Question Raptor Lake - Official Thread

[Hulk]

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.

 

[Exist50]

My demands have plummetted but I'd love a Meteorlake 5W 2-in-1 weighing under 2.5lbs and great battery life for a replacement to this XPS.

I think Lunar Lake will much, much closer to what you're looking for. Though 5W will be a stretch regardless. As for myself, I'm hoping to wait it out till (presumably) Nova Lake + Royal, but I'm keeping an eye out in case Zen 5 or Arrow Lake turn out well.

 

[poke01]

I think Lunar Lake will much, much closer to what you're looking for. Though 5W will be a stretch regardless. As for myself, I'm hoping to wait it out till (presumably) Nova Lake + Royal, but I'm keeping an eye out in case Zen 5 or Arrow Lake turn out well.

I am waiting for either a M3 or M4 MacBook to go along with my Raptor Lake PC. Why M Series for laptops for me? Well no one in the industry does idle power management like Apple.

 

[Henry swagger]

I am waiting for either a M3 or M4 MacBook to go along with my Raptor Lake PC. Why M Series for laptops for me? Well no one in the industry does idle power management like Apple.

Because apple cores are low clocked.. apple m1 is not that impressive.. clock them to 3.5 ghz they overheat to 108c 🙂

 

[lobz]

Because apple cores are low clocked.. apple m1 is not that impressive.. clock them to 3.5 ghz they overheat to 108c 🙂

Try pitching that word for word in the apple silicon thread, you're gonna be in for an unpleasant time 🤣

 

[poke01]

Because apple cores are low clocked.. apple m1 is not that impressive.. clock them to 3.5 ghz they overheat to 108c 🙂

yeah on second thought I might wait for Arrow lake or lunar lake. /s

 

[poke01]

Because apple cores are low clocked.. apple m1 is not that impressive.. clock them to 3.5 ghz they overheat to 108c 🙂

M2 reaches 108c while rendering a 8k Canon raw footage and running cinebench. Max tech is just click bait.

Who does that on a fanless macbook??
Get the 14" macbook pro with dual fans for that kind of workload.

HU found the M2 to have best battery life while consuming less watts in ST and MT.

 

[nicalandia]

There should be no difference in performance between a 12900K and the 13700K besides the few apps that take advantage of the extra L2 Cache, which so far have not been games according to QS sample gaming leaks.

Intel Core i7-13700K 16-core CPU spotted with 5.36 GHz clock, up to 17% faster than i7-12700K - VideoCardz.com

Intel Core i7-13700K is a 16-core Raptor Lake CPU All three K-Series processors from the upcoming Raptor Lake series have now been found in Geekbench benchmark database. Over the past few days we have seen numerous leaks featuring Intel mid-range Core i5-13600K 14-core CPU. A week sooner we...

videocardz.com

 

[Mopetar]

Odd that games wouldn't take advantage of the extra cache given what we know about the 5800X3D. Was there a dramatic latency increase or something like that which is causing it to offset the size or is the way the cache was structured that would be less optimal than alternatives?

 

[Thunder 57]

Odd that games wouldn't take advantage of the extra cache given what we know about the 5800X3D. Was there a dramatic latency increase or something like that which is causing it to offset the size or is the way the cache was structured that would be less optimal than alternatives?

We've seen this before, notably with Prescott. From the review: "If our cache latency figures are correct, it will take a 4GHz Prescott to have a faster L2 cache than a 2.8GHz Northwood. It will take a 5GHz Prescott to match the latency of a 3.4GHz Northwood. "

Also saw it again when they increased it to 2MB. This review actually has hard numbers. Hopefully we aren't seeing anything that bad, but it seems likely the access time went up a couple cycles or so.

 

[nicalandia]

Odd that games wouldn't take advantage of the extra cache given what we know about the 5800X3D. Was there a dramatic latency increase or something like that which is causing it to offset the size or is the way the cache was structured that would be less optimal than alternatives?

