Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

Page 223 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

itsmydamnation

Diamond Member
Feb 6, 2011
3,055
3,862
136
Is that specifically a characteristic of RDNA?
No its a trait of intermediate renderers, not that modern AMD/NV are true intermediate renderers any more. Tile based renderers are supposed to be more cache friendly , but as i said before the lines are really blurry these days. AMD/NV started to become dynamic ? renderers somewhere around maxwell / GCN 1.x.
 
Jul 27, 2020
26,612
18,313
146

Timmah!

Golden Member
Jul 24, 2010
1,568
924
136
Should i expect Zen5 reveal in few days during CES? Or not happening until Computex later this year?
 

Hitman928

Diamond Member
Apr 15, 2012
6,660
12,298
136
Should i expect Zen5 reveal in few days during CES? Or not happening until Computex later this year?

I'd be surprised if they don't have something on Zen5 at CES, but meaningful information I don't know. If, as some have suggested, Zen5 will launch during spring time, then Computex would be too late to hear about it and I would expect solid info at CES. If they aren't launching until end of 1H24, then Computex would be the perfect time to do a full announcement.
 

FlameTail

Diamond Member
Dec 15, 2021
4,384
2,761
106
I'd be surprised if they don't have something on Zen5 at CES, but meaningful information I don't know. If, as some have suggested, Zen5 will launch during spring time, then Computex would be too late to hear about it and I would expect solid info at CES. If they aren't launching until end of 1H24, then Computex would be the perfect time to do a full announcement.
I think Strix Point is what will be announced at Computex. According to the leaked roadmap, it will come in mid-2024, so perfectly in time for Computex?

Then Zen 5 should come earlier than that, so I guess CES yes.
 

Ajay

Lifer
Jan 8, 2001
16,094
8,112
136
I think Strix Point is what will be announced at Computex. According to the leaked roadmap, it will come in mid-2024, so perfectly in time for Computex?

Then Zen 5 should come earlier than that, so I guess CES yes.
CES *seems* early, so I wonder if AMD will hold their own special event.
 
  • Like
Reactions: Tlh97 and Saylick

Hitman928

Diamond Member
Apr 15, 2012
6,660
12,298
136
AMD has an event on January 8th at 7am PST/ 9am CST/ 10am EST / 3pm GMT:


Appears to be a repeat of the last AI focused event, but could possibly include some new announcements.

Looks like AMD doesn't have a keynote session this year so I don't expect there will be meaningful Zen5 info to come out of CES. They will probably have some invite only, closed door session where they show some stuff but probably nothing for the public.
 

RTX2080

Senior member
Jul 2, 2018
334
533
136
Should i expect Zen5 reveal in few days during CES? Or not happening until Computex later this year?
Since these companies are all about AI recently I don't know if there would be any forecast reveal about Zen5 client at CES, maybe it would be released all of a sudden in the coming months? Client cpu is the least important thing right now. All I know is mass production has been started and produced smallish amount of Zen5 cpus two months ago.
 
Last edited:
  • Like
Reactions: lightmanek

Thibsie

Golden Member
Apr 25, 2017
1,094
1,275
136
Since these companies are all about AI recently I don't know if there would be any forecast reveal about Zen5 client at CES, maybe it would be released all of a sudden in the coming months? Client cpu is the least important thing right now. All I know is mass production has been started and produced smallish amount of Zen5 cpus two months ago.

From a revenue PoV, clearly.
From a marketing PoV though, it is different. And AMD can't benefit so much from Radeon image than Nvidia with RTX.
 

Hans Gruber

Platinum Member
Dec 23, 2006
2,507
1,345
136
It's in AMD's best interest to release Zen5 ASAP. Intel has a better IPC, bad power efficiency and Intel has 20A coming later in the year. It's also in AMD's interest to scrap APU's and put in really good RDNA3 Integrated graphics with their Zen5 CPU's. I hope they go back to 65w and 105w TDP for Zen5. That is the sweet spot for power efficiency.
 

TESKATLIPOKA

Platinum Member
May 1, 2020
2,696
3,260
136
It's in AMD's best interest to release Zen5 ASAP. Intel has a better IPC, bad power efficiency and Intel has 20A coming later in the year. It's also in AMD's interest to scrap APU's and put in really good RDNA3 Integrated graphics with their Zen5 CPU's. I hope they go back to 65w and 105w TDP for Zen5. That is the sweet spot for power efficiency.
I think they know better than you what they should or shouldn't do.
BTW, didn't you hear about Strix Halo? There you will get a pretty strong IGP, but you should expect 100W or higher TDP.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
27,115
16,027
136
It's in AMD's best interest to release Zen5 ASAP. Intel has a better IPC, bad power efficiency and Intel has 20A coming later in the year. It's also in AMD's interest to scrap APU's and put in really good RDNA3 Integrated graphics with their Zen5 CPU's. I hope they go back to 65w and 105w TDP for Zen5. That is the sweet spot for power efficiency.
Better IPC in what vs what chip ? desktop, NO, server NO, HEDT NO, laptop (7840u vs several different Intel) NO. Care to qualify that ?
 
  • Like
Reactions: Tlh97 and Jimzz

StefanR5R

Elite Member
Dec 10, 2016
6,593
10,392
136
I hope they go back to 65w and 105w TDP for Zen5. That is the sweet spot for power efficiency.
Several parts with 65 W and 105 W default TDP¹ exist in the Ryzen 7000 line, and the Ryzen PRO 7000 line consists entirely of parts with 65 W default TDP. Also, there is still ECO mode and the ability to reconfigure TDP and PPT limit yourself.
As for Zen 5, per-core power budgets which make most sense in given scenarios are bound to go up from Zen 4 to 5 due to the latter's wider microarchitecture.

________
¹) a.k.a. 88 W and 142 W default PPT limit