• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Question Zen2 CPU prices Increasing; Not a good sign for Zen3?

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

eek2121

Senior member
Aug 2, 2005
813
907
136
Basically people are seeing the price increase and reading into it as stuff that isn't there. Meanwhile its just supply and demand...
This. I feel like people need to go back to school and take an economics class or something.
The XT launch is from the exact same wafers as Matisse in 2019. It tells us nothing about Zen3/Vermeer on N7+.
Not according to AMD. "The refined AMD Ryzen™ 3000XT Series processors use an optimized 7nm process to offer more of what PC enthusiasts love. "

If they are different is really unknown because we've really heard nothing from TSMC about what these process changes actually entail.
The XT series appears to have a much improved boosting profile. My wife’s 3600XT has no issues maintaining high (4.5+ ghz) clocks.

Why people are assuming Zen 3 will be a dud baffles me. Even with no IPC increases, we already have engineering samples floating around with peak clockspeeds hitting 4.8+ Ghz.
 
Feb 4, 2009
29,147
9,718
136
Before Zen+ came out, we had good discounts on Zen CPUs. Likewise on Zen+ CPUs when Zen2 was coming out.

Zen3 is supposed to come out before the end of the year, and:
1) we have no leaks, and
2) Zen2 CPU prices (3600) have risen from $154.99 to MSRP ($199.99) in a matter of weeks.

What this tells me, is that Zen3 sucks, and that retails know it.

Edit: Let me explain. If Zen3 were better than sliced bread, then Zen2 would be (relatively) unsellable, at least at the same price-points, and thus, Zen2 CPUs should be discounted commensurate with the disadvantage relative to zen3 CPUs, before release, to clear stock.

The fact that Zen3 is getting ever-closer to release, and Zen2 prices are going UP, points me to a VERY disasppointing Zen3 release.
Could be a supply & demand thing. Markets go thru weird periods as in maybe AMD reduced production so they don’t have too many old parts on the shelf but reduced production too much this there are shortages.
Also AMD may no longer want to be the value player. They may simply want better margins on their product which would suck but also be good in a way that I won’t go into here.
Basically it’s too early to say
 

x_marX

Member
Apr 23, 2020
61
6
41
IMO, it's just pure supply vs demand!

It's not just AMD processors, motherboard (X570) prices too are increasing over here. It's possible to get an Intel i7 10700+Z490 combo cheaper than the Ryzen 3700X+X570.

I'm in the middle of an upgrade and its frustrating to see AMD stuff getting pricier day by day.
I never wanted an Intel pc but it looks like it's the way to go if price per fps is important.
 

DrMrLordX

Lifer
Apr 27, 2000
17,031
6,002
136
Not according to AMD. "The refined AMD Ryzen™ 3000XT Series processors use an optimized 7nm process to offer more of what PC enthusiasts love. "
I hope you understand that that's corporate speak for "we binned the chiplets". The 3600XT saw the biggest increase over the 3600X (versus 3900XT which saw very little increase over 3900X) for this reason. 3900X was already pretty tightly-binned on one chiplet. 3800X from 2019 was also well-binned, and in some circumstances could do many of the same things wrt overclocking that the 3600XT can do now.
 
  • Like
Reactions: Tlh97

therealmongo

Member
Jul 5, 2019
83
157
66
I hope you understand that that's corporate speak for "we binned the chiplets". The 3600XT saw the biggest increase over the 3600X (versus 3900XT which saw very little increase over 3900X) for this reason. 3900X was already pretty tightly-binned on one chiplet. 3800X from 2019 was also well-binned, and in some circumstances could do many of the same things wrt overclocking that the 3600XT can do now.
I would be in agreement with you 100%, however, as more data is coming in these XT CPUs are acting differently when you attempt to use the EDC bug.

Unsure why this would be the case if the improvements were only due to binning.

Am not saying that its not binning, just the results ive seen there is a possibiliy that some changes have been made ...
 

therealmongo

Member
Jul 5, 2019
83
157
66
EDC bug is firmware-related. The boost algo has obviously been tweaked.
Are you saying this from a position of authority or is that just your best guess as any changes to the algorithm could have been made to detect silicon quality as we know that is part of the equation ?
 
