Question Raptor Lake - Official Thread

Page 44 - 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
4,223
2,013
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

tamz_msc

Diamond Member
Jan 5, 2017
3,795
3,626
136
I edited the original post. Yes, the 7800X with PBO will be able to reach 2,300 points in ST on GB5
PBO does not affect ST performance as far as I am aware; and no, it will take the 7950X, assuming it actually has a 5.7 GHz boost according to the WCCFTech rumor, to match this Raptor Lake score in GB5 ST. That assumes Geekbench scales perfectly with frequency.
 
  • Like
Reactions: controlflow

nicalandia

Diamond Member
Jan 10, 2019
3,330
5,281
136
PBO does not affect ST performance as far as I am aware; and no, it will take the 7950X, assuming it actually has a 5.7 GHz boost according to the WCCFTech rumor, to match this Raptor Lake score in GB5 ST. That assumes Geekbench scales perfectly with frequency.

This PBO 5800X it's getting 1718 at 4.6 Ghz Boost. You can extrapolate the difference on Speed and IPC and it should be a match to the Raptor Lake ST score.

1659800982792.png
 
  • Haha
Reactions: controlflow

tamz_msc

Diamond Member
Jan 5, 2017
3,795
3,626
136
This PBO 5800X it's getting 1718 at 4.6 Ghz Boost. You can extrapolate the difference on Speed and IPC and it should be a match to the Raptor Lake ST score.

View attachment 65469
I have done the extrapolation. It will take the 7950X @ 5.7 GHz to match the ST integer score of the 13900K sample, assuming perfect scaling with frequency.

Do not look at the overall score - it includes AES, which is irrelevant for 99% use cases.
 

Asterox

Golden Member
May 15, 2012
1,026
1,775
136
People on most on these sites long ago stopped looking at applications people actually use and benchmarks that were really relevant.

If people really looked at what was relevant to what they do, they'd probably just be looking at 4 areas. Web browser performance, MS Office performance, gaming performance, and overall system performance like PCMark - which also give one a good idea of how good a particular chipset / motherboard will work.

But no, cinebench and pov-ray are like the first thing up on every review. I see it as part of a cognitive dissonance that has infected the whole PC media over the last few years. Relative to those other categories, virtually no one uses cinebench or pov-ray.

I don't use any of those media apps listed, not one of them. But I do use a web browser...

View attachment 65467

Also you can try Jetstream, related to the below first link or comparison.Browser or browsing testing can be complicated.

- any Antivirus will slowdown your browsing vs Windows+browsing without active or installed Antivirus

- stock light browser(only two or three addons), vs browser crammed with ton of addons big difference



Old comparison R5 5600X vs i5 12400 also has intresting results.

 

pakotlar

Senior member
Aug 22, 2003
731
187
116
People expecting AMD to be sandbagging and get >>35% avg ST and >>40% avg MT uplift at stock to beat raptor lake are setting themselves up for disappointment. 7950X and 13900K look like they will be nigh indistinguishable both in terms of power and performance, 5-10% either way on all metrics (my guess is 13900K faster but with slightly worse efficiency). Fanboys will buy what their favorite company and claim the other solution is doggerel, everyone else should flip a coin or look at platform features (like longevity and AVX512 which is why I will look at 7950X most closely).

It’s wonderful that after Krzanich years of stagnation we have 2 fierce competitors and 2 good solutions.
 

Det0x

Golden Member
Sep 11, 2014
1,028
2,953
136
OK now if 2300 ST can be had without too much trouble, Raptor Lake starts looking attractive.
Seems like 2300ST is already reachable today with a Alder Lake @ 5.5ghz
1659806329207.png
LIMITATIONS
  • Use Geekbench 5.4.5. and HWinfo v7.26
  • Maximum Frequency/cache limitation 5500MHz
  • Disabling CPU cores/HT/SMT NOT allowed.
  • A VALID Geekbench 5 link is required.
  • A CPUZ 2.01 or newer Validation link is required, registered on your HWBOT username.
  • A verification screenshot is required, using the official wallpaper, GB 5 score, CPUZ tabs for CPU & Memory and HWinfo.
  • Only members of the rookie, novice, enthusiast league may participate.
  • No Extreme cooling allowed (chiller, Single Stage, Cascade, Dry ice, LN2)

 
Last edited:
  • Like
Reactions: lightmanek

pakotlar

Senior member
Aug 22, 2003
731
187
116
Talk about Being Delusional...:rolleyes:
At least on my 5950X PPT is pegged under load to 100% of the bios limit. I expect 7950X to be the same. That PPT is 230W. The PL2 on 13900K will be 250W. That’s 10%. I expect raptor lake to be faster and take 10% more power. Maybe a bit worse than that.

