Intel's upcoming 10nm and beyond

I have shared, to deaf ears.

What I state doesn't need any background except for having used computers for the last decade or more. I'm sure you're capable of following if you wanted to.

I used to do what you're asking from me, but there is no point here. I have learned to make my point as succinctly as possible.

 

i donno guys. amd looks promising and zen2 might really tilt the crown that lasted for a decade.

intel's 10nm so far has been garbage efficiency and garbage yield, i dont see how they can improve especially they clearly showed that desktop do not get 10nm until 2021 for it's 10nm+, meaning they know 10nm has junk frequency and simply won't please desktop overclockers. on the other hand, server side of thing at lower frequency might get a much better efficiency on first iteration of 10nm, so far the 10nm lenovo book as debunked it so we'll have to see how intel solves that problem.

finally if both 10nm and zen 2 fail to hit expectation we can always wait

 

Sure it looks promising on paper and 'might' can be said about anything else too if the will was there.

Of Intel's 10nm, we've only seen part one so far...

If we're actually hoping for the best from both sides, then everything looks rosy.

Looking at the same info and calling it to fail just because you can't imagine it going better and then looking at the other side and seeing beauty is your choice, of course.

I tend to look at things a little more balanced and less emotionally.

 

Although I've seen this previously, I can't find the specific reference about Intel converting 3 of the 4 10nm FAB's (previously assigned to produce 10nm wafers) to 14nm to meet production requirements.

Intel turns 10nm production back into 14nm
29. October 2018 Florian Maislinger
https://www.pcbuildersclub.com/en/2018/10/intel-turns-10nm-production-back-into-14nm/

This Wikipedia page lists Intel's manufacturing sites and has details as to which sites are for 10nm / 14nm production:

List of Intel manufacturing sites
https://en.wikipedia.org/wiki/List_of_Intel_manufacturing_sites

I'd previously heard that Intel has 15 active wafer fabs, the list above shows 14.

How Many Manufacturing Fabs Does Intel Have?
https://www.intel.com/content/www/us/en/support/articles/000015142/programs.html


Global Manufacturing at Intel
https://www.intel.com/content/www/us/en/architecture-and-technology/global-manufacturing.html

 

nope, its not on paper. demo has been shown by amd a 65w 8core skus beating 9900k at full core turbo 4.7ghz in CB15, which the 9900k PL2 is at least using ~120w.

this shows at least few things, ipc and frequency has improved and even with a lower frequency, zen2 has had a decent ipc boost enough to be on par with intel.

 

You're right if we compare like for like, the unreleased Ryzen desktop is beating a year old Intel design. And we take AMD's word for it, which we won't let Intel get away with. Yawn.

When Intel answers to what actually ships from AMD, we can do a real comparison then, but now? You're just showing your bias, again.

Over on mobile platforms, the situation hasn't changed. Intel is actually leading today. See my links previously to the AMDfied ThinkPad.

You keep trying to see clearly with those glasses on, but they aren't doing you any favors it seems.

 

not at all. see, this is where you are wrong. I am comparing exactly like for like and since intel's new 10nm that showed up in lenovo book was worth basically trash (not 1yr old intel design mind you, not 14nm+++ mind you), I am comparing zen 2 es vs intel's latest 10nm, which was worse than their 14nm+++. besides, intel's core arc is like 10 yrs old so effectively if you wanted, you could simply say that im comparing 10yr old cpu vs ryzen's future design.

i could careless either way though no matter how you talk about it, since ultimately Id use whats available out there to compare.

bias or what not, our value differs and thats just how it is.

funny it came from you. i'm sure @ajc9988 remembers very clearly when he had to stood on his own against the both of us a year ago when it was zen+ vs intel 14nm++.

at the time Intel had the performance crown in ipc/frequency and I stood by it regardless of intel's ethic, value, power efficiency etc. I guess you're the one thats blind, couldn't even see your own glasses.

 

Just to be clear, this is what you're comparing against?

