Intel Rocket Lake (14nm) Review: Core i9-11900K, Core i7-11700K, Core i5-11600K | Hacker News Reader

Intel Rocket Lake (14nm) Review: Core i9-11900K, Core i7-11700K, Core i5-11600K

anandtech.com

57 pointspantalaimon23 comments

How on earth does Intel think 9 skus that are all 6 cores / 12 threads over a difference of $100 makes any sense? 9 products from $157 to $262 that are all not really that different.

And 5 different products all called 11700 and 5 different products called 11900 for a total of 10 products with 8c / 16t spanning from $298 to $539?

The main market for Intel is not the consumer, it's the HP's and Dell's of the world.

Even still those are selling to consumers. How do you expect an HP or a Dell to figure out how to sell a user on an i5-11500 instead of the slower & more expensive i5-11600T?

Why isn't the i5 lineup just 11600 = the 125w part, 11500 = 65w part, and 11400 = 35w part? Is there really enough gap in product stacks to justify a different 11400T vs. 11500T at a $10 difference?

It's about silicone lottery. They could ignore best chips and sell them along bad chips at the same price. Or they could market best chips as another product with a higher price. And that's the whole spectrum.

Sure, but we're talking a $10, or ~5%, difference. Either these bin identically (since Intel is charging an ~identical amount for them), or one of these will be permanently out of stock making it functionally not exist in the stack anyway. The actual profit difference here will be even less as simply having that extra sku costs overhead. So what practically changes if the 11500T just didn't exist?

i7-11700T vs. i7-11900T would be silicon lottery in play, that's $323 vs. $439 for a +100mhz clock bump. Although then there's the question of "what is this 35w desktop CPU market that will pay for a $439 CPU?"

it doesn't really make sense for dell to offer a choice between T and non-T for the same model. the only reason to use a T part is to stay under a power/TDP budget.

Is it possible that this proliferation of SKUs is how Intel gives each of HP, Dell, Lenovo, Acer, etc exactly what they are asking for? (I presume, if you buy enough Intel CPUs, you have purchasing managers who are constantly chatting back and forth with sales engineer types at Intel)

As a customer, if you’re buying as few as 20,000 PCs a year, you’ll be able to get some sort of “investment” from Intel.

The reality is they segment based on availability and price. Nobody really cares.

HP has a little more wiggle room because of their AMD line. Dell gets dibs on parts.

What boggles my mind are the inconsitencies. E.g. a i9 10900F is clocked lower than a i9 10850K

Wouldn't you think higher number = faster, especially when core count is the same?

This is not really new, it's because of the letter at the end of the product name. Traditionally Ks are power hungry overclockable chips while F are more energy-conscious.

EDIT: Never mind I just confirmed your point. T are the energy-conscious ones, not sure what F are. This is all very confusing.

F means it lacks an IGP. in a certain sense you were right; the F variant sacrifices frequency for lower TDP vs a K variant. but there are also KF parts that are unlocked and also lack an IGP. and of course there's a different system entirely for laptop parts.

I have been incredibly skeptical of frequency numbers. Case in point, the AMD FX-9590 had a 4.7 GHz base an 5.0 GHz boost back in 2013 [0]; and the Ryzen 9 5950x has a 3.4 GHz base an 4.9 GHz boost [1].

To me, the frequency has become a "blob". It means nothing. If anything, I see more frequency and I think more power draw and less stability.

Real performance is a function of IPC and frequency and that is the only thing that really matters.

[0]: https://www.amd.com/es/products/cpu/fx-9590

[1]: https://www.amd.com/en/products/cpu/amd-ryzen-9-5950x

none of those numbers fully describe the performance of a CPU, but idk why you would be "skeptical" of frequency numbers in particular. IPC itself is somewhat arbitrary. in the real world, IPC constantly varies with the mix of instructions, cache-friendliness of the program/data, memory performance, etc.

> none of those numbers fully describe the performance of a CPU, but idk why you would be "skeptical" of frequency numbers in particular

Because I have seen quite a few consumers get guided too much by the "more frequency, more better" motto.

Moreover, it doesn't stop there. I have seen papers where the only thing cited is the frequency of the processors, and no other reference to the manufacturer/architecture/version of the processors used.

> IPC itself is somewhat arbitrary. in the real world, IPC constantly varies with the mix of instructions, cache-friendliness of the program/data, memory performance, etc.

Absolutely, I would also be skeptical of an "IPC" metric. Performance is the only thing that counts.

> To me, the frequency has become a "blob". It means nothing.

That seems reasonable across microarch versions, but presumably is totally reasonable within a product-release-cycle? i.e. to compare the FX-9590 to an FX-8350 based off clock speed. If it's the same core count and microarch, presumably clock is about the only difference, aside from other features like PCIE lanes, cache, etc.

> That seems reasonable across microarch versions, but presumably is totally reasonable within a product-release-cycle

Sure, but the performance doesn't scale linearly with the frequency. So how can we account for that? Also, the i9 10900F and i9-10850K have the same turbo. How long can they sustain in for? On how many cores? What frequency is left afterward? This gets even weirder with more recent technologies that push the frequency higher depending on power delivery stability and thermal constraints.

I agree; the frequency is indicative of performance within a product lifecycle. But I find it too meaningless to give an approximate delta on the performance. And this assumes the process you are running is not too much memory bottlenecked or I/O bound and this delta in performance can be measured. (And the setup can sustain higher thermals).

While not perfect, I prefer to look up benchmarks that resemble my workload and call it a day. Processors are too complicated for me to try to compare with simple numbers.

Well said. That’s exactly your I feel about it at this point. The thermal and power properties and how they interact with newer boost algorithms is unpredictable. I try to find my workload benchmarks and judge off that. For me that would be VS but any compilation comparison suffices.

Decent IPC increase, but really good liquid cooling is required in order to get it. Prebuilts like Dell will likely use tiny coolers that will limit the performance of these parts.

Apple's M1 illustrates how far behind Intel is in everything, and AMD's upcoming Zen 4 is supposed to have a staggering 20% increase in IPC.

It is a good thing Intel has massive market inertia, because the next few years are going to be rough for them from a tech standpoint.

Actually to me this generation seems like its even a regression compared to the previous due to the reduction of cores and increase of Power consumption.

Does look good for Intel :/

> It is a good thing Intel has massive market inertia

On desktop they don't. Although they got "lucky" in that demand right now is at an all time high so literally anything will sell.

That's fine, but is there an Intel chipset yet with PCIe 4 and even PCIe 5? AMD had PCIe 4 two years ago. How about Thunderbolt 4?

Yes. The new X590 supports PCIe 4 when paired Rocket Lake and has TB4 support.

Reply on news.ycombinator.com