team blue Ice Lake Details Released

[gutz00]

https://www.anandtech.com/show/14664/testing-intel-ice-lake-10nm

 

Intels Next Gen Processors code named Ice Lake have been benchmarked. 

Do you think they can compete with third gen Ryzen?

Personally I don't think they can. It's great to see Intel hit 10nm but we see a dip in Base Frequency. Personally not a trade off I'd like to see. 

[TheNaitsyrk]

1 minute ago, gutz00 said:

https://www.anandtech.com/show/14664/testing-intel-ice-lake-10nm

 

Intels Next Gen Processors code named Ice Lake have been benchmarked. 

Do you think they can compete with third gen Ryzen?

Personally I don't think they can. It's great to see Intel hit 10nm but we see a dip in Base Frequency. Personally not a trade off I'd like to see. 

Intel chips always were good overclockers except 6800K f that chip.

 

Also, jump from 14nm to 10nm will increase IPC and lower clock speed shouldn't impact it much. 

 

IPC improvements will help.

 

Also, don't forget these are really low TDP chips, I imagine something what could be next gen 9900K will have higher clock speed, maybe even 4.5Ghz or so?

 

Then you OC on top.

 

Looking forward to see what they will release and learn on their mistakes by not releasing early.

 

Also I expect lower TDP, (or they will make it higher to boost clocks).

[DrMacintosh]

All I care about is integrated Thunderbolt and the much improved Iris Graphics. These should bring a big performance improvement to the 13” MacBook Pros, especially in regards to eGPU performance since there won’t be a latency penalty. 

 

The T2 chip already can do a lot of what Intel added to these chips, like HEVC encode/decode, but the T2 can’t do d*ck to improve underwhelming graphics performance. 

 

Hoping that we see a refresh of the Mac with these CPUs in September. 

[mr moose]

I'll wait for more tests before commenting on how it performs compared to anything else, however this does seem to suggest that 10nm for Intel isn't the sinking ship people are making out.  But then I have said before you can never guess what Intel will do next or bring to the table, so no one should be surprised eitherway.

[Guest]

18 minutes ago, gutz00 said:

Personally I don't think they can. It's great to see Intel hit 10nm but we see a dip in Base Frequency. Personally not a trade off I'd like to see. 

Lets go back to 90nm pentium 4 then. 

[BiG StroOnZ]

Here are all the official slides from Intel, for those interested:

 

Spoiler

 

Source: https://videocardz.com/press-release/intel-launches-first-10th-gen-intel-core-processors

[porina]

31 minutes ago, gutz00 said:

Do you think they can compete with third gen Ryzen?

Yes from an architecture point of view, but the problem is that desktop parts based on this are not on the radar. After mobile they said they will do server, which I guess could include a dotted line to HEDT (generation after rumoured Cascade Lake as next HEDT). Mainstream desktop versions could be a long way off.

 

31 minutes ago, gutz00 said:

Personally I don't think they can. It's great to see Intel hit 10nm but we see a dip in Base Frequency. Personally not a trade off I'd like to see. 

As others have said, Intel are targeting power limited mobile parts first. We don't know how they do on the top end without that constraint.

 

13 minutes ago, mr moose said:

I'll wait for more tests before commenting on how it performs compared to anything else, however this does seem to suggest that 10nm for Intel isn't the sinking ship people are making out.  But then I have said before you can never guess what Intel will do next or bring to the table, so no one should be surprised eitherway.

The architecture improvement claims started around their architecture day earlier in the year. The difference now is, we have some benchmarks run by persons outside Intel confirming that to an extent, even if hardware was under Intel control (pre-release, not everything is representative yet) and the testers only had limited time.

[mr moose]

10 minutes ago, porina said:

even if hardware was under Intel control (pre-release, not everything is representative yet) and the testers only had limited time.

Hence why I'll wait for more testing.

[Guest]

27 minutes ago, Genwyn said:

wow, a pc tech strawman

dont see that too often

 

very different hardware, very different technologies at the time

clock speeds mattered way more then than they do now

lower wattage, high effeciency, high core count cpus to more adequately compete with ryzens midranged options would be great for the market

not to mention beneficial for the laptop market

this isn't the first time I thought that /s wasn't necessary 

[Unixsystem]

Barely beating out Whiskey Lake on a new arch with faster memory under an extremely controlled test environment isn't really a great sign. That said, until we actually have shipping product for trustworthy reviewers to look at on their own time, this data is largely irrelevant. 

 

The real question to me is what they mean by "shipping this year". Do they mean mass production in every major high end laptop for the holidays, or are they going to pull another Cannonlake and push it into one SKU in one market just to say "see, I told you we'd ship it". 

 

Edit: I'm actually curious how the IPC stacks up against Zen 2. I'd be pretty funny if we wind up in a situation where both have near identical IPC and clockspeed. What will the fanboys argue about then? 

[ravenshrike]

16 minutes ago, Waffles13 said:

What will the fanboys argue about then? 

Seeing as the yields on 10nm still suck royally? Cores.

[Unixsystem]

Just now, ravenshrike said:

Seeing as the yields on 10nm still suck royally? Cores.

That last line was /s. 

 

You and I both know that they'll argue about the color of the packaging if that's all they're given to work with. 

