Raptor Lake 2022 October at the latest

chameleoneel

Supreme [H]ardness
Joined
Aug 15, 2005
Messages
5,942
These scores are reportedly from a 12600k engineering sample. 6 P-cores + 8 E-cores. beats 12700k in multithread by ~ 1400 points---which is not a trivial win. And will probably beat it in gaming. Pretty incredible.

1658162457673.png
 

LukeTbk

2[H]4U
Joined
Sep 10, 2020
Messages
2,342
Not sure how legitimate this is, on the multicore side you have 10980xe 18c-36T performance above an 12700K while on the single thread it is below a 11400F ?

If it achieve above 12700K on multithread despite a massively lower single thread performance (1387 is juste 71% of 1,939) and despite less P-Core, that could be a really good sign on how well it handle/use the E-core or handle heat or something of the sort, because that seem a bit strange, could be an benchmark issue, a engineering chips that do not clock well or boost past a very low level that affect the ST performance a lot but not the MT or me not understanding/reading the numbers correctly.
 

chameleoneel

Supreme [H]ardness
Joined
Aug 15, 2005
Messages
5,942
Not sure how legitimate this is, on the multicore side you have 10980xe 18c-36T performance above an 12700K while on the single thread it is below a 11400F ?

If it achieve above 12700K on multithread despite a massively lower single thread performance (1387 is juste 71% of 1,939) and despite less P-Core, that could be a really good sign on how well it handle/use the E-core or handle heat or something of the sort, because that seem a bit strange, could be an benchmark issue, a engineering chips that do not clock well or boost past a very low level that affect the ST performance a lot but not the MT or me not understanding/reading the numbers correctly.
Its Windows 10, which does not have the thread director for the the Big/Little architecture. Its very possible that during their test run, it incorrectly ran the single thread test on an E-core, instead of a P-core.
 

chameleoneel

Supreme [H]ardness
Joined
Aug 15, 2005
Messages
5,942
Techpowerup tested the e-cores from a 12900k at 1089 points for Cinibench R23 single core. 1387 for Raptor Lake, would be a nice improvement.
 

LukeTbk

2[H]4U
Joined
Sep 10, 2020
Messages
2,342
Its Windows 10, which does not have the thread director for the the Big/Little architecture. Its very possible that during their test run, it incorrectly ran the single thread test on an E-core, instead of a P-core.
If that the case:
https://www.anandtech.com/bench/CPU-2020/2975

1,347 would be really good, 12900K e-core only scoring 1068 according to the above and could explain why it achieve to beat a 12700k despite having a more e-core heavy setup.

That would mean e-core with 9900K core performance.
 

AgentQ

[H]ard|Gawd
Joined
Jun 16, 2000
Messages
1,589
Scores for all 3 -K variants showed up on Geekbench. The 13700K (8P + 8E) is 5% faster single core, but 4% slower than the 12900K (8P + 8E). The tests were done using DDR4-3600 memory, though. Interested in seeing scores using DDR5.
Supposed 13700K + DDR5 benchmarks spotted on Geekbench: https://browser.geekbench.com/v5/cpu/16190485

Single-core is within margin of error of the DDR4 results, but multi-core gets a huge boost. 16,542 on DDR4 to 19,811 on DDR5.

Now I'm very interested in seeing 13900K + DDR5 results.
 

Comixbooks

Fully [H]
Joined
Jun 7, 2008
Messages
18,677
Intel Raptor Lake to launch on October 17th
According to Enthusiastic Citizen (ECSM_Official) better known for reviews featuring unreleased hardware, Intel has now set a launch date for its new desktop CPU series.
It is said that Intel Raptor Lake processors will launch on October 17th, which is almost 3 weeks after the official introduction during the Innovation event. On this date, Intel will launch its K-series CPUs from 13th Gen Core series “Raptor Lake-S” along with Z790 high-end motherboards.
 

Grebuloner

[H]ard|Gawd
Joined
Jul 31, 2009
Messages
1,256
Is anybody planning to install a 240v circuit for this?
Hah! I've considered it after seeing too many rumors of Nvidia's lax attitude towards the 40 series power consumption, but I'm having a current issue that the largest standard UPSes only go to 900W. Time to give up on power loss protection?
 

