Intel Core i9-11900(k) + i7-11700(k/kf) 8c16t Rocket Lake Desktop CPU Benchmarks and Pricing Leaked: (Update #8)

[Bombastinator]

4 hours ago, Tedny said:

if pricing below 5000 series, it maybe will be good, if it not... it's anyway time to next gen upgrade soon, so why buy something that will be old way to soon 

There are people with LGA boards out there who could have use for it.  They’ve got to compete on price as well as function though. That became clear with ryzen2.

 

4 hours ago, RejZoR said:

This is good. It means AMD will be working hard on further IPC improvements and higher clocks. We've been waiting for such market situation for years You know, AMD and Intel poking each other for our money.

I concur.  AMD does have a ftm advantage, but they’re going to have to run like bunnies to keep it.

[Bombastinator]

4 hours ago, jagdtigger said:

They had to gimp it so it  does not need a chiller......

This may actually be possible.  We’ll know when it comes out.

[porina]

 

27 minutes ago, Bombastinator said:

There are people with LGA boards out there who could have use for it.  They’ve got to compete on price as well as function though. That became clear with ryzen2.

It is always more complicated than a single factor like price. Once they become available, if AMD haven't allocated more production to Zen 3 by then, Intel can simply win by having a product that people can buy once it is launched.

[yian88]

Who even cares about 2%, 7% these things are not affordable anyway, ryzen is where the hype is at.

[svmlegacy]

18 hours ago, williamcll said:

11700H maybe?

This is an engineering sample. It has no name, and will probably not be tied to any release CPU.

[spartaman64]

19 hours ago, BiG StroOnZ said:

 

Some food for thought from the source link in the OP:

 

 

that seems sort of misleading advertising a low clock that its never intended to run at just to advertise a low tdp number

[Bombastinator]

52 minutes ago, spartaman64 said:

that seems sort of misleading advertising a low clock that its never intended to run at just to advertise a low tdp number

It will need to be reviewed when it comes out.  Manufacturers claims have always been massaged by marketing at least a little.  It’s an argument for not pre-ordering one of these but there’s a pretty good argument for never pre-ordering ANYTHING except in very special circumstances.

[tim0901]

On 12/21/2020 at 12:11 PM, porina said:

...if there are no further delays. My gut feeling is Rocket Lake is likely to meet the currently communicated Q1, but I have less confidence that Alder Lake will be on time. Look at Ice Lake server, unless that remains on the pre-SF 10nm where Alder Lake is SF or ESF (without looking it up).

Ice lake server (Ice Lake-SP) is on what was originally called 10nm+, but has since been rebranded as 10nm. Tiger Lake is on 10nm Superfin, whilst Alder lake is on 10nm enhanced superfin. Yes my head hurts.

 

The bigger change with Alder lake though is heterogeneous compute - aka big.LITTLE. The advantages of this on a desktop CPU are... complex, especially given it will likely have the same problems as AMD had with it's CCX chiplets in that software (the Windows scheduler) won't be able to utilise it properly at first.

 

Rant about how a big.LITTLE design could make a lot of sense for desktop cpus:

Spoiler

From slides Intel has shown off, vector compute (AVX) seems to be making it's way to their little cores (Gracemont). This is great news and should allow them to be more useful in highly multithreaded workloads.

 

Also, smaller cores are more efficient at a given power load. We know that zen cores (for example) can be power throttled at times - this is clear when you measure power consumption vs cores loaded - a 5950x reaches peak overall power consumption at an 8-10-core load, despite being a 16-core processor:

 

 

it's clear here that the individual cores have under half the power available to them when undergoing a full-core load vs an 8 core load. There is potential performance left on the table here, but it can't be extracted because of power budgets and heat - this is one of the reasons why multicore scaling always kinda sucks even in perfectly multithreaded workloads.

 

And so this brings up an interesting question: is there much point having 16 high-powered cores if you can't make full use of all of them at once anyway? Given that the whole point of 'little' cores is to be more power-efficient than their 'big' partners, could it be that - by making better use of that limited power budget - they can offer significant benefits during a multithreaded workload?