Last edited:

DrMrLordX

Lifer
Apr 27, 2000
17,031
6,002
136
Are you saying this from a position of authority or is that just your best guess as any changes to the algorithm could have been made to detect silicon quality as we know that is part of the equation ?
No, and no. TSMC already showed us the node improvements from N7 to N7+, N7P, and N6. None of those improvements show up in the XT chips, and the best of the best 3800X chips already exhibited max OC headroom similar to some of the XT chips (notably the 3600XT). XT chips obviously have a different boost algorithm - read the specs. What you aren't seeing is higher clocks @ isopower or lower power @ isoclocks which would be indicative of a move to something like N7+.

Trying to extrapolate a node change out of a fix for the old EDC bug is fruitless. There are bugs in silicon and firmware all the time, and it would make no sense for AMD to leave that bug in there when tweaking the boost algorithm to fit the tigher binning standards.
 
  • Like
Reactions: Tlh97

thesmokingman

Platinum Member
May 6, 2010
2,305
230
106
Not according to AMD. "The refined AMD Ryzen™ 3000XT Series processors use an optimized 7nm process to offer more of what PC enthusiasts love. "

If they are different is really unknown because we've really heard nothing from TSMC about what these process changes actually entail.
They're the same chips just binned at the 3950x level. If you look closely they have the same single core boost targets and the boost algorithm is massaged for higher all core boost and longer.
 

A///

Senior member
Feb 24, 2017
829
583
106
Also AMD may no longer want to be the value player. They may simply want better margins on their product which would suck but also be good in a way that I won’t go into here.
Basically it’s too early to say
Quite possibly. Intel are doing their marketing for them. "You can build a cheaper and faster**** Intel system over an AMD!"


****On a few esports titles and get destroyed elsewhere.
 

therealmongo

Member
Jul 5, 2019
83
157
66
No, and no. TSMC already showed us the node improvements from N7 to N7+, N7P, and N6. None of those improvements show up in the XT chips, and the best of the best 3800X chips already exhibited max OC headroom similar to some of the XT chips (notably the 3600XT). XT chips obviously have a different boost algorithm - read the specs. What you aren't seeing is higher clocks @ isopower or lower power @ isoclocks which would be indicative of a move to something like N7+.

Trying to extrapolate a node change out of a fix for the old EDC bug is fruitless. There are bugs in silicon and firmware all the time, and it would make no sense for AMD to leave that bug in there when tweaking the boost algorithm to fit the tigher binning standards.
Am not attempting to extrapolate anything, was justing posing the question.

Is adding a metal layer accepted as a node change?

Thanks
 

DrMrLordX

Lifer
Apr 27, 2000
17,031
6,002
136
Is adding a metal layer accepted as a node change?
That's up to the individual fab. Nobody outside the individual fab company gets to decide what really is or isn't a new node. I don't think TSMC added a metal layer to N7 since last year, though. Or if they did, they didn't make it widely-enough known for me to notice.
 

skline

Junior Member
Sep 9, 2009
11
10
81
Before Zen+ came out, we had good discounts on Zen CPUs. Likewise on Zen+ CPUs when Zen2 was coming out.

Zen3 is supposed to come out before the end of the year, and:
1) we have no leaks, and
2) Zen2 CPU prices (3600) have risen from $154.99 to MSRP ($199.99) in a matter of weeks.

What this tells me, is that Zen3 sucks, and that retails know it.

Edit: Let me explain. If Zen3 were better than sliced bread, then Zen2 would be (relatively) unsellable, at least at the same price-points, and thus, Zen2 CPUs should be discounted commensurate with the disadvantage relative to zen3 CPUs, before release, to clear stock.

The fact that Zen3 is getting ever-closer to release, and Zen2 prices are going UP, points me to a VERY disasppointing Zen3 release.
VirtualLarry, the "no leaks" might be the edict from Dr. Su. She impresses me as a no B.S. straight shooter who lets her products speak for themselves, once they are released.

Furthermore, AMD is not a charity. Prices on cpus fluctuate.
 

alexruiz

Platinum Member
Sep 21, 2001
2,751
393
126
I think it is the exact opposite:

Zen3 is going to demolish what is currently available, and as such, AMD will price them accordingly.
That means that the price gap between Zen3 and current Zen2 will be big. That gap has to be bridged, and to achieve it, Zen2 prices will go up.
 