That's What happens when your L2 size goes Up by 60% but your associativity remains the same(10Way for Client , but 16-Way for Servers)

 

[FangBLade]

That's What happens when your L2 size goes Up by 60% but your associativity remains the same(10Way for Client , but 16-Way for Servers)

Yeah, the reason why Zen 3 benefits so much of extra cache is because latency increase is minimal, excellent implementation, and Zen 4 l2 cache was already tested and it is also excellent, minimal latency increase while having 2x size, can't wait to see how Zen 4 scales in gaming with 3dcache + double l2.

 

[Exist50]

Seems kind of odd to assume the latency increased significantly, especially given that games did show frametime improvements. Perhaps something to do with the test setup (GPU bound)? That, or the difference between 1.25 and 2MB doesn't cover any new working sets.

 

[Thunder 57]

Seems kind of odd to assume the latency increased significantly, especially given that games did show frametime improvements. Perhaps something to do with the test setup (GPU bound)? That, or the difference between 1.25 and 2MB doesn't cover any new working sets.

Have they? I don't think I've seen any actual games benchmarked. I wouldn't be surprised if it went up 1-2 cycles. I'd like to see things other than Geekbench as well. Guess I'll see what else might be out there.

 

[Exist50]

Have they? I don't think I've seen any actual games benchmarked. I wouldn't be surprised if it went up 1-2 cycles. I'd like to see things other than Geekbench as well. Guess I'll see what else might be out there.

From the exact same source as nicalandia is using to claim no improvement in games, there was a substantial improvement in minimum frametimes.

Intel Core i9-13900K engineering sample tested in games shows improvement in minimum framerates - VideoCardz.com

Intel Raptor Lake faster in gaming ExtremePlayer shares more benchmark data on Intel’s upcoming Raptor Lake processor. Hardware reviewer recently acquired a sample of Intel 13th Gen Core CPU. This processor featuring 24-cores and 32-threads with unlocked power limits, which basically means it’s...

videocardz.com

Now what to make of that? I really don't know. But I'm not convinced those results show a significant increase in latency. You'd think that would produce a greater variety of scores, no?

 

[Thunder 57]

From the exact same source as nicalandia is using to claim no improvement in games, there was a substantial improvement in minimum frametimes.

Intel Core i9-13900K engineering sample tested in games shows improvement in minimum framerates - VideoCardz.com

Intel Raptor Lake faster in gaming ExtremePlayer shares more benchmark data on Intel’s upcoming Raptor Lake processor. Hardware reviewer recently acquired a sample of Intel 13th Gen Core CPU. This processor featuring 24-cores and 32-threads with unlocked power limits, which basically means it’s...

videocardz.com

Now what to make of that? I really don't know. But I'm not convinced those results show a significant increase in latency. You'd think that would produce a greater variety of scores, no?

Ah yes I remember seeing that now, thanks. The minimums were impressive. I still think tradeoffs had to be made somewhere, otherwise why not go with the larger L2 to begin with? They're both Intel 7. Maybe the compromise was power, as those leaks indicated an increase there. I still think it's possible there was a latency increase that was possibly/probably negated by the increase in frequency. That would give you a similar latency (in ns) but you get the increase in size to help as well.

 

[Exist50]

I still think tradeoffs had to be made somewhere, otherwise why not go with the larger L2 to begin with?

I think cost was probably the dominant reason, with some minor consideration for power. Will be interesting to see if it shows up in mobile, or if they stick with the current L2 capacity.

 

[LightningZ71]

I was always unde4 the impression that minimum frame rates were largely controlled by "last mile" memory efficiency. In this case, higher speed dram and reduced latencies in the memory controller might be the overriding factor here. The marginal L3 increase might also be showing up here too.

 

[pakotlar]

13700K: 19,811 Geekbench 5 MT with DDR5-5200. Puts it well above 12900K (17,287) and 5950X (16,506).