See:
https://www.tomshardware.com/news/intel-10nm-processor-core-i3-8121u,37054.html

This is the only released 10nm platform so far (I referred to it as gen 1, above).

 

aye

 

You know you are comparing 2017 tech released 2018 (a year ago) to what is currently not available yet.

But I'll take your word for, you're not biased at all.

 

its the only 10nm product out, kinda the reality. sux doesnt it, your favorite company has garbage to show

 

So again things are starting to take a turn for the worse. If you two can't play nice, I'm going to have to step in...

 

No, it's not the only thing that is out. Just that what is out can't be had yet by you or me for a few more weeks or so.

If you make decisions based on over 2-year-old results, well, what can I say?

I don't see that? Just having a conversation, no?

 

It's a back and forth now, but things tend to escalate when you start to take jabs (like saying "your favorite company has garbage to show") instead of keeping up with posting reasonable information.

 

Yes, and I have been guilty of that in the past.

These days, I am much more in control, but thanks again for the fair warnings.

 

basically i took a look at anandtech's review on the 10nm in laptop and efficiency isnt all that better than 14nm. if 10nm has improved since then thats great but intel still has nothing to show on the "newer" 10nm yet.

 

I like Intel as a company more (since they introduced Pentium M back in 2003 anyway), but honestly with all the security flaws uncovered last two years which mostly affect Intel, while AMD and ARM CPUs are either immune or take less of a performance hit after mitigations, I'd definitely prefer my next machine to have an AMD cpu.

 

Ever since I found out how badly Intel behaved in order to cut AMD out (by bribing OEM's among other things and having to pay a ridiculously pitiful fine as a result) I hadn't really liked Intel. I considered them for a new laptop at some point, but only because AMD didn't have any kind of presence to speak of... but then I weighted my options, decided to hold off and see what AMD will do with Zen (and they hadn't disappointed).

Intel (to me) doesn't really offer anything in the performance area that would make me want to get their CPU's right now.
Sure, all their CPU's come with iGP's (which is a bonus for laptops) and they clock higher than AMD for the time being, but I'm not too bothered with that bit (and the performance difference between same core parts doesn't exactly justify the hefty price premium Intel comes with - and lets face it, the performance diff. is not a huge one to begin with in games... and AMD has a pretty good parity in multithreading with same core parts as Intel - especially when Ryzen is paired with high clocked RAM that has low latencies). Plus, as you say, the security vulnerabilities (not to mention the price) in Intel CPU's make me wonder how can anyone currently justify getting them - but then I remember just how big of a presence Intel has in the market and mind-share in general.

I mean sure, the average person might not care too much about security vulnerabilities, nor would they necessarily be affected, still, I wouldn't want to take those chances... and data centers in all probability wouldn't want that either (as they value security among other things and could be considered to be at far greater risk - makes me wonder if the people who exploit various software weaknesses use Intel vulnerabilities to make their jobs easier given how widespread they are).

I'll consider Intel in the future if their practices, security and overall technology improves.
Right now, they're on a 'time out' as far as I'm concerned (plus, I really have no need to get anything new given my Acer Helios - lol, who would have thought that I'd go from Acer 5930G to a high-end Acer Helios 500 with all AMD desktop grade hw inside - I thought I would have used/kept the Asus GL702ZC for a long time, but given how many failures it ended up with (shame on you Asus for botching an otherwise great design), best choice for me was to force the retailer to give me a refund and get Acer Helios with all AMD hw (plus I got it for cheaper than the Asus, and I hadn't regretted it because the 'monster' is really cool temperature wise and emits a very low background hum which is due to the fans rotating at low speed and is actually soothing, but otherwise, not really audible or distracting - heck even when I fire up an intensive game the thing stays both quiet and cool).

10nm Intel parts looked to be interesting until they started delaying the node and parts perpetually. Right now, at least several factors prevent them from bringing out 10nm parts:
1. Yields (which seem to be abysmal and suitable only for a few mobile parts).
2. Security vulnerabilities (for which I think it was said they need a completely new uArch - and I think this may have played a part for Intel bringing in Keller - at the very least it wouldn't surprise me).