 

The answer to that will depend on the answer to a different question: how close can a 'little' core get to a power-limited 'big' core? According to Intel, the answer to that is "pretty close". Their current Lakefield hybrid processors apparently show a performance profile that looks a bit like this:

 

 

which predicts that at 50% power budget, a Sunny Cove 'big' core consumes the same amount of power as a Tremont 'little' core running flat-out, while also providing ~15% more performance. Which doesn't sound too great for the little Tremont cores... until you realise that they are a third of the size of the big Sunny Cove core:

 

(there are four cores in that Tremont Atom section)

 

If you're chasing multicore performance - which you are when you're talking 8+ core processors - replacing a couple of big cores with six little cores in the same space seems to be a bit of a no-brainer. At 33% of their power budget, three Tremont cores should handily outperform the single 50% power budget Sunny cove core that would consume the same amount of power (3x ~50% relative perf > 1x ~80%) - provided of course that the workload is sufficiently multithreaded. But if it isn't multithreaded, then it shouldn't matter - you've still got big cores to take care of those tasks for you. This lines up with the leaked configurations, which suggest that the low core count desktop CPUs would only have big cores - ie. the little cores are for those who need the multicore performance.

 

But who knows how this will actually play out, only time can truly tell.

 

Quote

While I'm not aware of any official communication, it seems the logical time to move to DDR5, so that'll be interesting.

Agreed. Nothing confirmed yet, but there have been a few leaks that have claimed it's coming.

[Bombastinator]

9 minutes ago, tim0901 said:

Ice lake server (Ice Lake-SP) is on what was originally called 10nm+, but has since been rebranded as 10nm. Tiger Lake is on 10nm Superfin, whilst Alder lake is on 10nm enhanced superfin. Yes my head hurts.

 

The bigger change with Alder lake though is heterogeneous compute - aka big.LITTLE. The advantages of this on a desktop CPU are... complex, especially given it will likely have the same problems as AMD had with it's CCX chiplets in that software (the Windows scheduler) won't be able to utilise it properly at first.

 

Rant about how a big.Little design could make a lot of sense for desktop cpus:

  Reveal hidden contents

From slides Intel has shown off, vector compute (AVX) seems to be making it's way to their little cores (Gracemont). This is great news and should allow them to be more useful in highly multithreaded workloads.

 

Also, smaller cores are more efficient at a given power load. We know that zen cores can be power throttled at times - this is clear when you measure power consumption vs cores loaded - a 5950x reaches peak overall power consumption at an 8-10-core load, despite being a 16-core processor:

 

 

it's clear here that the individual cores have under half the power available to them when undergoing a full-core load vs an 8 core load. There is potential performance left on the table here, but it can't be extracted because of power budgets and heat - this is one of the reasons why multicore scaling always kinda sucks even in perfectly multithreaded workloads.

 

And so this brings up an interesting question: is there much point having 16 high-powered cores if you can't make full use them all at once anyway? Given that the whole point of 'little' cores is to be more power-efficient than their 'big' partners, could it be that, by making better use of that limited power budget, they may offer significant benefits during a multithreaded workload?

 

The answer to that will depend on the answer to a different question: how close can a 'little' core get to a power-limited 'big' core? According to Intel, the answer to that is "pretty close". Their current Lakefield hybrid processors apparently show a performance profile that looks a bit like this:

 

 

which predicts that at 50% power budget, a Sunny Cove 'big' core consumes the same amount of power as a Tremont 'little' core running flat-out, while also providing ~15% more performance. Which doesn't sound too great for the little Tremont cores... until you realise that they are a third of the size of the big Sunny Cove core:

 

(there are four cores in that Tremont Atom section)

 

If you're chasing multicore performance - which you are when you're talking 8+ core processors - replacing a couple of big cores with six little cores in the same space seems to be a bit of a no-brainer. At 33% of their power budget, three Tremont cores should handily outperform the single 50% power budget Sunny cove core that would consume the same amount of power (3x ~50% relative perf > 1x ~80%) - provided of course that the workload is sufficiently multithreaded. But if it isn't multithreaded, then it shouldn't matter - you've still got big cores to take care of those tasks for you. This lines up with the leaked configurations, which suggest that the low core count desktop CPUs would only have big cores - ie. the little cores are for those who need the multicore performance.

