Question Raptor Lake - Official Thread

Page 85 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Hulk

Diamond Member
Oct 9, 1999
4,525
2,519
136
Since we already have the first Raptor Lake leak I'm thinking it should have it's own thread.
What do we know so far?
From Anandtech's Intel Process Roadmap articles from July:

Built on Intel 7 with upgraded FinFET
10-15% PPW (performance-per-watt)
Last non-tiled consumer CPU as Meteor Lake will be tiled

I'm guessing this will be a minor update to ADL with just a few microarchitecture changes to the cores. The larger change will be the new process refinement allowing 8+16 at the top of the stack.

Will it work with current z690 motherboards? If yes then that could be a major selling point for people to move to ADL rather than wait.
 
  • Like
Reactions: vstar

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
26,171
15,317
136
If you pretend that 350W is a stock value, then that's simply wrong. We have more than enough confirmation on that by this point.
I know its not stock. What I really want to see is stock vs stock with an aio (and maybe air also), AND max vs max with an AIO, no custom water.

I was commenting that anyone that uses the 350 watt setting and says "it wins" when not comparing against the max from the competitor is unrealistic. I seen this here before. Any "wins" should be at the same conditions with the same cooling. Maybe even the same watts, but in reality, stock vs stock or max vs max is OK. But insane wattage on either test is unacceptable. IMO.

Even LN2 vs LN2 is an OK test, but does not mean much, as nobody can afford to run that except for a suicide run. And custom or chilled water is VERY expensive, so very few people will be using that. We need tests that are what a mild enthusiast can do, not an insane overclocker. That has its place yes, but most people want realistic results.
 

Hulk

Diamond Member
Oct 9, 1999
4,525
2,519
136
Honest question, does Photoshop even benefit from e-cores?

I haven't specifically checked but I doubt it. My point is this. When I'm working in PS I can hear the fan on my CPU spinning up for like 4 or 5 seconds and then shutting down as I work. It's a "bursty" workload. Running the P's at 5+GHz in my opinion and for my usage pattern doesn't make sense for long periods of time. But, when I'm interacting with the computer in a way that waiting for operations slows down my creative workflow I like having the extra compute even at the expense of a momentary loss of efficiency.

In addition, as I've written here many times if I am transcoding video while editing in PS, the e's are efficiently and admittedly rather slowly taking care of this operation in the background while I continue to work with all 8 P's at my disposal. It works well for me.

Having only 4 e's kind of stinks. 16 would be pretty amazing.

Again, this is just my experience. Different strokes for different folks.
 
  • Like
Reactions: DAPUNISHER

DrMrLordX

Lifer
Apr 27, 2000
22,065
11,693
136
I know its not stock. What I really want to see is stock vs stock with an aio (and maybe air also), AND max vs max with an AIO, no custom water.

When AMD's "max" is 230W and Intel's is 350W, that's maybe not a fair comparison unless the target audience is okay with speccing out appropriate cooling. AMD users willing to pony up for the same cooling may be able to use PBO and squeeze out even more performance, who knows?

I haven't specifically checked but I doubt it.

I didn't think so. Stuff like Photoshop already does pretty well on Alder Lake so I doubt that Raptor Lake will show major improvement. Could be wrong though.
 

Det0x

Golden Member
Sep 11, 2014
1,299
4,234
136
If you pretend that 350W is a stock value, then that's simply wrong. We have more than enough confirmation on that by this point.
I think the problem arises when IDL forum posters act like this is performance numbers you will get a with a "stock" 13900k, failing to mention its done with unlimited power limits. (basically overlocked)
How many times have i seen people talk about 40k MT score in Cinebench r23 and not a pip about that 350w power consumption needed and what cooling is required to run that..

I'm totally fine comparing a unlimited Raptor Lake vs a PBO Zen4.
..or stock raptor 253w vs 230w Zen4

But not this 350w vs 230w comparison that is currently doing the rounds

Dont you think that would be a fairer comparison ?








Use of acronym IDL to insult the Intel side. Not allowed, don't do it again.


esquared
Anandtech Forum Director
 
Last edited by a moderator:

coercitiv

Diamond Member
Jan 24, 2014
6,677
14,275
136
I didn't think so. Stuff like Photoshop already does pretty well on Alder Lake so I doubt that Raptor Lake will show major improvement. Could be wrong though.
Photoshop scales well with ST performance for most users, core count increase is a mixed bag, doubly so when you already have 8 P cores. Raptor Lake will show improvements with higher clocks and maybe cache afinity in some benchmarks, but the E core spam will be largely irrelevant for the average PS user.
 
  • Like
Reactions: scineram and Kaluan

Asterox

Golden Member
May 15, 2012
1,039
1,823
136
Honest question, does Photoshop even benefit from e-cores?

No, or example where you can see which CPU has the highest Singlecore performance.