Also, Intel seems to be talking more about 7nm... so it wouldn't surprise me if they just skipped 10nm for the most part (by limiting it to mobile parts) and then focus their attention on 7nm instead.
Either way, AMD seems to have more 'breathing room' to recapture market and mind share, and hopefully get industry devs to optimize for their CPU's and GPU's from the start.

 

Buying tech with your 'heart' is, on the surface, romantic and altruistic, and makes you use both sides of your brain. But hardly a good business case can ever be made for doing so.

REAL FUTURE-PROOFING:
Always buy the most current platform you can, put in the best CPU and the most RAM possible for that platform from day 1, and then use and enjoy until your needs change and/or your cash flow situation allows you to buy something better (and better = after testing the new potential platform vs. your current one in the same identical workloads and also after testing any future workloads you may need on both the old vs. the new too and making an unbiased decision accordingly..).​

Being noble about the tech company you choose is laughable by the very companies you're being noble to. They just see you as a PIN to your savings account, after all. I suggest you treat them the same to get the most bang for your buck.

Thinking about inconsequential and at this point, theoretical security hacks that won't ever affect normal users, business practices from the long dark past that may also be going on in the company you want to 'reward' with your business too, (but just haven't been brought to light yet) and buying on the promise of 'future-proofing' when stats show otherwise, for years now, is just the kind of wrench in otherwise impartial thinking/decision making that all companies strive for. Because when you're confused, and/or overloaded with information, or otherwise using inputs to the decision that have no bearing on your use, you're most likely to not use your head at that point.

Which is exactly what any company wants you to be at. Make an irrational decision with your heart and defend it to the death. Great for them, huh.

In my case, Intel hasn't always provided the best new path in each new iteration (and I've stated that many times before across the forums here). But using the 'real future-proofing' process above, neither did I 'have' to buy/validate a new platform at those times either.

No one is forcing anyone to buy a new platform at any particular point in time. When the time comes that you decide to do so, decide like a business entity would, and not as a teenager might.

After all, the only one that has to be truly impressed with your new platform is you.

Let the big business figure out how to do that successfully and keep their investors happy too. Along with worrying about positioning, market share and everything else that doesn't directly affect a users experience with one platform or another.

 

https://www.nextplatform.com/2019/05/13/intel-gives-moores-law-a-makeover/

"I ntel is not the only one doing this. AMD famously used their own chiplet packaging and interconnect technology with its EPYC processors. That approach likely enabled AMD to jump to Taiwan Semiconductor Manufacturing Corp’s 7nanometer process node a lot quicker than it would have had Intel gone the monolithic die route. In fact, Renduchintal estimates their approach will enable Intel to accelerate the use of a given process node by up to two years as a result of the kind of flexibility afforded by packet-level integration."

I guess what they're saying is that by going smaller dies with chiplets, they can more easily manufacture new chips on the new nodes, and shave off significant development time for yields due to not using huge monolithic dies anymore. Makes sense and should certainly help them get back on track faster and cheaper.

https://www.pcgamesn.com/intel/10nm-tiger-lake-cpu-sisoft-database-benchmark

2019, 2020 MacBook Due For Intel 'Ice Lake' Boost As Computex Nears

https://www.forbes.com/sites/brooke...10nm-upgrades-as-computex-nears/#7b65cfa4b8d8

 

That is a correct interpretation. By going with smaller chiplets, you increase yields drastically.

As an example, let's look at Intel's 28-core behemoth on the 14nm++ node. With each "+", Intel decreases density, which allows for higher clocked frequencies. Even with that, Intel's 28-core only achieves around 35% yields. Why? Because by going so large with a monolithic die, you increase the chance of a critical defect occurring on the die, which then makes the chip unusable. Meanwhile, if you make a smaller chip, such as a 4 or 8 core, especially if the architecture allows for shutting off cores completely, allowing for you to make it a 6-core chip, or similar, then you can get actual and effective yields up during the manufacturing process.

