• Hey there Guest! This holiday season you have a chance to win some cool AnandTech swag! Check out “The Holiday Hotlist” giveaway, which you can find here.

[Probably Fake] AMD 2019 Lineup

Shivansps

Platinum Member
Sep 11, 2013
2,261
42
126
#1
This is going around the social networks, it looks fake to me because of 3 important things:
1) Pricing looks as consumer prices not OEM prices
2) Duron Excavator @ 4.9Ghz?
3) Why on earth do you need to place OC numbers on a OEM product list?

So i decided to post it here to see what you guys think about it.
 

mohit9206

Senior member
Jul 2, 2013
925
15
116
#4
Is there going to be a 4 core 8 thread Zen 2? Or just 4/4 and 6/12?
 

EXCellR8

Platinum Member
Sep 1, 2010
2,631
15
126
#5
Ryzen 9 for $500...
 

BigDaveX

Senior member
Jun 12, 2014
303
8
101
#6
Why would they resurrect the Duron name after ~15 years, when Sempron has been their budget brand for most of that time?

Aside from that, this list feels like wishful thinking. A 64C Threadripper with a 5GHz base clock? Just seems a little hard to believe, know what I'm saying?
 
Apr 27, 2000
10,192
112
126
#7
Okay:

1). 20 CU on 7nm? This product shows up on exactly zero of their past roadmaps. AMD has already told us that the first 7nm APUs will show up in, what, 2020? Picasso, a 12nm APU, appears in 2019.
2). The boost clocks seem unrealistic given core counts. I mean, why would you reserve the highest boost clocks for the chips with higher core counts? XFR2 doesn't even have that kind of restriction, and certainly XFR3 won't either. I would expect boost clocks to be segmented between CPUs that feature XFR3 support and those that do not (namely, the non-X parts). I will observe that the exception here is the 2950X of the current lineup which has a listed max boost of 4.4 GHz (as opposed to the 2700x which boosts to 4.3 GHz), but that is more due to the binning of Threadripper dice than anything else. The 2990WX certainly has lower boost clocks than the 2950X!
3). The "base" clocks seem completely ridiculous. 5 GHz base for the 64t Threadripper product? Completely absurd. And really, 4.3 GHz as the base clock for the 3800x seems silly as well.
4). Dual socket for Threadripper? What?
5). XV running @ 5 GHz on 28nm? No.
 

scannall

Golden Member
Jan 1, 2012
1,250
47
136
#9
As was pointed out, official docs aren't going to have estimated overclocking for example. And aside from rampant speculation there is no indication of more than 8 cores on the consumer products. Further, 20 CU's seems pretty iffy as well.
 

BigDaveX

Senior member
Jun 12, 2014
303
8
101
#10
As was pointed out, official docs aren't going to have estimated overclocking for example. And aside from rampant speculation there is no indication of more than 8 cores on the consumer products. Further, 20 CU's seems pretty iffy as well.
I actually could see AMD making an APU with 20 CUs, given that it'd be four fewer CUs than are used by the Subor Z+'s custom APU on a smaller manufacturing process. However, it'd more likely be a high-end laptop chip, not a $239.99 desktop chip.
 

NostaSeronx

Platinum Member
Sep 18, 2011
2,188
9
106
#11
My list:

EPYC2&Castle Peak (X499?) -> 7nm Chiplet
Matisse&Picasso => 12nm standard: chip w/ IO.

However, Vermeer and Renoir are not launching in 2020, but in 2019.
October-December 2019 => Vermeer/Renoir <-- 7FF TSMC
January-March 2019 => Matisse/Picasso <-- 12LP(still 78CPP/9T) refresh/rebrand

22FDX/12FDX roadmap is supposedly going to be announced May 2019, with 22FDX coming July 2019.
 

Batmeat

Senior member
Feb 1, 2011
669
3
81
#12
Okay:

1). 20 CU on 7nm? This product shows up on exactly zero of their past roadmaps. AMD has already told us that the first 7nm APUs will show up in, what, 2020? Picasso, a 12nm APU, appears in 2019.
2). The boost clocks seem unrealistic given core counts. I mean, why would you reserve the highest boost clocks for the chips with higher core counts? XFR2 doesn't even have that kind of restriction, and certainly XFR3 won't either. I would expect boost clocks to be segmented between CPUs that feature XFR3 support and those that do not (namely, the non-X parts). I will observe that the exception here is the 2950X of the current lineup which has a listed max boost of 4.4 GHz (as opposed to the 2700x which boosts to 4.3 GHz), but that is more due to the binning of Threadripper dice than anything else. The 2990WX certainly has lower boost clocks than the 2950X!
3). The "base" clocks seem completely ridiculous. 5 GHz base for the 64t Threadripper product? Completely absurd. And really, 4.3 GHz as the base clock for the 3800x seems silly as well.
4). Dual socket for Threadripper? What?
5). XV running @ 5 GHz on 28nm? No.
This —-^
 

Spartak

Senior member
Jul 4, 2015
203
30
86
#15
My list:

EPYC2&Castle Peak (X499?) -> 7nm Chiplet
Matisse&Picasso => 12nm standard: chip w/ IO.

However, Vermeer and Renoir are not launching in 2020, but in 2019.
October-December 2019 => Vermeer/Renoir <-- 7FF TSMC
January-March 2019 => Matisse/Picasso <-- 12LP(still 78CPP/9T) refresh/rebrand

22FDX/12FDX roadmap is supposedly going to be announced May 2019, with 22FDX coming July 2019.
Your fantasy is running prime95 again I see.

Matisse is Zen2, there is zero chance this will be released on 12LP. Picasso will be the 12nm Zen+ APU update for 19Q1-2 and Matisse will be introduced in 19Q4. Renoir is Zen2 APU that follows in 20Q1-2 and for Vermeer you are a full year off.
 

HurleyBird

Golden Member
Apr 22, 2003
1,656
11
106
#16
Probably fake, but I'm all for dumping the lame Sempron name and bringing back Duron.
 
Oct 10, 1999
64,781
59
126
#17
Would be cool if true, but....
 

dark zero

Platinum Member
Jun 2, 2015
2,488
0
61
#19
The Duron/Athlon part I see it very real... But not going over 4.9 Ghz. That is real fake.

Duron going excavator? Is OK, but can't reach over 4Ghz at 28nm, the only chance to go up that part is going on 22 FDX. @NostaSeronx , your dream is true then.

The other Duron is Sempron in fact and makes sense.
 

Similar threads


ASK THE COMMUNITY