 

But who knows how this will actually play out, only time can truly tell.

 

Agreed. Nothing confirmed yet, but there have been a few leaks that have claimed it's coming.

For what it’s worth I thought the rant on big.LITTLE was extremely cool.

[porina]

1 hour ago, tim0901 said:

Ice lake server (Ice Lake-SP) is on what was originally called 10nm+, but has since been rebranded as 10nm. Tiger Lake is on 10nm Superfin, whilst Alder lake is on 10nm enhanced superfin. Yes my head hurts.

Intel: Cannon Lake? Nah, it never existed...

 

1 hour ago, tim0901 said:

The bigger change with Alder lake though is heterogeneous compute - aka big.LITTLE. The advantages of this on a desktop CPU are... complex, especially given it will likely have the same problems as AMD had with it's CCX chiplets in that software (the Windows scheduler) won't be able to utilise it properly at first.

 

Rant about how a big.LITTLE design could make a lot of sense for desktop cpus:

While it made more obvious sense on mobile applications, this is the first description I've seen making a case for it benefitting desktop. I was struggling to find a scenario where it gave a benefit, and this looks like it. Like so many enthusiasts, I'm more used to thinking in a power-unrestricted way. Generally what you can run one core at, you can run many cores at. Never mind power usage. For most other real world scenarios, power will be limited. Similarly we tend to think about max performance, not perf/watt, and it will be interesting to see how that goes.

[BiG StroOnZ]

Interesting little update (Ashes of the Singularity / AoTS i7-11700KF Benchmark Leak). I'm going to add this to OP, but still remain quizzical as mentioned in the tweet: 

Quote

 

Source 1

Source 2

Source 3

[Poison4K]

Dude what is that fucking power draw lmao

Also, no 10-core?

Guess Rocketlake is DOA unless it undercuts AMD, might as well just wait until Alderlake

[Kisai]

On 12/21/2020 at 5:19 AM, jagdtigger said:

They had to gimp it so it  does not need a chiller......

That's not what a non-K is. A non-K does not have unlocked clocks. K's do. T's are the "low power" desktop model that is intended for ITX/NUC types.

 

Just based on what is seen here, the K parts likely will need liquid cooling just like the Ryzen 9 x950 parts. That much TDP is just going to result in a lot of thermal throttle without a good cooler.

 

The i7 parts do not require liquid cooling.

 

 

[D13H4RD]

17 minutes ago, Kisai said:

Just based on what is seen here, the K parts likely will need liquid cooling just like the Ryzen 9 x950 parts. That much TDP is just going to result in a lot of thermal throttle without a good cooler.

They probably don't need liquid cooling, at least out of the box, provided MCE isn't enabled by default, which tends to run the CPU out of Intel's recommended spec.

 

Of course, if you have a K-series Intel SKU, it's likely that you may plan on overclocking, in which by then, a liquid cooler might be something to look into, although a good tower cooler might still do the job fine.

[jagdtigger]

2 hours ago, Kisai said:

Just based on what is seen here, the K parts likely will need liquid cooling just like the Ryzen 9 x950 parts.

Not fair comparison, this sucker is only 8 cores where as the 9 series is 12 at minimum.... Its more comparable to the r7 series which works just fine with an air cooler.

[porina]

A reminder to all about differences between how Intel and AMD treat TDP and power consumption in general.

 

Intel TDP = suggested (not enforced) long term power limit. You are guaranteed to get base clock as long as you use a cooler capable of meeting TDP cooling. Short term power limit may go higher for a certain amount of time as conditions allow. Enthusiast overclocking mobos tend to set unlimited power limit so the CPU can boost as far as it can. This is all allowable without voiding warranty and is NOT considered overclocking. Thermal limiting is essentially off below some high threshold. MCE is considered overclocking.

