Thursday, October 29th 2020

Intel Confirms Rocket Lake-S Features Cypress Cove with Double-Digit IPC Increase

Today, Intel has decided to surprise us and give an update to its upcoming CPU lineup for desktop. With the 11th generation, Core CPUs codenamed Rocket Lake-S, Intel is preparing to launch the new lineup in the first quarter of 2021. This means that we are just a few months away from this launch. When it comes to the architecture of these new processors, they are going to be based on a special Cypress Cove design. Gone are the days of Skylake-based designs that were present from the 6th to 10th generation processors. The Cypress Cove, as Intel calls it, is an Ice Lake adaptation. Contrary to the previous rumors, it is not an adaptation of Tiger Lake Willow Cove, but rather Ice Lake Sunny Cove.

The CPU instruction per cycle (IPC) is said to grow in double-digits, meaning that the desktop users are finally going to see an improvement that is not only frequency-based. While we do not know the numbers yet, we can expect them to be better than the current 10th gen parts. For the first time on the Intel platform for desktops, we will see the adoption of PCIe 4.0 chipset, which will allow for much faster SSD speeds and support the latest GPUs, specifically, there will be 20 PCIe 4.0 lanes coming from the CPU only. The CPU will be paired with 12th generation Xe graphics, like the one found in Tiger Lake CPUs. Other technologies such as Deep Learning Boost and VNNI, Quick Sync Video, and better overclocking tuning will be present as well. Interesting thing to note here is that the 10C/20T Core i9-10900K has a PL1 headroom of 125 W, and 250 W in PL2. However, the 8C/16T Rocket Lake-S CPU also features 125 W headroom in PL1, and 250 W in PL2. This indicates that the new Cypress Cove design runs hotter than the previous generation.
Source: Intel Newsroom
Add your own comment

74 Comments on Intel Confirms Rocket Lake-S Features Cypress Cove with Double-Digit IPC Increase

#51
Mouth of Sauron
I want "Xe gen12" - though I was living a lie that Xe is gen1 or gen2 max, but however IGP numbering goes, this will be my major selling point! Because it will rule or something!

Seriously, it's about time for Intel to lower themselves and stop producing one-size-fits-all type of (guess) APUs, because I don't need at all their worthless iGPU and never did, ok - companies are another thing, but individual users rarely do, and we all still have to pay for it, its development and probably flawed silicon products.

