Intel Rocket Lake 11000 retail reviews

[Mats]

Intel Core i7-11700K Review: Blasting Off with Rocket Lake

www.anandtech.com

There are two tables that stand out from the rest, everything else is less remarkable:
(It does take a lead in some benchmarks, but not like with AVX)

It seems to be a tad hot running.

AVX-512 is nice tho, but we've known that for TWO years now since Cannon Lake.

 

[mouacyk]

The improvements Intel has made to its processor seem to do very little in our gaming tests, and in a lot of cases, we see performance regressions rather than improvements.

+19% IPC for a whole lot of nothing. Ouch.

 

[Mats]

+19% IPC for a whole lot of nothing. Ouch.

Hey, at least it's not Skylake!
This is no surprise tho, Alder Lake will come soon. We thought we knew why, and now we know.

 

[v12dock]

Hot and mediocre I guess I was expecting a bit more... 14nm must really be holding Intel back.

 

[Mats]

14nm must really be holding Intel back.

With the old rumor about reduction of core count this was expected. Although I didn't expect it to be that hot.

 

[tussinman]

Prices for 11700 and 5800x are too high. If I had to buy now it would be the 10700F for $250.

 

[tabascosauz]

Intel Core i7-11700K Review: Blasting Off with Rocket Lake

www.anandtech.com

There are two tables that stand out from the rest, everything else is less remarkable:
(It does take a lead in some benchmarks, but not like with AVX)

It seems to be a tad hot running.

View attachment 191091

AVX-512 is nice tho, but we've known that for TWO years now since Cannon Lake.

View attachment 191092

A "tad"? That's Chernobyl right there

Did Intel forget to put an AVX offset in its pre-release firmware? No way they intended for it to run AVX-512 workloads at its normal 4.6GHz all-core turbo. It doesn't take a genius to put two and two together to figure out that 8 large cores doing AVX-512 on air, without any AVX offset whatsoever, are going to burn down the house.

And 104C? Did Intel move TjMax up just for this chip?

In all fairness, outside of that, things look nothing out of the ordinary. So it might just be pre-release shenanigans.

Inevitably someone's going to come in here protesting that Intel "figured it out" with Ice Lake and doesn't need to offset AVX anymore. To that end, it's fairly obvious that when the pitiful Ice Lake chip is only doing 3.3GHz in normal workloads, there's no need to put in an offset.

 

[oxrufiioxo]

And 104C? Did Intel move TjMax up just for this chip?

They did it for the 9900k as well you could raise TjMax to 105C on some boards.

Pretty shocked an 8 core is hitting 300w although I somewhat expected it.

 

[ViperXTR]

Prices for 11700 and 5800x are too high. If I had to buy now it would be the 10700F for $250.

Its still beyond 250 on our region but a lot cheaper than 11700 and 5800X

 

[Melvis]

Oh dear! Its worst then I thought! and using more then double the power for less over all performance, yikes.

 

[Valantar]

The overall performance of this is really disappointing. Wow. Barely any improvement at all over the previous generation, and nowhere near catching up to AMD. I was expecting this to run hot, but be competitive, likely surpassing Zen 3 at equivalent core counts. Instead we get this hot mess. Makes me very happy going for a 5800X, but sad for the CPU market overall.

To be fair though, those power draw numbers are taken out of context and are a bit apples and oranges - the 11700K number is AVX512 peak power, while every other chip is using AVX2. The 11700K consumed ~225W under AVX2 and ~150W under non-AVX loads in AT's testing.

 

[lynx29]

Intel Core i7-11700K Review: Blasting Off with Rocket Lake

www.anandtech.com

There are two tables that stand out from the rest, everything else is less remarkable:
(It does take a lead in some benchmarks, but not like with AVX)

It seems to be a tad hot running.

View attachment 191091

AVX-512 is nice tho, but we've known that for TWO years now since Cannon Lake.

View attachment 191092

I will never understand why you use a 4750g in this benchmark showing instead of a 5600x... but hey to each their own I guess. actually they need a 7700k in here as well or an 8700k for more reference comparisons. rocket lake uses way too much power though, my 5600x uses 1/3 the power almost and goes toe to toe with it.

 

[Valantar]

I will never understand why you use a 4750g in this benchmark showing instead of a 5600x... but hey to each their own I guess. actually they need a 7700k in here as well or an 8700k for more reference comparisons. rocket lake uses way too much power though, my 5600x uses 1/3 the power almost and goes toe to toe with it.

You can check all of those in their Bench benchmark engine. The included comparisons are meant to all be 8-core chips for a like-for-like comparison.

 

[R0H1T]

+19% IPC for a whole lot of nothing. Ouch.

Doesn't apply across the board ~

In our testing, we saw the following:

• Single thread floating point: +19.0%
• Multi-thread floating point: +19.5%

Sounds great, right?

• Single thread integer: +13.0%
• Multi-thread integer: +7.3%

+19% IPC for a whole lot of nothing. Ouch.