 

AMD TDP presumably works in a similar way with regard to cooler and base clock. At stock, AMD CPUs have a power limit "PPT" that is higher than TDP e.g. for 65W TDP parts it'll actually limit to 88W PPT. The stock coolers included with those parts are insufficient if you want maximum possible boost due to thermal limiting, as the CPUs more gradually reduce clocks as temps increase unlike the Intel approach. AMD, unlike Intel, do treat changing the power limit as an overclock function.

 

An Intel CPU can often run cooler than a Ryzen CPU at same power due to the process differences, assuming equal cooling. The power density due to the smaller node works against heat transfer and it is more concentrated at the core. Without overclocking a decent high end air cooler as more than sufficient for mainstream (not HEDT) CPUs from both sides.

[bcredeur97]

We’re gonna have staggering gains for a couple more gens then it’ll stagnate again IMO

 

im trying to figure out which gen will be the “sandy bridge” Generation so I can replace my Intel 3960X XD 

[BiG StroOnZ]

A few more updates to this story (all will be added to the OP) ~

 

Intel Core i9-11900 Rocket Lake-S Sample tested on B560 motherboard

 

Quote

 

Asian-based leakers have no problems finding the engineering samples of the new CPUs. It appears that the vast majority of the samples that were sent out are Core i9-11900 non-K CPUs (labeled QV1J). Engineering samples are clocked lower than retail units or have some boost frequencies disabled. It is also worth noting that, the sample tested by a leaker from Bilibili appears to have Xe Graphics present, although they cannot be tested due to lack of a driver.

 

The CPU has been tested in Cinebench R15 and R20:

 

 

The processor appears to match Core i7-10700 or Core i9-9900K, depending on the workload. The final version of the CPU should perform better. The leaker also tested the CPU power consumption using AIDA’s AVX512 and AVX2 tests. The package consumption varies between 123 to 161 W depending on which instruction was set for the test. It was also determined through the motherboard’s BIOS that the CPU’s PL1 and PL2 values are set to 65W and 224W respectively.

 

 

 

Source 1: https://videocardz.com/newz/intel-core-i9-11900-rocket-lake-s-engineering-sample-tested-on-b560-motherboard

Source 2: https://www.guru3d.com/news-story/rocket-lake-s-processors-likely-have-similar-pl2-states-as-10th-gen-intel-core-i9-11900-rocket-lake-s-leak.html

 

Intel Core i9-11900K, Core i9-11900, and Core i7-11700 engineering samples emerge

 

Quote

 

The leaker from Chiphell claims that he just received the samples of three Rocket Lake CPUs. including the 11900K. The engineering samples are clearly clocked lower than retails units that we expect to appear in the leaks soon. The Core i9-11900K ES appears to have a base clock of 3.4 GHz and a turbo of 4.8 GHz. The rumored specs mentioned a 3.5 GHz base clock and thermal velocity boost up to 5.3 GHz, this technology however is not shown by the CPU-Z software. The unit is clearly a 125W TDP model, indicating this is a K series CPU. The Core i9-11900 and i7-11700 engineering samples have also been acquired by the leaker. Both CPUs carry a base clock of 1.8 GHz with the turbo at 4.4 GHz for the 11900 CPU and 4.4 GHz for the 11700. Both samples are 65W processors. According to the leaker, the samples were acquired for as much as 2800 Yuans (430 USD) for the 11900K, 2300 Yuans (350 USD) for the 11900 non-K sample, and 1600 Yuans (245 USD) for the Core i7 model. 