I don't want it and go make a variant without, which will be 10g less - no matter how integrated Xe became great... or didn't. Unless you make your 'One Ring, One Intel-HSA API initiative" worthy of Mordor and usable everywhere...
Posted on Reply
#52
londiste
Minus InfinityIt would need 50% IPC uplift, 30% power decrease, 30% price cut and even then I still wouldn't buy it.
Really? It could blow the competition absolutely out of water and you still would not buy these CPUs?
Posted on Reply
#53
R0H1T
You're not getting a 30% price cut from Intel, those days are long gone so even if they do a Conroe (sort of) the price will not be earth shattering!
Posted on Reply
#54
z1n0x
I can't wait for Ryan Shrout's stand-up comedy special.
Posted on Reply
#55
Melvis
dicktracyAMD's reign as "fastest gaming CPU" will only last for a few months lol.
and no one will care just like now, no one cares.
Posted on Reply
#58
bug
Owen1982STILL 14nm :wtf:
At this point I guess it's good news they didn't go back to 22nm :D
Posted on Reply
#59
InVasMani
Memory support for DDR4 3200MHz yet Intel's test configuration 4x16GB of DDR4 2933MHz. Alright so what is the reason for that it ain't like DDR4 3200MHz is difficult to obtain clearly something going on with that picture!? Is the DDR4 3200MHz support only for two DIMM sockets perhaps!? Is it to possibly alter TDP though idk how much altering memory impacts the overall TDP however imagine it is measurable in some form or another. Is it to bump up IPC claims by doing more per clock cycle at the expense of reduce bandwidth!? I'm not saying Intel is cheating, but I want to know the legitimate rational reason behind it. Boasting DDR4 3200MHz support then utilizing weaker memory in their own test configuration comes across as some type of red flag to me that there is a probably a gotcha on some of their claims. Perhaps they aren't being misleading, but it kind of appears like they could be at the same time.
Posted on Reply
#60
bug
InVasManiMemory support for DDR4 3200MHz yet Intel's test configuration 4x16GB of DDR4 2933MHz. Alright so what is the reason for that it ain't like DDR4 3200MHz is difficult to obtain clearly something going on with that picture!? Is the DDR4 3200MHz support only for two DIMM sockets perhaps!? Is it to possibly alter TDP though idk how much altering memory impacts the overall TDP however imagine it is measurable in some form or another. Is it to bump up IPC claims by doing more per clock cycle at the expense of reduce bandwidth!? I'm not saying Intel is cheating, but I want to know the legitimate rational reason behind it. Boasting DDR4 3200MHz support then utilizing weaker memory in their own test configuration comes across as some type of red flag to me that there is a probably a gotcha on some of their claims. Perhaps they aren't being misleading, but it kind of appears like they could be at the same time.
Intel doesn't have IF, super-fast DRAM is not required ;)
Posted on Reply
#61
efikkan
InVasManiMemory support for DDR4 3200MHz yet Intel's test configuration 4x16GB of DDR4 2933MHz. Alright so what is the reason for that it ain't like DDR4 3200MHz is difficult to obtain clearly something going on with that picture!? Is the DDR4 3200MHz support only for two DIMM sockets perhaps!? Is it to possibly alter TDP though idk how much altering memory impacts the overall TDP however imagine it is measurable in some form or another. Is it to bump up IPC claims by doing more per clock cycle at the expense of reduce bandwidth!? I'm not saying Intel is cheating, but I want to know the legitimate rational reason behind it.
This will be determined when the final stepping is ready. Comet Lake-S promised 2666 MHz, but were validated for 2933 MHz, so we'll have to see the final production version, but it will probably be 3200 MHz for all four DIMMs.
IPC doesn't go up by slowing down the memory.
Posted on Reply
#62
Blueberries
I never believed the conspiracy that Intel had better cores tucked away and continued to sell Skylake refreshes due to lack of competition but with the timing here it's hard to believe otherwise.
Posted on Reply
#63
TheoneandonlyMrK
BlueberriesI never believed the conspiracy that Intel had better cores tucked away and continued to sell Skylake refreshes due to lack of competition but with the timing here it's hard to believe otherwise.
Not sure I get you, so they stick with skylake due to a lack of competition, they would have continued evolving it through node's had that of panned out.
Seems like five years ago or more they started on skylakes replacement , it too hasn't gone to plan.
They definitely do need to start sticking to the plan though or they're not going to recover.
As for timing, they been an absolute Age on skylakes.
The skylake epoche, fitting.
Posted on Reply
#65
londiste
tiggerYikes 21% faster single core.
From the looks of it, 21% faster at 4.2GHz compared to 4.7GHz-ish 10700K. It is "only" 14% faster compared to 10700K that boosts to 5.1GHz. Basically ~38% when accounting for higher frequency of 10700K. That... does not sound very believable.
Posted on Reply
#66
efikkan
londisteFrom the looks of it, 21% faster at 4.2GHz compared to 4.7GHz-ish 10700K. It is "only" 14% faster compared to 10700K that boosts to 5.1GHz. Basically ~38% when accounting for higher frequency of 10700K. That... does not sound very believable.
Even if the scores are real, don't count on the reported clock speeds being accurate.
Also, the rated clock speeds of Comet Lake and Coffee Lake are highly optimistic. If you look at Gamer's Nexus' review of i9-10900K, it doesn't get anywhere near a stable boost at 5.3 GHz on a single core. This is similar to my experience with trying an i7-10700K at work (stock, power limit on), it's more like 4.6-4.7 GHz at sustained loads, don't think I've seen it hit 5.0 GHz, but I've not done a "proper" test either.

If Rocket Lake achieves similar real world clock speeds of 4.6-4.7 GHz (and a little lower under heavy AVX loads), and offers an ~18% IPC gain, it will pack a serious punch. Though I wonder if Intel have to keep inflating the rated clock speed…
Posted on Reply
#67
londiste
efikkanEven if the scores are real, don't count on the reported clock speeds being accurate.
My own experience so far has been that boost clock is what the benchmark actually measures and it should be accurate. Even more so, single thread scores of other 8-core CPUs like 10700K or 3800X clearly have a range that correlates very well with the clock speed determined by benchmark. If I had to guess, there is something about that Rocket Lake we do not know or simply benchmark cannot read. Earlier rumors did have 5.0GHz boost speed for the thing, so... :)
Posted on Reply
#68
sepheronx
im a little ticked that max is 8c/16t when I was hoping for at least 10c/20t for one of these.
Posted on Reply
#69
efikkan
londisteMy own experience so far has been that boost clock is what the benchmark actually measures and it should be accurate. Even more so, single thread scores of other 8-core CPUs like 10700K or 3800X clearly have a range that correlates very well with the clock speed determined by benchmark.
Sure, but you're talking about retail versions, not engineering samples, right?
There is no guarantee an engineering sample is reporting accurate clock speeds.
Posted on Reply
#70
Caring1
0.5% is still double digits right? :rolleyes::p
Posted on Reply
#71
OneMoar
There is Always Moar
01% is technically double digit
Posted on Reply
#72
The Von Matrices
sepheronxim a little ticked that max is 8c/16t when I was hoping for at least 10c/20t for one of these.
I noticed that too. This may cause a marketing problem. Uninformed consumers will buy the older 10-core CPU over a newer 8-core CPU just because of the bigger number of cores.
Posted on Reply
#74
bug
kane nas
Yes, I think everyone (at least on TPU) knows exact voltages and frequencies are set in stone last, so performance of engineering samples is just an indicator. A pretty good one, but just an indicator in the end.
Posted on Reply
Add your own comment
Apr 26th, 2024 03:46 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts