake. Was K12 something you worked on, and if so do you have any thoughts?

Capelio

James: in a hypothetical world where Intel brings you in to get their roadmaps back on track, what would you do in your first six months?