Given your aggressive word choice, what are you expecting and why?
 

shady28

Platinum Member
Apr 11, 2004
2,520
397
126
People expecting AMD to be sandbagging and get >>35% avg ST and >>40% avg MT uplift at stock to beat raptor lake are setting themselves up for disappointment. 7950X and 13900K look like they will be nigh indistinguishable both in terms of power and performance, 5-10% either way on all metrics (my guess is 13900K faster but with slightly worse efficiency). Fanboys will buy what their favorite company and claim the other solution is doggerel, everyone else should flip a coin or look at platform features (like longevity and AVX512 which is why I will look at 7950X most closely).

It’s wonderful that after Krzanich years of stagnation we have 2 fierce competitors and 2 good solutions.


Chipset bandwidth and features are actually what I am probably going to look at closely this time.

A lot of people don't look at those factors, but if you have more than just a simple setup like one m.2 and kb/mouse, it can make a difference.

The Z490 I have is miserable in this regard due to the DMI 3.0 x4 lanes to the chipset. A DMI 3.0 lane is basically like a PCIe 3.0 lane, bandwidth is ever so slightly faster on DMI but call it the same. Everything off the chipset, comes down to that interface.

Z590 was far better, with DMI 4.0 x4 (twice the bandwidth) and Z690 expanded to DMI 4.0 x8. From that perspective Z590 had basically the same bandwidth capability to the chipset as AMDs X570, which uses PCIe 4 x4 to its chipset, while the Z690 has significantly greater bandwidth (twice as much) by using twice as many lanes.

From what I can tell, Zen 4 will have PCIe 5 x4 to the X670 which matches the Z690's DMI 4.0 x8. On X670E, they may have PCIe 5.0 x4 to each of its two chipsets, but that isn't certain. Zen 4 has 24 PCIe 5.0 lanes, so 16 are used for GPU, and if you have two chipsets than presumably the other 8 are used on the two chipsets - so no PCIe 5.0 x4 off the CPU to an m.2 slot, it's going to be shared.

Alder Lake has 16 PCIe 5.0 lanes, 4 PCIe 4.0 lanes off the CPU, plus 8 DMI 4.0 lanes - this is equivalent to having 16 PCIe 5.0 and 12 PCIe 4.0 lanes.

I am hoping that the Z790 bumps DMI up to 5.0 and keeps 8 lanes to the chipset. If it does that, then a Z790 will be superior to an X670E.

At the risk of ticking off a certain group of people, I should probably point out that the AMD B550 is one of the worst at this now. That chipset is still on a PCIe 3.0 x4 link to the CPU. Regardless of what you can connect to the board, it's all going to get packed into that interface, and that sucks. It's exactly like having a Z170 from many years ago.

The Intel B660M is actually a DMI 4.0 x4 interface to the chipset - same as an X570, and way way better for supporting PCIe 4 x4 m.2 and multiple USB 3.2 v2 and v2x2 off that shipset.

1659807931250.png
 
Last edited:

Abwx

Lifer
Apr 2, 2011
10,947
3,457
136
. I expect raptor lake to be faster and take 10% more power. Maybe a bit worse than that.

At 10% more power it will still be vastly less efficent as well as lower performing.
There s no real core count advantage for RPL since one Zen 4 C + SMT has about as much throughput/Hz than two E cores.

The differenciation is in the full node advantage for Zen 4, expect RPL to use 350W as a mean to match a 7950X@230W.
 
  • Like
Reactions: Kaluan

shady28

Platinum Member
Apr 11, 2004
2,520
397
126
  • Like
Reactions: controlflow

pakotlar

Senior member
Aug 22, 2003
731
187
116
If that is true AMDs high end chipsets are garbage. I can't imagine they would do this. All the other connectivity they list here I would call lipstick on a pig.

View attachment 65478

It's all constrained at the same point as the X570 - and inferior to a Z690 :

View attachment 65479

Probably a practical step to get the consumer chipset out. We’ll probably get future boards with better IO. Consumer increasingly looks like not AMDs focus, which makes me a bit sad as I don’t want to spend money on workstation or server parts.
 

pakotlar

Senior member
Aug 22, 2003
731
187
116
At 10% more power it will still be vastly less efficent as well as lower performing.
There s no real core count advantage for RPL since one Zen 4 C + SMT has about as much throughput/Hz than two E cores.