DDR5 memory makes Intel Raptor Lake CPUs at least 11% faster in Geekbench - VideoCardz.com

Intel Raptor Lake gets DDR5 boost Two Intel Raptor Lake CPUs have now been tested with DDR5 and DDR4 memory. A fresh Core i7-13700K 16-core CPU score is now available on Geekbench website, this time, however with DDR5 memory. The processor has now been tested with two nearly identical...

videocardz.com

Totally speculating here, but I wonder if the larger cache makes Raptor Lake less sensitive to DDR5’s higher latencies or otherwise better able to make use of DDR5 bandwidth.

edit: having trouble finding 12900K + DDR5 5200 bench. I found one with ddr5 4800 (anandtech, but on windows 10) and that gave 18,500 or so, so the 13700K here is likely identical to 12900K with ddr5 5200, in Geekbench 5.

 

[tamz_msc]

12900K with DDR5-5200. Honestly, I don't expect much of a difference between the 13700K and 12900K. 13700K will be slightly better overall.

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an Intel Core i9-12900K processor.

browser.geekbench.com

 

[IntelUser2000]

I think Lunar Lake will much, much closer to what you're looking for. Though 5W will be a stretch regardless. As for myself, I'm hoping to wait it out till (presumably) Nova Lake + Royal, but I'm keeping an eye out in case Zen 5 or Arrow Lake turn out well.

If it's using the Intel process sure. I like the company, but only if it's good. But since I don't NEED it, I also like getting inflection point upgrades. Like Haswell. I am willing to put in the above and beyond tweaks needed to allow with to reach ultra low idle. Of course if MTL is a sidegrade, sure I will likely skip it. If the new SoC configuration allows a big boost, yea it's possible that's my next.

 

[Exist50]

If it's using the Intel process sure. I like the company, but only if it's good. But since I don't NEED it, I also like getting inflection point upgrades. Like Haswell. I am willing to put in the above and beyond tweaks needed to allow with to reach ultra low idle. Of course if MTL is a sidegrade, sure I will likely skip it. If the new SoC configuration allows a big boost, yea it's possible that's my next.

Maybe should move this discussion to the "future" thread, but I've become a lot more optimistic about Lunar Lake recently. I think it might well deserve to be called Haswell's spiritual successor.

That said, the whole "If it's using the Intel process" thing might be a problem... Why is that a requirement for you? I'm honestly surprised you care either way.

 

[HurleyBird]

and Zen 4 l2 cache was already tested and it is also excellent, minimal latency increase while having 2x size, can't wait to see how Zen 4 scales in gaming with 3dcache + double l2.

Zen 4 L2 was tested? Where?

 

[Det0x]

Zen 4 L2 was tested? Where?

I guess he is referring to this:

L1 read = 29924.5 / 6 = 4987 GB/s (+ ~200mb/sec)
L1 write = 17524.5 / 6 = 2920 GB/s
L1 copy = 30061.4 / 6 = 5010GB/s

L2 read = 19985.6 / 6 = 3330 GB/s (+ ~1000mb/sec)
L2 write = 19100.7 / 6 = 3183 GB/s
L2 copy = 18840.3 / 6 = 3140 GB/s

L3 read = 10257.7 / 6 = 1709 GB/s (+ ~200mb/sec)
L3 write = 9693.1 / 6 = 1615 GB/s
L3 copy = 9107.5 / 6 = 1517 GB/s

vs

https://twitter.com/x/status/1549256374936170497

 

[poke01]

The current idle wattage of a Alder lake P chip is 4 watts avg. Hopefully MTL improves this. maybe Raptor will improve this but I doubt.
Tested on Galaxy Book 2. high is 7 watts and low is 4 watts. Intel should also target efficiency for ultrabooks like Apple and AMD do.

https://www.notebookcheck.net/Intel-Core-i7-1260P-Processor-Benchmarks-and-Specs.589539.0.html

 

[coercitiv]

The current idle wattage of a Alder lake P chip is 4 watts avg. Hopefully MTL improves this. maybe Raptor will improve this but I doubt.
Tested on Galaxy Book 2. high is 7 watts and low is 4 watts.

That's the power consumption of the entire unit in idle, not just the Alder Lake P package.