Now, for AMD, and this is only used as a comparison for what Intel is going to adopt, they were achieving in the mid-80s or higher on yields on GF 14nm, a very mature node (same with 12nm). Now, with moving the I/O and IMC off of the core die, if a critical defect hit any of those, the chip would be useless. Even non-critical defects could slow the memory controller's capabilities or you would have reduced PCIe lanes, etc. So by moving those hard to shrink components off, you reduce the R&D costs to reach a new node, reduce the risks of critical and non-critical defects hitting the die, and reduce the die size considerably, allowing for the core die made on the new node to be smaller so each wafer produces both a higher actual and higher effective yield. This is why AMD is rumored with Zen 2 to be achieving a 70% yield on TSMC's 7nm process. That comes up to, if the 70% is actual yield, approximately 506 good dies per wafer, which each wafer for 7nm (or Intel's 10nm) costs around $11,000. That means AMD is yielding around $22.50 per core die on 7nm! That would be 126 32-core CPUs per wafer! Meanwhile, Intel's 35% yield on a mature wafer that costs around $6000 is a problem. That is why Intel's extreme core count CPUs cost so much.

Now, that doesn't mean Intel's problems are solved by disintegrating the cores into separate dies with chiplets. Intel is rumored to, even with 2 and 4 core chips, have problems with yields on 10nm. They are working on it, but until they show more than 4-core chips, which isn't really even planned for this year as they are primarily using it on mobile "U" and "Y" chips and will have a Xeon server offering that will sit in between the 14nm Cooper lake and the 14nm Cascade lake. This means they are not likely doing a full lineup of 10nm chips for servers, and no plans were shown to suggest Rocket Lake will be on 10nm.

That, however, does not mean they are not working on accelerating 7nm. By all accounts, they may have 7nm up and running for volume in 2021 (protected forward looking statements under securities law). If they have that, theoretically, they may be able to achieve a higher density than TSMC's 5nm, which will be offered in volume in 2020, but that AMD is not expected to adopt until 2021, while 3nm may not be offered by TSMC until 2023 (unless they successfully move 5nm fabs over to 3nm ahead of the new fab they are building for 3nm). Samsung, however, is expected to offer 3nm in 2021. Intel, TSMC, and Samsung are the last standing cutting edge fabs. IBM sold to GF. GF, after not getting AMD 7nm deal mentioned shopping for a buyer months ago. So, that leaves those big three.

If Intel cannot get 7nm to go off without a hitch, then there is a real possibility of there becoming just two large fabs: TSMC and Samsung. Hell of a thing to think about, right?

Now, that is not discussing Intel's architectural engineering, which is awesome. Rather, it is saying that they have major process issues at the moment, which if not resolved, may require them to go fabless.

 

i think some people just got to admit, AMD went and prepare the smart way of doing things from yrs ago now we are seeing the results while intel is currently doing what AMD was doing 3-4 yrs ago, planning for new arch due to market milking they only got themselves to blame.

performance is another topic, intel may be able to do 8 cores at 5.5ghz on some water cooling and may still have a decent lead in gaming. i'll take a 12 or 16 cores at 4.6ghz from AMD instead. the frequency temptation gone way down, by having 2x more cores and a lesser price while having similar IPC imo is worth it.

 

Once again, your assumptions are showing.

If Intel is currently doing this today, they didn't just wake up and said 'lets copy AMD' yesterday. They've been working on it for a very long time too. Agreed that they got a little sidetracked with all that's happened in-between, but they're getting back on track with each passing day.

So, I think that some people just have to admit that the top two will be very close to the other if they want to succeed long term. Nobody is getting game-changing future tech from an alien race yet that they keep hidden in the basement; yeah not even the fruity company could sustain that magical belief for forever...

The hard work still needs to be done and as the best minds keep trading offices between Intel and AMD, the final products will continue to look like copies of the other.

I will still pick the productivity champion, not a cheaper, but ultimately weaker solution, that just doesn't make business sense or otherwise for a one time cost in return for many years of ROI to be gained though.