Randall Stephens

[H]ard|Gawd
Joined
Mar 3, 2017
Messages
1,282
Hah! I've considered it after seeing too many rumors of Nvidia's lax attitude towards the 40 series power consumption, but I'm having a current issue that the largest standard UPSes only go to 900W. Time to give up on power loss protection?
Naw, time to “innovate” and bring back dual psu cases. I remember the days of the 5.25 bay power supplies…
 

Burke888

Gawd
Joined
Aug 1, 2004
Messages
550
I've got a 9900k right now, clocked at 5.0Ghz on all cores. I game using a 4k 144Hz panel with an RTX 3090. Would Raptor Lake or the new Ryzen 7000 series offer any significant improvement? I've got the upgrade bug really bad, so much so that I'm debating picking up a 12900k now.
 

Darkswordz

Gawd
Joined
Jan 5, 2016
Messages
730
I've got a 9900k right now, clocked at 5.0Ghz on all cores. I game using a 4k 144Hz panel with an RTX 3090. Would Raptor Lake or the new Ryzen 7000 series offer any significant improvement? I've got the upgrade bug really bad, so much so that I'm debating picking up a 12900k now.
No point spending $575.00+ on a 12900K when it is about to be made obsolete by the 13700K, which should offer equal performance for less money.

Or, if you're in the $575.00+ market, why not just wait for the 13900K?

scores.png
 

Grebuloner

[H]ard|Gawd
Joined
Jul 31, 2009
Messages
1,256
But the difference between the 12900k and the 13700k is so small does it even make a difference?
Two differences, in fact:
1. Better memory support (if using DDR5)
2. Couple hundred bucks less or, thinking of it differently, couple hundred bucks to go towards another part.
 

xDiVolatilX

Limp Gawd
Joined
Jul 24, 2021
Messages
479
Two differences, in fact:
1. Better memory support (if using DDR5)
2. Couple hundred bucks less or, thinking of it differently, couple hundred bucks to go towards another part.
The memory support is great.

But what are the prices? Is the 13700k going to be the same price as the 12900k?
 

Burke888

Gawd
Joined
Aug 1, 2004
Messages
550
https://www.guru3d.com/news-story/i...-and-without-power-management-settings,3.html

It looks like based on this R23 Cinebench score that the 13900k scores about the same as a 12900ks for singled threaded performance at 2290? :( With this in mind, would it be better to go with the new AMD series? Or is the IPC gain from AMD's new CPUs speculated to still not even match the 12900k? I'm guessing here, but I can see perhaps Intel having a slight IPC edge late October 2022. But then sometime next year, AMD perhaps taking the crown and still using the same motherboard they released in September '22. Whereas any future Intel releases would require a new motherboard. Curious as to others opinions.
 

Grebuloner

[H]ard|Gawd
Joined
Jul 31, 2009
Messages
1,256
https://www.guru3d.com/news-story/i...-and-without-power-management-settings,3.html

It looks like based on this R23 Cinebench score that the 13900k scores about the same as a 12900ks for singled threaded performance at 2290? :( With this in mind, would it be better to go with the new AMD series? Or is the IPC gain from AMD's new CPUs speculated to still not even match the 12900k? I'm guessing here, but I can see perhaps Intel having a slight IPC edge late October 2022. But then sometime next year, AMD perhaps taking the crown and still using the same motherboard they released in September '22. Whereas any future Intel releases would require a new motherboard. Curious as to others opinions.
Alder and Raptor are essentially using the same architecture with larger caches on Raptor, so single core scores at the same clocks should be similar, and multicore will be better (also, remember: prerelease hardware is being unofficially tested, use a big salt shaker!). If AMD delivers on their slides, they'll at least match IPC with Alder P-cores, and theoretically the 16-core 7950X could pull ahead of the 8/16 13900k depending on the workload.

You might as well keep saving up your money and wait for everything to be officially released and compared since it's not too far away, now.
 

Domingo

Fully [H]
Joined
Jul 30, 2004
Messages
20,961
The new AMD stuff is just around the corner. Seems like a prime moment to see what both sides bring to the table before making a decision. Pretty sure I'll be making a new scratch build this winter and my plan is to go with whoever has the best product for my use case.
 

chameleoneel

Supreme [H]ardness
Joined
Aug 15, 2005
Messages
5,942
Two differences, in fact:
1. Better memory support (if using DDR5)
2. Couple hundred bucks less or, thinking of it differently, couple hundred bucks to go towards another part.
more cache, as well. So far, seems to equal at least 10% better minimum frames. And in some games that are more cache sensitive, its 20% better.

And that's with QS and ES samples. Final silicon and firmwares could improve it. Intel usually improves latency in the first couple of bios updates, etc.
 
Last edited:

saedrin

2[H]4U
Joined
Jan 5, 2009
Messages
2,112
Looking to pick up a higher end SKU when they drop as a new container host. Looks like ECC will still be supported on the W series chipset.
 
Top