[jiyeon]

49 minutes ago, BiG StroOnZ said:

Intel SKU names just become a mouthful. At least it adheres to the usual naming convention though.

[MMKing]

Umm cool i guess. But why would anyone, who already own a Whiskey lake laptop, upgrade to an Ice Lake laptop? The performance uplift is nearly meaningless.

 

I am curious if this is truly the best they can do, or if they are trying to pull the same shit they did for nearly a full decade beginning with the Core-i, where performance year over year was pathetic.

[Taf the Ghost]

4 hours ago, mr moose said:

I'll wait for more tests before commenting on how it performs compared to anything else, however this does seem to suggest that 10nm for Intel isn't the sinking ship people are making out.  But then I have said before you can never guess what Intel will do next or bring to the table, so no one should be surprised eitherway.

These are about all they can yield on the node and they're looking at nearly a 1 Ghz clock regression. If it's not a AVX load, it's pretty much clock for clock the same as the advanced Skylake parts. And most of that improvement seems to be from moving from 2133 to 3733 memory. (Benefit probably taps out around 3000 memory.)

 

Beyond yields being horrible, the node is a massive clock regression and the AVX units burn way too much power. The only real value for these parts are the fact the iGPU has finally been upgraded. They'll work, but 10nm is still a disaster. Icelake does seem like it has potential, but no one really is going to see it until maybe Alder Lake.

[BiG StroOnZ]

 

[mr moose]

26 minutes ago, Taf the Ghost said:

 but 10nm is still a disaster.

 

Yer, I'm still going to wait for the dust to settle on all this before I start making calls.

[Taf the Ghost]

56 minutes ago, mr moose said:

 

Yer, I'm still going to wait for the dust to settle on all this before I start making calls.

Icelake doesn't seem like much interesting, but the iGPU + updated encoders seems like it'll be the key feature. Intel stuffs enough creature features on their products that those should be enough to at least look nice. Volume will still be in 14nm parts though.

[williamcll]

Might be enough to beat ryzen 3 3200G

[RejZoR]

9 hours ago, TheNaitsyrk said:

Intel chips always were good overclockers except 6800K f that chip.

 

Also, jump from 14nm to 10nm will increase IPC and lower clock speed shouldn't impact it much. 

 

IPC improvements will help.

 

Also, don't forget these are really low TDP chips, I imagine something what could be next gen 9900K will have higher clock speed, maybe even 4.5Ghz or so?

 

Then you OC on top.

 

Looking forward to see what they will release and learn on their mistakes by not releasing early.

 

Also I expect lower TDP, (or they will make it higher to boost clocks).

Manufacturing process doesn't inherently improve IPC. It gives space to make architectural changes to achieve IPC boost, but without architectural changes you can downsize it to single atom and it wouldn't make it any more efficient per clock.

[realpetertdm]

Intel's still fucked in the desktop market though, unless they pull out massive improvements before mid 2020 AMD's gonna leave them in the dust with Zen 3

[porina]

8 hours ago, Waffles13 said:

The real question to me is what they mean by "shipping this year". Do they mean mass production in every major high end laptop for the holidays, or are they going to pull another Cannonlake and push it into one SKU in one market just to say "see, I told you we'd ship it". 

The indication is they have volume availability of the CPUs for multiple vendors to sell whatever models they want to. 

 

8 hours ago, Waffles13 said:

Edit: I'm actually curious how the IPC stacks up against Zen 2.

Hard to say without a direct like for like comparison but they should be close for general use. There may remain some specific biases in particular use cases.

 

8 hours ago, Waffles13 said:

I'd be pretty funny if we wind up in a situation where both have near identical IPC and clockspeed. What will the fanboys argue about then? 

Fanboys don't need anything more than mention of a brand to start an argument.

 

8 hours ago, MMKing said:

Umm cool i guess. But why would anyone, who already own a Whiskey lake laptop, upgrade to an Ice Lake laptop? The performance uplift is nearly meaningless.

Why would anyone upgrade from n-1 gen tech? The vast majority don't. People tend to wait multiple generations until there is a big enough jump to be worth getting.

 

7 minutes ago, realpetertdm said:

Intel's still fucked in the desktop market though, unless they pull out massive improvements before mid 2020 AMD's gonna leave them in the dust with Zen 3

How much uplift are you expecting from Zen 3 over Zen 2? Zen 2 relative to Zen(+) was a fairly big step. I'm not saying don't expect improvements, but I wouldn't expect as big a jump again as low hanging fruit has been picked. I suspect future IPC improvements will be in smaller increments, leaving the main trick of throwing more cores at things. My wish list for future Zen would be unified L3 cache at CCD level (basically ditch CCX and make CCD the smallest logical unit) and un-cripple the CCD write bandwidth. These limitations have been a pain to deal with on Zen 2.

[Likwid]

I hope it will support HDMI 2.0 out of box without any adapters

[TheNaitsyrk]

7 hours ago, RejZoR said:

Manufacturing process doesn't inherently improve IPC. It gives space to make architectural changes to achieve IPC boost, but without architectural changes you can downsize it to single atom and it wouldn't make it any more efficient per clock.

I didn't know we have a CPU manufacturer / maker in here.

[pas008]

59 minutes ago, TheNaitsyrk said:

I didn't know we have a CPU manufacturer in here.

manufacturer?

many companies outsource the manufacturing part for their cpus/chips/etc