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

Page 857 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

CouncilorIrissa

Senior member
Jul 28, 2023
561
2,212
96

Hitman928

Diamond Member
Apr 15, 2012
6,328
11,132
136
Yes, but they have a lot more TSVs than what Zen5 does. The better BSPD implementations have ~a TSV for every transistor, literally billions of them per die. Zen5 has ~tens of thousands.

Right, because obviously N4P doesn't have BSPD support yet but theoretically that doesn't stop you from using TSV's to bring power into a die, just like they used TSVs to bring power into the cache chiplet in Zen 3 and Zen 4. Because the TSVs are much bigger, you can power many FETs off of them, but the disadvantage is that instead of a density increase, you get a density decrease because you have to bring the power up to a normal metal line and then distribute it to multiple FETs surrounding the TSV through the normal metal/contact routing. However, like I said previously, I wasn't thinking about the die sharing issue so it makes this approach highly doubtful. I also thought that someone had pointed out that the TSVs had been distributed into more of the die now rather than just in the L3 area, but I guess I was mistaken about that.

Putting that aside, there is still no explanation for what the TSVs are doing if indeed the CCD is on top and flip chip.

What if 9950X3D CCD is sandwiched between V-cache dies?

Theoretically possible, but I thought we had leaks that already showed the cache amount being the same as prior generations? Plus, that would significantly increase costs and decrease production throughput on an already slower process, so I highly doubt this is the case.
 
Reactions: Joe NYC

StefanR5R

Elite Member
Dec 10, 2016
5,968
8,941
136
9800X3D listed, 5.2 boost.
Recent GB6 runs also top out at 5.25.
This *may* all be just guesswork, based on claims from various 3rd parties, not from original information. (Including the part number 100-100001084WOF.)

They chose to list the stepping as "GNR-B0" BTW.
 

moinmoin

Diamond Member
Jun 1, 2017
5,094
8,098
136
This *may* all be just guesswork, based on claims from various 3rd parties, not from original information. (Including the part number 100-100001084WOF.)

They chose to list the stepping as "GNR-B0" BTW.
Yeah, it's very likely placeholder data that is being corrected as more info comes out.
 

arcky

Junior Member
Oct 21, 2024
14
50
46
Reactions: Tlh97 and Mopetar

Saylick

Diamond Member
Sep 10, 2012
3,605
8,075
136
9800x3d delidded, no V-Cache in sight

View attachment 110431

Rather odd that they didn’t flag this as an “exclusive”, which they usually do in the past to take credit for being the first source… but I’ll give credit where credit is due and say that they rightfully have the first leak.

But then you see idiotic statements like this and you’re reminded of their incompetence.
The most interesting thing about this new stacking design is that it opens the room for another 3D V-Cache stack to be employed over the Zen 5 CCD, so you could theoretically get two stacks. This would make for a true stacked chiplet design for the desktop segment. Going such a route will be a bit tricky since power, cost, and thermals have to be kept in mind, but there is definitely a possibility, and maybe we will see such products in the future.
Now why the heck would AMD sandwich the CCD between TWO cache dies?! C’mon! If you want to have multiple cache dies, you either stack them all on top or on the bottom because of how TSVs work. JFC, WTFTech writers are given a sneak peak and they come to the silliest of conclusions. *facepalm*

Edit: Wow, I found the one sane person in the comments section, and they happen to have the same thoughts as me. Of course, stupid Hassan deflects blame by basically saying “Oh, poor me. Please don’t shoot the messenger, that’s just what my source told me!” I swear to God, these hacks don’t deserve getting exclusive leaks because more than half the time they don’t use critical thinking. It’s like that opening scene in Rush Hour where Chris Tucker’s character, who is a detective, is trying to buy weapons from a guy and gets shown C4 in the back of the guy’s car and he’s all like, “Wow, that’s beautiful, man…. What am I looking at?”