 

https://www.pcgamesn.com/intel/xeon-sapphire-rapids-granite-rapids-leaked-huawei

A pissed off Huawei is leaking Intel's road map..

10nm 26 Core Ice Lake coming between Q1-Q2 2020 with DDR5 and PCI-E 5.0 coming Q1 2021. NICE!

 

I'll believe it when I see it, and it's allegedly Q2.

Now, I'm wondering since you get 4 working 28-core chips on a mature 14nm++ node, working out to a 35% yield rate, how they will achieve this 26-core part on a process they barely got working for 2 and 4 core low power chips for laptops and mobile?

As to PCIe 5 and DDR5, Zen 3 Epyc server parts are said to have that and launch in Q1. So that's not special. AMD will have beat them to PCIe for consumers and PCIe for servers, considering Cooper lake is releasing Q1 against the rumored Zen 3. And consumer chips don't get PCIe 4.0 until someone next year.

So, welcome to the party on something that AMD was rumored to have during the same time frame. Except AMD seems to be making good on their release. Intel's recent record....

Sent from my SM-G900P using Tapatalk

 

yep, assumption for sure, though more likely to be real AND better than anything intel have and will offer in the next yr or two for sure. 10 cores 14nm++ will run damn hot for sho, though 10 cores at 5.1ghz is impressive, im over it because my laptop can't use it, it can barely even do 8 core with normal fan profile.

@ajc9988 i have little doubt intel wont go full out given how much they know they're behind at least for a yr or two. we know 10nm is definitely coming and server is going to get the best and first of it because its not frequency dependent and unlike mainstream care about more performance, first gen 10nm wont have those high clocks which will be just fine.

those 10nm yields are low anyway, hopefully they can hit it in the 20-30% LOL but i seriously doubt it especially with 28 cores, on a brand new node that suppose to come out 4 yrs ago.

but yes this is probably another PR stunt if charlie was to see it to keep the share prices up and give future hope, the similar stunt they pulled with lenovo with 10nm laptop and that was it, nothing else afterwards. they want people ot keep hope and not buy into AMD's product, some will probably sitll be swayed by intel's bs, fanboism runs deep.

 

You have to act upon known knowledge, not on speculation of the unknown. To do so is irrational.

Future proof, purchase a CPU and board that can not just drop in an upgraded CPU, that is irrational as well.

 

https://www.engadget.com/2019/05/26/intel-10nm-ice-lake-cpu-gaming/

Intel showing off their 25w TDP Ice-Lake-U with 11th gen graphics. Faster than the latest Ryzen 25w APUs in both CPU and graphics. The Ryzen APU was great for it's graphical performance, but now Intel has single core, multi core, and graphics performance ahead above Ryzen. 10nm shaping up to be promising fast. I am looking forward to an Ultra book from Intel on 10nm soon.

 

As someone else pointed out:

"

Not bad, not bad.

However, the specs were wildly different in regard to RAM speed:

* Ice Lake-U LPDDR4X-3733
* Intel 8565U 16GB DDR4-2400
* AMD 8GB DDR4-2400 (single channel? really Intel?)

And also there's a strange mix of low/med settings as if Intel tried to cherry-pick the titles to show the best performance [gains].

Absolute most consumer laptops won't be running DDR4-3733 RAM which is hugely expensive."

Sent from my SM-G900P using Tapatalk

 

It could be Huawei is counting on Intel failing to deliver a 10nm product on a partner roadmap once again, showing how Intel isn't trustworthy.

As suggested by the trustworthy Huawei -> that just leaked proprietary NDA'd technical data of a trusted partner.

It is interesting how Intel is showing different roadmaps to different partners, with desktop CPU's left out of some while showing them to others.

I hope Intel really does deliver 10nm deskop CPU's and laptop H CPU's in production quantities by 2022.