 

The leaker also posted a Cinebench R20 score of the i7-11700 sample. The CPU allegedly scored 529 points in the single-core benchmark with 4672 points in the multi-core test:

 

 

Source 3: https://videocardz.com/newz/intel-rocket-lake-core-i9-11900k-i9-11900-i7-11700-engineering-sample-cpu-z-screenshots-leak

[Bombastinator]

2 hours ago, bcredeur97 said:

We’re gonna have staggering gains for a couple more gens then it’ll stagnate again IMO

 

im trying to figure out which gen will be the “sandy bridge” Generation so I can replace my Intel 3960X XD 

Not sure of that.  Gallium arsenide still hasn’t even shown itself, and it’s been lurking for something like 30 years.  People were talking about finally going there, then SOI got invented. It’s still waiting.

[porina]

4 hours ago, Bombastinator said:

Gallium arsenide still hasn’t even shown itself

Doubt it'll ever go mainstream and take over silicon. It has its niche use cases but for it to be economic over silicon... just can't see it happening unless something really radical happens.

[BiG StroOnZ]

Got another noice Rocket Lake update (will add to OP) ~

 

Intel Rocket Lake Core i9-11900K outperforms Ryzen 9 5950X in leaked Ashes benchmark. Intel could steal back its gaming performance crown from AMD, according to the leaked benchmark.

 

Quote

Intel's next flagship desktop processor, the Core i9-11900K, may only have eight cores, but it is already comparing well against the AMD Ryzen 9 5950X in gaming benchmarks. The Core i9-11900K holds upwards of an 8% advantage over AMD's 16-core Vermeer chip, even with essentially the same boost clock as its Comet Lake-S predecessor. The processor has been benchmarked with a GeForce RTX 2080 Ti, allowing further comparisons to be drawn between it and other recent desktop processors. According to the new entries, the Core i9-11900K scores around 6,075 points with a GeForce RTX 2080 Ti and 32 GB of RAM. The processor also achieved a framerate of 61.9 FPS on the Crazy_1080p preset running the DirectX 12 preset. By contrast, a similar Core i9-10900K system scored 5,700 points, even though it has two more cores than the Core-11900K. The Core i9-11900K still managed to outscore the Ryzen 9 5950X by a healthy margin. As the screenshot below shows, the Ryzen 9 5950X loses out by nearly 500 points and 5 FPS in the same preset.

 

 

 

 

The images above show the Ryzen 9 5950X getting 57 fps on the 1080p Crazy preset using an RTX 2080 Ti, against the Rocket Lake-S Core i9-11900K scoring 63 fps using the same GPU and with the same presets.  The only notable difference between the two systems is that the Ryzen spec is equipped with 16GB of memory while the 11900K has 32GB. This is unlikely to have skewed the results, however. The results would indicate that the Intel Core i9-11900K is running around 10.5% faster than its opponent, but it's worth bearing in mind that if this is indeed a genuine 11900K, it’s still an early engineering test sample and isn't indicative of the full performance of the final chip!

 

Source 1: https://www.notebookcheck.net/The-Intel-Core-i9-11900K-beats-the-Ryzen-9-5950X-in-new-Ashes-of-the-Singularity-benchmarks.509533.0.html

Source 2: https://www.techradar.com/news/intel-could-steal-back-its-gaming-performance-crown-from-amd-according-to-leaked-benchmark

[jagdtigger]

43 minutes ago, BiG StroOnZ said:

Intel could steal back its gaming performance crown from AMD, according to the leaked benchmark.

Like anyone with a brain would pay the huge intel tax just to get 5 extra fps.....

[BiG StroOnZ]

6 minutes ago, jagdtigger said:

Like anyone with a brain would pay the huge intel tax just to get 5 extra fps.....

 

According to current pricing information the i9-11900k would actually be $369 cheaper than Ryzen 9 5950X...

[jagdtigger]

2 minutes ago, BiG StroOnZ said:

 

According to current pricing information the i9-11900k would actually be $369 cheaper than Ryzen 9 5950X...

Knowing intel its a monolithic design with pretty horrific yields so highly doubt it.

[Blademaster91]

An i9 at i7 pricing? Well it is more of an i7 with 8 cores instead of 10 with the 10th gen i9.

But I doubt it, unless Intel is really desperate to get the gaming lead back.