 

Kaluan

Senior member
Jan 4, 2022
504
1,074
106
That s 43% more power, from 240 to 343W, so basically the same perf/watt at these conditions than ADL.
Their 45% was actually spot on (236 to 343 = 45,3% inc)

Raptor Lake may be quite a bit more efficient than Alder Lake under X and Y circumstances, but at least under 100% Intel spec v Intel spec (5,5GHz/4,3GHz P/E core sustained v 4,9/3,9), Raptor Lake comes slightly short of linear scaling. A sign IMHO, that Raptor Lake is (spec-wise) overtuned even more so than Alder.

One also has to factor in the quality of the samples tested. The ADL one (emphasis) seems like a great sample, no idea about the Raptor Lake one. But Intel has been known to seed reviewers with their top silicon before. Not that we have any way of knowing for sure were they got their sample from, but Intel seems to be running a incredibly loose ship... suspicious loose.
 
  • Like
Reactions: DAPUNISHER

Exist50

Platinum Member
Aug 18, 2016
2,452
3,102
136
I think the problem arises when IDL forum posters act like this is performance numbers you will get a with a "stock" 13900k, failing to mention its done with unlimited power limits. (basically overlocked)
How many times have i seen people talk about 40k MT score in Cinebench r23 and not a pip about that 350w power consumption needed and what cooling is required to run that..

I'm totally fine comparing a unlimited Raptor Lake vs a PBO Zen4.
..or stock raptor 253w vs 230w Zen4

But not this 350w vs 230w comparison that is currently doing the rounds

Dont you think that would be a fairer comparison ?
Who in this thread or the original source has claimed those numbers as representative of performance at stock? Where are you claiming to see this behavior? Because as things stand, it reads mostly like a strawman. Ironically so given the effort needed to convince some people about each chip's actual TDP...
 

TESKATLIPOKA

Platinum Member
May 1, 2020
2,523
3,037
136
.....
I'm totally fine comparing a unlimited Raptor Lake vs a PBO Zen4.
..or stock raptor 253w vs 230w Zen4
....
Isn't the max socket power 230 W for AM5?
7900x and 7950x has 170W TDP and 230W PPT, right?
Then even If you activate PBO the limit still will be 230W?

I want to see a comparison between a standard 13900K vs 13900T. How much lower It needs to be clocked to fit in 35W sustained.
AMD most likely won't release such a low TDP product for desktop, so 13900T will be the most efficient CPU for desktop out of the box.
The question is price and If It's worth the lost performance.
 

Det0x

Golden Member
Sep 11, 2014
1,299
4,234
136
Isn't the max socket power 230 W for AM5?
7900x and 7950x has 170W TDP and 230W PPT, right?
Yes default PPT values for the 7950x should be 230PPT according to all leaks.
Stock value for my 5950x is 142w, but i have no problem drawing over 300w from the AM4 socket when i want.. (think ive been upto 360w PPT)

Then even If you activate PBO the limit still will be 230W?
That depends on the "stock motherboard PBO limits".
Simply turning on PBO allows the cpu to consume more than stock 142 PPT limits for the 5950x
I'm very sure the same will be the case with highend x670e motherboard..

Like i said above, i'm all for comparing performance with unlocked/higher powerlimits between Raptor Lake and Zen4, but it seem disingenuous to only allow Intel to run with unlimited power limits compared to a stock PPT Zen4 for the "performance crown".

Either run CPU's at stock powerlimits (253w vs 230w).
..Or if you want to enable the unlimited power option for the 13900k, you should also enable PBO for the 7950x.

But it seem like certain parties dont want this... I wonder why ?
It makes me wonder what the 7950x would score with PBO enabled when we know it reaches 39k at stock 230w.
1663527305642.png

*edit*
I just tested auto PBO limits for my x570s MSI unify x max running a 5950x --> 1000w PPT limit
Unlimited power value for Intel is 2048w i think ? or was it 4k w ?
1663527564614.png
 
Last edited:

biostud

Lifer
Feb 27, 2003
18,700
5,433
136
If that difference in price between N7 dies and N6 dies is even remotely true, and TSMC's own statements about the portability of N7 designs to N6, its almost baffling that AMD hasn't done an N6 version of Zen3 as an in place upgrade to their line of N7 based Zen3 products just to decrease their own cost per working CCD. I realize that there would still be an R&D overhead to such a move, but, in volume, it should be more than made up in short order unless TSMC is flat out lying about the ease of porting those designs. We already see the improvements in Rembrandt's CCX over Cezanne while using what is essentially the same design. With a higher power budget, it seems logical that desktop CCDs should show an even better MT performance improvement.
They do have 6xxx mobile zen3+ APU on 6nm :p
 

DrMrLordX

Lifer
Apr 27, 2000
22,065
11,693
136
As @maddie pointed out, it was and (presumably) still is more cost-effective for AMD to stick to 8c Vermeer CCDs on N7 so long as they continue to produce Milan. If or when that consideration changes somehow, maybe we will see N6 Rembrandt-like CCDs on AM4. It would be a nice followup product to Vermeer, and it would fill in the gap that AMD has left by not selling Rembrandt on AM5 yet.
 
  • Like
Reactions: Kaluan

Kocicak

Golden Member
Jan 17, 2019
1,090
1,141
136
"Informal" reviews from people who did not sign any contract with Intel are not subject to any regulation. If somebody deleted it, it most likely contained some mistakes.