The differenciation is in the full node advantage for Zen 4, expect RPL to use 350W as a mean to match a 7950X@230W.

You do realize that they could not possibly set a 350W PL2 and retain compat with boards that are specced to 230W PL2? 250W was probably chosen to ensure they didnt break existing board support.

Furthermore the PL4 has dropped by around 50W.
 

Abwx

Lifer
Apr 2, 2011
10,947
3,457
136
You do realize that they could not possibly set a 350W PL2 and retain compat with boards that are specced to 230W PL2? 250W was probably chosen to ensure they didnt break existing board support.

Furthermore the PL4 has dropped by around 50W.

230W is for AM5 MBs, there s no such limitation for Intel s MBs, it s up to the MB manufacturers to set a limit, if the CPU exceed the motherboard limitation it s throttled.
 

shady28

Platinum Member
Apr 11, 2004
2,520
397
126
Probably a practical step to get the consumer chipset out. We’ll probably get future boards with better IO. Consumer increasingly looks like not AMDs focus, which makes me a bit sad as I don’t want to spend money on workstation or server parts.

All of them are fine if you just have a CPU, a GPU, and one m.2 SSD. That can all be driven from the CPU and they all have plenty of bandwidth for a wireless card or NIC off the chipset.

But many have more than one m.2, a couple of SATA drives, and use the USB ports for something other than a keyboard/mouse. I have a 4TB external USB 3.1, and right now have a 128GB thumb drive plugged in along with two SATA drives.

When I run backups across these devices, I can see the performance drop due to the Z490's DMI 3.0 x4.

I actually believe that is a far more common scenario than say, running cinebench.
 

Abwx

Lifer
Apr 2, 2011
10,947
3,457
136
230W is the PL2 on high end 12th gen Intel parts, nothing to do with AM5. Raptor Lake is 250W.
350W is a number you made up, because you have no idea what you’re talking about.

To get 25-30% higher perfs than ADL they need to add 50W for the 8 added e cores, we re already at 300W with same frequencies for e and P cores than RPL...

And then if they want to clock everything a little higher than RPL, both e and P cores, another 50W is really not a lot...
 

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
To get 25-30% higher perfs than ADL they need to add 50W for the 8 added e cores, we re already at 300W with same frequencies for e and P cores than RPL...

And then if they want to clock everything a little higher than RPL, both e and P cores, another 50W is really not a lot...

This is assuming Raptor Lake has the exact same power scaling as Alder Lake, which by all accounts isn't true. Raptor Lake will have better performance per watt than Alder Lake (likely through node tweaks), which is why they are increasing core count and frequency.
 

pakotlar

Senior member
Aug 22, 2003
731
187
116
All of them are fine if you just have a CPU, a GPU, and one m.2 SSD. That can all be driven from the CPU and they all have plenty of bandwidth for a wireless card or NIC off the chipset.

But many have more than one m.2, a couple of SATA drives, and use the USB ports for something other than a keyboard/mouse. I have a 4TB external USB 3.1, and right now have a 128GB thumb drive plugged in along with two SATA drives.

When I run backups across these devices, I can see the performance drop due to the Z490's DMI 3.0 x4.

I actually believe that is a far more common scenario than say, running cinebench.

I agree
This is assuming Raptor Lake has the exact same power scaling as Alder Lake, which by all accounts isn't true. Raptor Lake will have better performance per watt than Alder Lake (likely through node tweaks), which is why they are increasing core count and frequency.

Yep, this. Probably not node tweaks, but maybe better yields let them bin chips for lower voltage, probably what you mean. Cores will also require a few watts less when accessing memory (since they’ll spend more time in L2), and there could be a number of IMC and other uncore tweaks to further reduce power consumption. We’ll see. Rumors could also be wrong and they really do target 300W or more, but that would be a disaster both in terms of lack of support from 12th gen mobos that have weaker power delivery, and in terms of heat generation (and presumably mobile battery life).

Given people’s reports on how well ADL undervolts, my guess is that to ensure high ADL yield on Intel 7, they set very conservative voltages on higher end parts. I think Rocket Lake will be a nice step forward in efficiency, but we’ll see, nothing known until launch.
 
Last edited: