• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Question Raptor Lake - Official Thread

Page 192 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Hulk

Diamond Member
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.
 

This isn't new news, but Videocardz has the full lineup of Raptor Lake Refresh. All the way down to a 2+0 "Intel Processor".
 
Profanity is not permitted in the tech forums
To be fair, CS2 in its current state is a mess, it's getting updated constantly, and drivers are still being optimized for it.

What holds true for CS2 today may or may not in six months.

Big CS fan here. Runs pretty much like shit, now lag has improved a bit but it's still in a pretty bad state.
 
Agree, I have a few RPL and ADL rigs too and they have been absolutely rock solid and performant. I really don't get the complaints about the hybrid architecture. There were some early issues but they are very very few and far in between now. I would expect that to only get better in the future.
Garbage software like Denuvo isn't a good argument for why hybrid architecture isn't viable.


DRM sucks and is garbage.

However software that is good like Star Citizen and Counterstrike 2 also have issues which means the hybrid arch is not good.

We have been in SMP world for like 20 years now and the ecosystem is not ready for the most major change in a long while that was unnecessary. Just get more P cores no e-waste cores. AMD is able to do it with similar performing cores why not Intel. Oh cause their node is too big.
 

Hopefully we should see some benchmarks soon!
We pretty well know what the performance will be, except for the 14700k, which gets 4 more e cores. The only question is will there be any improvement in power usage due to slight process or binning improvements.
 
Maybe a more refined IMC that allows much easier OC or tighter RAM timings? One can hope.
I’m not expecting much in the way of memory bandwidth improvements.

The memory stability above DDR5-7200 isn’t just dependent on the CPU. A lot of the performance is dependent on the motherboard vendor and the memory IC themselves. Signal reflection is the primary limiting factor once you get to the higher memory speeds.

For anybody installing RPL-R into a Z690 board there will be practically no difference. Same holds true with people using IC’s that aren’t Hynix.

If you’ve got a solid 2-Dimm Z790 motherboard you’ll probably be able to push an extra ~400MTs but that’s probably it.

I don’t think we’ll be moving the needle much until ARL introduces CKD.
 
Raptor Refresh on Tuesday, right?

Have we ever had 3 generations of Intel processors that run on the same mobo?

I had a 12700K. Then a 13900K. I got a bad part as it was restarting under load without voltage increase. Intel warrantied it. Now I'm running a 13600K, which actually has been quite nice. If pricing is right and it can hit same clocks as 13900K at lower power I might pick up a 14900K at my local MC. This will be the last CPU for this board so I might as well go to the top so I can hold off on another build longer.
 

Here's the official slides. RCP is the same compared to 13th Gen.
So basically Intel has a better stepping here. They had two options, add a "50" to the number of the 13th gen series or call this a new generation. They chose the latter I assume because they want 14th gen desktop and mobile to launch more or less around the same time.
 
7950X
7950X3D
14900KF
13900KS
13900K
Integer Math
228,193 MOps/Sec
214,554 MOps/Sec​
219,583 MOps/Sec​
213,857 MOps/Sec​
211,279 MOps/Sec​
Floating Point Math
139,180 MOps/Sec​
130,149 MOps/Sec​
151,639 MOps/Sec​
156,928 MOps/Sec
154,471 MOps/Sec​
Find Prime Numbers
356 Million Primes/Sec​
501 Million Primes/Sec
268 Million Primes/Sec​
265 Million Primes/Sec​
243 Million Primes/Sec​
Random String Sorting
102,433 Thousand Strings/Sec
95,637 Thousand Strings/Sec​
84,188 Thousand Strings/Sec​
93,125 Thousand Strings/Sec​
90,373 Thousand Strings/Sec​
Data Encryption
49,776 MBytes/Sec
47,155 MBytes/Sec​
44,390 MBytes/Sec​
48,649 MBytes/Sec​
47,660 MBytes/Sec​
Data Compression
842,542 KBytes/Sec
785,435 KBytes/Sec​
721,262 KBytes/Sec​
829,269 KBytes/Sec​
812,358 KBytes/Sec​
Physics
3,102 Frames/Sec​
5,139 Frames/Sec
3,733 Frames/Sec​
3,516 Frames/Sec​
3,116 Frames/Sec​
Extended Instructions
62,383 Million Matrices/Sec
58,486 Million Matrices/Sec​
42,935 Million Matrices/Sec​
48,928 Million Matrices/Sec​
47,539 Million Matrices/Sec​
Single Thread
4,312 MOps/Sec​
4,168 MOps/Sec​
4,939 MOps/Sec
4,764 MOps/Sec​
4,661 MOps/Sec​
Average CPU Mark​
63335
62653​
59384​
61860​
59696​
 