 
Last edited:

Hitman928

Diamond Member
Apr 15, 2012
6,328
11,132
136

Hitman928

Diamond Member
Apr 15, 2012
6,328
11,132
136
Rather odd that they didn’t flag this as an “exclusive”, which they usually do in the past to take credit for being the first source… but I’ll give credit where credit is due and say that they rightfully have the first leak.

But then you see idiotic statements like this and you’re reminded of their incompetence.

Now why the heck would AMD sandwich the CCD between TWO cache dies?! C’mon! If you want to have multiple cache dies, you either stack them all on top or on the bottom because of how TSVs work. JFC, WTFTech writers are given a sneak peak and they come to the silliest of conclusions. *facepalm*

Edit: Wow, I found the one sane person in the comments section, and they happen to have the same thoughts as me. Of course, stupid Hassan deflects blame by basically saying “Oh, poor me. Please don’t shoot the messenger, that’s just what my source told me!” I swear to God, these hacks don’t deserve getting exclusive leaks because more than half the time they don’t use critical thinking. It’s like that opening scene in Rush Hour where Chris Tucker’s character, who is a detective, is trying to buy weapons from a guy and gets shown C4 in the back of the guy’s car and he’s all like, “Wow, that’s beautiful, man…. What am I looking at?”

View attachment 110432

Sandwiching the CCD in between Vcache dies is theoretically possible. From a stacking perspective, there's no technical reason you couldn't do it, but you'd have to have 2 different cache chiplet designs. In Zen 3 and Zen 4, the CCD already sits beneath the Vcache chiplet, so the thermals being prohibitive of this approach is not really valid. With that said, I think this approach is unlikely, even if theoretically possible, but who knows what AMD is researching in their labs.
 

Hans de Vries

Senior member
May 2, 2008
338
1,145
136
www.chip-architect.com
9800x3d delidded, no V-Cache in sight

View attachment 110431


What you see on top is the "Support silicon to match the 2D CCD height"
The 3D V-Cache is below that (Red) and the thinned CCD (blue) is at the bottom.
( The slide can be found on the previous page )

 

dr1337

Senior member
May 25, 2020
420
701
136
9800x3d delidded, no V-Cache in sight
From a quick google this is exactly what other X3D parts look like, both 5800X3D and 7800X3D just look like solid black squares with no apparent structural silicon.

I'm pretty sure those special pictures where you can see through the dies are either photoshop or use some kinda IR/UV/Xray photography.
 

Hans de Vries

Senior member
May 2, 2008
338
1,145
136
www.chip-architect.com
The only change which seems possible is that they have flipped the 3D V-cache die so that the TSV's of the CCD now connect directly to the top metal layer of the 3D V-cache.

In this case the 3D V-cache die itself would no longer need TSV's with the result that the total TSV length is halved, reducing the resistance by a factor two.
 

Hitman928

Diamond Member
Apr 15, 2012
6,328
11,132
136
Using the images of Frizchens Fritz:

This shows the signal / power interconnects on the AM5 Zen 5 substrate and the ball grid array of the Zen 5 CCD that goes on it.

View attachment 110435

I do not know how you would put the 3D V-Cache die between those two ???

Maybe AMD engineers can do wonders?


That makes more sense rather than the TSVs to nowhere. It is interesting though that the TSVs seem to no longer have the little logic circuits next to any of them.
 
Reactions: Tlh97 and Joe NYC

Hitman928

Diamond Member
Apr 15, 2012
6,328
11,132
136
The only change which seems possible is that they have flipped the 3D V-cache die so that the TSV's of the CCD now connect directly to the top metal layer of the 3D V-cache.

In this case the 3D V-cache die itself would no longer need TSV's with the result that the total TSV length is halved, reducing the resistance by a factor two.

Wasn’t it already face to back stacking?

 

Saylick

Diamond Member
Sep 10, 2012
3,605
8,075
136
what is the point of v-cache when big part of die is igpu which is not necessary?
???? Wat. The iGPU is on the IO die.

The point of V-cache is to increase the odds of an L3 cache hit, which reduces the demand on RAM. This has performance and power consumption benefits for workloads that don’t fit within the normal 32 MB of L3.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |