First professional review of Intel's Rocket Lake (11700K) is out.. and it's a disaster

[TruthBomb]

Oh dear. A few weeks ago I bought a 10700K (new build) and I had slight buyers regret knowing the new chips would soon be out soon. Not anymore. Guess I'm all set. 

[Loki0111]

10 hours ago, pythonmegapixel said:

That's the kind of thing I expected, though there can't be an actual blanket "don't sell AMD bribe" because almost every major OEM has some AMD models; also that would probably lead to a massive antitrust case.

 

It's probably done significantly more subtly than that.

It honestly wouldn't shock me. Intel has actually been convicted for doing exactly that previously. In 2009 Intel has to pay AMD something like $1.25 billion in an anti-trust settlement in the US and another $1.45 billion in an anti-trust settlement in the EU in 2014 for doing exactly the same thing again. Plus the settlement in Japan for doing the same again thing, etc. I think in one case Intel actually stole tech they got access to from AMD's records during the discovery process in court and used it in their upcoming products then got nailed for that.

 

In short Intel never seems to have any problem throwing around money to try and muscle vendors away from their competitors.

 

https://www.networkworld.com/article/2239461/intel-and-antitrust--a-brief-history.html

[LAwLz]

Sigh...

AMD laptops aren't rare because of a conspiracy theory from Intel. They are rare because:

1) Supply issues from AMD.

2) Zen was pretty bad for laptops up until very recently. AMD had desktop processors figured out before laptop CPUs. See this review as an example. Intel's mobile CPUs are like 2-3 generations ahead of their desktop CPUs.

3) AMD doesn't have, or at least haven't had until recently, as good relationships with laptop manufacturers, nor have they had the staff to help with designing the laptops. These things take time, so you need multiple generations of good products before the relationship and expertise are built up and you start getting major design wins.

[Loki0111]

3 minutes ago, LAwLz said:

Sigh...

AMD laptops aren't rare because of a conspiracy theory from Intel. They are rare because:

1) Supply issues from AMD.

2) Zen was pretty bad for laptops up until very recently. AMD had desktop processors figured out before laptop CPUs. See this review as an example. Intel's mobile CPUs are like 2-3 generations ahead of their desktop CPUs.

3) AMD doesn't have, or at least haven't had until recently, as good relationships with laptop manufacturers, nor have they had the staff to help with designing the laptops. These things take time, so you need multiple generations of good products before the relationship and expertise are built up and you start getting major design wins.

I am not defacto stating that is untrue in this situation. I am simply stating Intel has been found guilty in court of doing exactly this type of thing to AMD multiple times before and has to pay several large settlements out over the years when they got nailed.

 

In fact the majority of the time when police raid their offices its usually a safe bet it was something anti-trust related to AMD.

[porina]

Another tech site has also purchased and published testing results:

https://www.hardwareluxx.de/index.php/artikel/hardware/prozessoren/55649-prozessor-non-grata-rocket-lake-s-als-core-i7-11700k-im-vorab-test.html

 

I'm not familiar with their testing process and as I'm reading it machine translated from German that doesn't help.

 

Would say one thing caught my eye, they state the PL2 value is 229W. I assumed a value much lower than that earlier, using a similar ratio seen on previous gen CPUs.

 

Edit: I just searched around for the PL2 value, and different "news" sites seem to be claiming different values. So I'd treat that with caution until we get something more official.

[leadeater]

2 hours ago, porina said:

Edit: I just searched around for the PL2 value, and different "news" sites seem to be claiming different values. So I'd treat that with caution until we get something more official.

Yep best to wait, checked the 11th Gen Technical documentation on Intel's website which was updated in Jan and the desktop processors are not in it. I'd be reluctant to believe any PL2 figures given without some detailed information on BIOS settings, it's a configurable value so very hard to trust what the motherboards are doing. Just safer to wait for official Intel spec.

 

That being said 10th Gen i9 was 250W PL2 and i7 was 229W PL2, without AVX offsets. So basically it's not that surprising to see the 11700K using peak ~224W in an AVX2 workload as that is what a 10700K would use if it weren't hitting some other limitation first. I would be interested to see the same test Anand did with a raised PL2 to see where it tops out at, similarly to how the 10700K tops out at 204W so is not PL2 limited under that workload.

[Fnige]

On 3/7/2021 at 12:58 AM, leadeater said:

First gen Intel Atom?

Explain? Cant find anything on it

[JediFragger]

1 hour ago, SlimyPython said:

Explain? Cant find anything on it

It was bad but redeemed in later shrunken revisions with more than 1 core (2005/06 is ringing a bell, a fair while ago now).

[porina]

https://www.anandtech.com/show/16549/rocket-lake-redux-0x34-microcode-offers-small-performance-gains-on-core-i711700k

 

Anandtech have obtained another mobo with a newer microcode than the one they used in the early review. They tested it and on average there is a +1.8% improvement over before. They question if it is due to the newer microcode, or if the different mobo vendor did things differently. L3 cache latency was lower, and AVX-512 peak power and performance was a bit lower.

 

Apparently the new test results will be added to the original review "soon" if people want to poke at the numbers further.