PASSMARK SCORES ABOVE ^^^

7950X wins a lot.

7950X3D cache just kills it in Find primes and Physics.

13900KS still claims a win in FP and won't go down without a fight, humiliating 14900KF in quite a few of the tests.

 
7950X
7950X3D
14900KF
13900KS
13900K
Integer Math
228,193 MOps/Sec
214,554 MOps/Sec​
219,583 MOps/Sec​
213,857 MOps/Sec​
211,279 MOps/Sec​
Floating Point Math
139,180 MOps/Sec​
130,149 MOps/Sec​
151,639 MOps/Sec​
156,928 MOps/Sec
154,471 MOps/Sec​
Find Prime Numbers
356 Million Primes/Sec​
501 Million Primes/Sec
268 Million Primes/Sec​
265 Million Primes/Sec​
243 Million Primes/Sec​
Random String Sorting
102,433 Thousand Strings/Sec
95,637 Thousand Strings/Sec​
84,188 Thousand Strings/Sec​
93,125 Thousand Strings/Sec​
90,373 Thousand Strings/Sec​
Data Encryption
49,776 MBytes/Sec
47,155 MBytes/Sec​
44,390 MBytes/Sec​
48,649 MBytes/Sec​
47,660 MBytes/Sec​
Data Compression
842,542 KBytes/Sec
785,435 KBytes/Sec​
721,262 KBytes/Sec​
829,269 KBytes/Sec​
812,358 KBytes/Sec​
Physics
3,102 Frames/Sec​
5,139 Frames/Sec
3,733 Frames/Sec​
3,516 Frames/Sec​
3,116 Frames/Sec​
Extended Instructions
62,383 Million Matrices/Sec
58,486 Million Matrices/Sec​
42,935 Million Matrices/Sec​
48,928 Million Matrices/Sec​
47,539 Million Matrices/Sec​
Single Thread
4,312 MOps/Sec​
4,168 MOps/Sec​
4,939 MOps/Sec
4,764 MOps/Sec​
4,661 MOps/Sec​
Average CPU Mark​
63335
62653​
59384​
61860​
59696​
These scores look busted - the 14900KF had to have horrible cooling since it's losing to the 13900K in a lot of MT workloads.
 
These scores look busted - the 14900KF had to have horrible cooling since it's losing to the 13900K in a lot of MT workloads.
Yea, something is fishy. No way the 13900ks should be faster than 14900k. That said, I really hate to see Intel going back to this refresh protocol. Looks like they are going to do the same thing for ARL, but maybe they will be able to make more improvement there, since ARL is the first of a new generation.
 

Found these slides VERY interesting:

1697462226143.png
1697462250826.png
1697462272880.png

Guess Intel should never be under-estimated. They managed to introduce something new even in a boring launch.
 
It's available to order on Newegg. For reference, they are charging $17 more for the 14600K over the 13600K. And yet the improvement is very minimal based upon clocks.
 

Found these slides VERY interesting:

View attachment 87238
View attachment 87239
View attachment 87240

Guess Intel should never be under-estimated. They managed to introduce something new even in a boring launch.
Well, that should help the MT benchmarks for the 14900k/kf... will be interesting if they extend that tool down the stack or leave it exclusive to those SKUs.
 
Back
Top