Last edited:

Thibsie

Member
Apr 25, 2017
135
141
86
I think that was the whole point of the XT release.
I dunno if a price increase or 'just' a way to get prices a little bit more in line with previous (introduction) prices.

This is (or was? We'll see) in photo industry were many (but Canon mainly) did constantly recycle the same design with a button change here or there just to warrant a model number change and get pricing back to 'new' stuff.

Well, it works until int doesn't of course.
 

scannall

Golden Member
Jan 1, 2012
1,773
1,218
136
I think that was the whole point of the XT release.
The whole point of the XT release was to keep in the news cycle around Intel's 10 series release. Not so much 'New and Exciting'. More 'Hey, Intel doesn't get all the coverage.' When Zen 3 gets released I'm sure Intel will release something or other to the market around the same time, for the same reason. Nothing wrong with any of that btw.
 

Chicken76

Senior member
Jun 10, 2013
237
25
91
There's one aspect that nobody seems to take into account: by now that 7nm process is yielding better, so there are fewer partially defective dies. Why would they disable two cores and sell a good die for $150 in an R5 3600 when they can put it in an R7 3700X and sell it for $300, or put the few defective dies they do have into an R9 3900X and sell them for $500?
 

A///

Senior member
Feb 24, 2017
829
583
106
I think it is the exact opposite:

Zen3 is going to demolish what is currently available, and as such, AMD will price them accordingly.
That means that the price gap between Zen3 and current Zen2 will be big. That gap has to be bridged, and to achieve it, Zen2 prices will go up.
This is what I believe to be true, too. There was a slight price increase from Zen+ to Zen 2. I estimate the 12 and 16 core processors, if offered, will see an increase of at least $80 and $150, respectively, over Zen 2 launch MSRP. $580 and $900.

However, don't for a moment assume Rocketlake will be cheap. It'll be priced accordingly, and Alderlake will likely be expensive regardless of its performance against a future Zen 4-6.

The whole point of the XT release was to keep in the news cycle around Intel's 10 series release. Not so much 'New and Exciting'. More 'Hey, Intel doesn't get all the coverage.' When Zen 3 gets released I'm sure Intel will release something or other to the market around the same time, for the same reason. Nothing wrong with any of that btw.
I am inclined to say you may be onto something here. AMD slipped the RX5300 out the door and it's been or was in the news cycle the last two days. It's a low end card meant to up against the 1650 or its super variant. It's not terribly important, but it keeps AMD GPUs in the new cycle, especially when their competitor, nv is set to announce and release new products in a few hours.
 
Last edited:
  • Like
Reactions: Tlh97

Jimzz

Diamond Member
Oct 23, 2012
4,389
181
106
You all do know there is a shortage of 7nm CPUs right? AMD has run out of 7nm Renoir chips. OEMs are canceling laptops as they can't get the chips for at least another month.
I've been watching laptops as I was going to buy one for my kid but many are disappearing. I don't mean out of stock, I mean removed from sites with no return dates. Lenovo even listed a laptop they just came out with recently as discontinued.

Acer and some others have removed AMD 4000series laptops from their site. Dell has cut back as well. So this is not just 1 OEM seeing this issue but all major ones.
 

sdifox

No Lifer
Sep 30, 2005
83,618
8,731
126
You all do know there is a shortage of 7nm CPUs right? AMD has run out of 7nm Renoir chips. OEMs are canceling laptops as they can't get the chips for at least another month.
I've been watching laptops as I was going to buy one for my kid but many are disappearing. I don't mean out of stock, I mean removed from sites with no return dates. Lenovo even listed a laptop they just came out with recently as discontinued.

Acer and some others have removed AMD 4000series laptops from their site. Dell has cut back as well. So this is not just 1 OEM seeing this issue but all major ones.

yeah, there is usually a lot of laptop on sale for back to school but there is no stock. I want a 4800H laptop myself...
 

Jimzz

Diamond Member
Oct 23, 2012
4,389
181
106
yeah, there is usually a lot of laptop on sale for back to school but there is no stock. I want a 4800H laptop myself...

Yea I expected sales to be good but also thought many would still turn their noses up at anything AMD. Guess I and AMD were wrong. I was able to get a order in at costco for a 4500u system but think it might get canceled.

Just hope this does not turn OEMs off from AMD chips due to shortage.
 

ASK THE COMMUNITY