Doesn't apply across the board ~

In our testing, we saw the following:

• Single thread floating point: +19.0%
• Multi-thread floating point: +19.5%

Sounds great, right?

• Single thread integer: +13.0%
• Multi-thread integer: +7.3%

Have no idea what's happening with the nested quotes

 

[londiste]

To be fair though, those power draw numbers are taken out of context and are a bit apples and oranges - the 11700K number is AVX512 peak power, while every other chip is using AVX2. The 11700K consumed ~225W under AVX2 and ~150W under non-AVX loads in AT's testing.

Outside the extended time, non-AVX loads are spec power consumption. 125W PL1, 156W PL2 (PL1 at TDP and PL2 at 1.25x TDP) but Tau is obviously oversized at 400 or so seconds.

I really wish Anandtech would have tested this thing at spec as well rather than only whatever limitless way motherboard defaults to for AVX2/AVX512.
Also, at this point I would really love to see frequency and power measurements for each individual test.

 

[thesmokingman]

Doesn't apply across the board ~


Have no idea what's happening with the nested quotes

Yea, the IPC claims are like... kinda BS.

Oh. While Intel’s claim of +19% is technically correct, it only seems to apply to math-heavy workloads. The benefits of non math-based throughput are still better than average, 7-13%, but vary rarely do Intel’s big claims come with an easily identifiable asterisk.

I swear, Intel is hardly ever forthcoming with the facts.

I just wrote this below at anand. It's really hilarious, the 5800x is not THE power miser in AMD's stack but it sure looks like one here. And in Anand's words the AVX2 loads were light, smh.

It's shocking that it uses 50% more peak power in avx2 loads than the equivalent 5800x. When we put that into context on AMD side, the 5800x is the single cpu in AMD's stack with the highest TDP per core. Considering that it has the same TDP as the 5900x and 5950x but only has 8 cores to spread that TDP around. These Intel cpus make the 5800x look like a power miser lmao in comparison!

 

[HD64G]

And there were people believing this gen of Intel CPUs will win over Zen3.

 

[R0H1T]

Its' a good thing Intel didn't release any 10c/20t parts ~

 

[thesmokingman]

Hey, at least it's not Skylake!
This is no surprise tho, Alder Lake will come soon. We thought we knew why, and now we know.

Alder Lake... what was or is the expectation?

 

[Valantar]

Outside the extended time, non-AVX loads are spec power consumption. 125W PL1, 156W PL2 (PL1 at TDP and PL2 at 1.25x TDP) but Tau is obviously oversized at 400 or so seconds.

I really wish Anandtech would have tested this thing at spec as well rather than only whatever limitless way motherboard defaults to for AVX2/AVX512.
Also, at this point I would really love to see frequency and power measurements for each individual test.

Yeah, spec level tests would be nice, though it's understandable that they don't do so given the 2x increase in workload for a setting level that is essentially nonexistent on consumer motherboards. And it's not like Intel discourages this behaviour from their motherboard partners, after all. Still, it would be very interesting to see.

Alder Lake... what was or is the expectation?

Similar IPC increase as Ice Lake to Tiger Lake. Which is to say relatively minor, but there. And a bunch of atom cores for no reason anyone can understand, at least on desktop.

 

[R0H1T]

Alder lake is a white elephant (whale?) & totally not worth the wait!

I imagine Intel's only propping it up because they can't compete with AMD's "true core" counts.

 

[thesmokingman]

Alder lake is a white elephant (whale?) & totally not worth the wait!

I imagine Intel's only propping it up because they can't compete with AMD's "true core" counts.

It's gonna be a fiasco since windows won't know what to do with big vs little. Intel has a ton more pull with Microsoft, but then again this is Intel. They're not great at new things.

 

[Valantar]

Alder lake is a white elephant (whale?) & totally not worth the wait!

I imagine Intel's only propping it up because they can't compete with AMD's "true core" counts.

To be honest, AMD's 12- and 16-core chips have relatively little utility to consumers. A lot to a small selection (and they're invaluable to people with scalable workstation workloads but not astronomical hardware budgets), but most people don't need more than 8 for the next few years, and 6 is still perfectly fine even for relatively heavy workloads. So if Intel could bring a great 8-10-core architecture to market they would do very well in all the high volume markets. The problem is that these days they can't even do that.

 

[Tomgang]

Well this is one review. I will have to see others first before making a final conclusion.

But so far it really looks grim for Intel. Most disappointing performance. Can barely keep up with 5800X in games (I expected at least Intel to win here) and heat and power consumption. Holy shit, that is bad as in really bad.

If other reviews confirms this most disappointing performance. I will for sure go Zen 3. Actually all ready getting hardware home for a zen 3 build.

 

[R0H1T]

To be honest, AMD's 12- and 16-core chips have relatively little utility to consumers.

It is (not much useful) but then 12c/20t or whatever combinations they come up with is hardly better. I know I said in the past big.Little would be fun to watch on the desktops but I sure as hell ain't buying one for posterity's sake!