• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

AMD Zen 5 Technical Deep Dive

Problem is now your spinning a completely new IO die for a specific segment of the market

Currently this is why their chiplet approach works as they can mix and match parts to suit from bare office PCs all the way up to highest end server parts with very little changing between them. For AMD this is the best in terms of mass productions benefits. I believe for their whole current offerings its 2 IO dies, 2 CCD types and 1 Monolithic die to cover all these?

Instead of what you are saying in regards to reworking of the IO Die. I suspect there will be a repurpose of some of their mobile offerings to fill the low end office pcs, especially with the explosion of USFF PCs becoming more and more common in office spaces.
There might be some overlap - my point is two-fold - in that this could be combined to scale up the Ryzen 5 and 7 range to encompass c cores also, be it 4c + 4/6/8 combos. If Intel continue to throw increasing numbers of e cores at the desktop to shore up the thread counts, AMD will at some point need to do something.
The other option is to do chiplets with either all c cores or a mix on the desktop parts, but as is the case with the current AM5 platform I don't see that making it price into the low-end/low-cost price point - AMD will be merely redefining their version of it at a higher cost price.

To be honest the repurposing of mobile dies into USFF devices isn't new for either Intel or AMD, but the current Pheonix dies used for the 8000 series desktop parts are a bit 'meh' in terms of utilising the desktop platform, especially if you wanted to use a PCIe x16 card for anything. And at the current pricing, Intel will still shift more i3's to people who don't care about how good the iGPU is, which when you get to the 8300G isn't the huge step offered by the 8700/8600 parts, and at that price point the i3 is a better CPU also.

Unfortunately, 'AI' is the new buzzword and I suspect the likelihood is the next IO die they come out with will probably have their 'not-actually-AI' logic block added to it before anything else so as to satisfy Microsoft's dumb requirements that require 'still-not-really-AI' and that it be performed by dedicated logic instead of processed by perfectly capable iGPU or dGPU hardware...
 
Is this package power or has this reviewer excluded cIOD consumption? It should be a static 10-15W for cIOD alone. Intel can idle down to a watt.


Both tests are from the CPU power plugs. So as raw as they can get to actual CPU power for both Intel and AMD.


"We measured power draw directly from the dual 8-pin connections on our motherboard. For the 7800X3D, that gave us a peak of 119 watts at load and 29 watts idle.

Read more: https://www.tweaktown.com/reviews/10384/amd-ryzen-7-zen-4-7800x3d-cpu/index.html"



For the 14900K they measured idle of 14W, but not single digit.
 
Surprising there's still no plans in this generation to bring the 'c' cores into their main desktop product offering (the 8000g series being both last gen zen4 and also more of a laptop part).

I've often thought AMD could kill two birds with one stone by doing an alternate IO die configuration loosing one of the chiplet infinity fabric channels on the IO die and instead splicing on some zen4c/5c cores - sure they'll end up with a slightly bigger IO die (likely a few mm² bigger) but they could then offer lower cost quad-core AM5 CPUs using the IO die cores only (and AM5 really needs some cheap low-end parts to exist once AM4 really is dead), and have 4c + 8 core desktop products (and whatever derivative SKUs they'd want).
Admittedly the quad-core AM5 using just 'c' cores would be an odd product outside of just being a value proposition - tons of IO capability compared to the 8000g series but not a huge amount of CPU power to exploit it, but I could see home NAS / niche appliance / embedded usage scenarios where the additional 8000g Radeon CUs wouldn't be utilised and a smaller essentially 'monolithic' die with better IO options would be desireable.

I reckon there's a large amount of bare minimum office PCs being sold that utilise new Intel Core-i3 chips instead of AM4 4000/5000g CPUs due to AM4 being seen as older platform, the newer i3's having better IPC and (even though Intel's process tech is behind) comparatively acceptable power usage. Eventually that existing AM4 inventory is going to end or not be used for most new sales.

Obviously they can still make the normal IO die for the 16 core CPU configurations they currently push out - not suggesting they abandon that.
It's worth noting that by the Zen 3 era, AMD has said that their yields are so good now that they often have to artificially gimp the "worst" bins to put into lower tier products, compared to the earlier Zen 1 - 2 generation where actual defective dies were used for lower tier products. There was also that limited run of incorrectly fused bins that allowed lucky owners to get double the cores/threads like the old silicon lottery.

Assuming that's still true now, it's more beneficial for AMD to just keep producing and maximizing gains in the more lucrative middle to high end of the market, while letting AM4 take the low to middle end of the market. Moreso now that they gave AM4 an minor life-extension with the new X3D parts.

There definitely is a need for some low-end options, but I suspect AMD might just eventually repurpose some previous-gen parts for the low end; whether it's some old Zen 3 or Zen 4 stuff put onto a chip that fits the needs of the low end, since the demands there would usually be much lower. About the only reason to go with something beefier is if AMD also wants to add Copilot+ capabilities to the budget/low end of the spectrum, just to further eat Intel's lunch, in which case they may as well just go with thin clients or mini-PCs reusing their mobile chips.
 
The Ryzen 9000 desktop processors are exciting, too, but we really wish it came with an NPU and beat the 7800X3D, because Arrow Lake-S will probably do both, at unknown power levels though.

Gonna be hard to match even this moderate uplift with the roughly the same IPC increase but a clock regression.
 
I am happy to see that finally there will be laptops on the market which are capable of casual gaming without dedicated graphics cards. When I bought my laptop in 2020, it wasn't the case. But it appears there aren't / weren't many people like me.
 
AMD thinks that they can get away with a 6-year-old CPU being sold as brand new to the masses.
Are you having an Ozzy Osbourne moment? :kookoo: Nothing of what you just stated is correct. AMD is not selling ANY of their 6year old CPU models as new. The newest 5000 series models are a refresh of a 3 year old range. If you're talking about the socket, AM4 is 8years old, almost 9. Yet it is still selling well because it still PERFORMS well. Yes, yes.

It never is. You can't put the sensor in the hottest part, because then the sensor circuitry would take up so much space that the temps won't reach maximum in that area.

What's done is to put several sensors in less "busy" locations and interpolate the temperatures, of course including "enough" safety margin
Boys & girls; this comment is what we call the "Conversational Bitch-slap". It is at the same time the rarely seen, much feared W1zzard burn! :fear:

I live for moments like this. :rockout::peace::lovetpu:
 
Are you having an Ozzy Osbourne moment? :kookoo: Nothing of what you just stated is correct. AMD is not selling ANY of their 6year old CPU models as new. The newest 5000 series models are a refresh of a 3 year old range. If you're talking about the socket, AM4 is 8years old, almost 9. Yet it is still selling well because it still PERFORMS well. Yes, yes.
Oldest CPU tech AMD sells is the refreshed Zen 2 on the Mendocino (xx2x) mobile range and that's originally from 2019 (remember Intel shouting foul play?), but yeah, it performs just what it needs to in the price point it's sold at. At least refreshed that they are, they use RDNA instead of Vega.
 
Boys & girls; this comment is what we call the "Conversational Bitch-slap". It is at the same time the rarely seen, much feared W1zzard burn! :fear:
Wait, what? I never intended for my comment to be rude
 
A small typo which no one seems to have caught.

Microsoft isn't far behind with the NPU 4 of Lunar Lake peaking at 48 TOPS
 
Back
Top