 

Intel’s 10nm Ice Lake CPUs Achieves 18% IPC Improvement Over Skylake
https://wccftech.com/intel-10nm-ice-lake-18-pc-ipc/

 

It's nice to see progress in Intel's 10nm production asperations, as an upgrade to the year ago short run of 10nm CPU's, this time there are more cores and a working iGPU:

"As far as the raw specs go, the preliminary lineup of ICL (Ice Lake) CPUs are going to be 9W, 15W and 28W designs for the mobility market (as per usual) and will feature clock rates up to 4.1 GHz boost.

Cores will max out at 4 with hyper threading enabled (Ice Lake has hardware level mitigation for spectre and family of attacks). The lineup will be broken into the familiar Core i3, i5 and i7 segments and graphics will be clocked at 1.1 GHz.

DDR 3200 is now supported by default without the need for overclocking of the memory controller. L2 cache size is up to 8MB."

It's all BGA, no LGA parts this year (or next):

"The packaging is broken into two main categories: 15W and 9W. The 15W Type 3 package will have 1526 balls with a pitch of 0.65mm while as the 9W Type 4 package will have 1377 balls across a pitch of 0.43 mm.

To the uninitiated, the balls in a BGA packaging are the same thing as pins in the standard LGA socket.

Both types feature a new thin magnetic inductor array."

So there won't be an enthusiast part, no upgrades or even mixing motherboards - you'll likely need to buy a built laptop - maybe a built desktop, less likely a motherboard with the CPU pre-soldered on.

There doesn't seem to be detail for the 28w part, at least not in this article. Perhaps that will follow later next year? Or it's a power upgraded Type 4 BGA?

As Intel said, as mentioned in the article, this is the first real upgrade in performance - even in the liliputing sizes it will come in at first 9w / 15w since 2015.

That's 4-5 years later when in full production availability...and Intel is still 1-2(-3?) years away from delivery of the higher power full performance even matching today's highest performance Intel CPU's.

I hope to see Intel beat it's 10nm efforts with better performing 7nm CPU's before 2-3 years are up.

 

Considering the 6700K and 9900K have almost exactly the same IPC this is going to be a huge step forward for Intel when they release the high power desktop product. Exciting times.

Last year people said 10nm was dead or cancelled. The dual core no iGPU was a failure. Next month they are bringing a 4 core 8 thread laptop with a faster iGPU than even AMD Vega graphics laptop APUs. Looks like Intel is going to show AMD how an APU is done right.

 

AMD come with 15% IPC increase for their new chips. Around same also for Intel. Not much will change between them if not AMD continue increase it next year.

Intel Unveils 10th-Gen Core Processors, 10nm Ice Lake, 18% IPC Improvement, Sunny Cove Cores, Gen11 Graphics, Thunderbolt 3
https://www.tomshardware.com/news/i...ics-sunny-cove-thunderbolt-3-usb-c,39477.html

 

The problem here is 2-3 years for high end desktop CPU's gives AMD 2-3 iteration upgrades to the 7nm process if not even encroaching on 5nm.

 

But, the Intel 10nm parts are being delivered only as liliputing CPU's, low power only - 9w / 15w, low clocks, no high core count high frequency high power desktop 9900k level or HEDT level performance Intel 10nm CPU's, 10nm CPU's are low power laptops / desktops only.

The 15% IPC from Ryzen 3 (Zen 2) is delivered across a wide range of high performance desktop, data center, 65w -> 105w maybe more in the 16c/32t Ryzen 3 CPU's, and perhaps even higher core count + higher power Threadripper and Epyc CPU's.

There's really no comparison between AMD 7nm CPU's and Intel's 10nm CPU's until Intel delivers full power 10nm desktop / data center in the same performance range as today's 14nm CPU's - and hopefully in an even higher performance range than current desktop / data center CPU's. That comparison is still years away from being possible.

And so far, after years of Intel repeatedly missing delivery on 10nm desktop / data center CPU progress, Intel's guesstimates are just that - very rough estimates.

Until Intel delivers to the market full production of 10nm CPU's for desktop / data center in the same - and higher quantity - as 14nm is shipping today, we can't count on Intel's full performance desktop / data center 10nm delivery dates.

 

The 10nm desktop / data center full performance CPU products as promised for years by Intel are dead - they were never delivered.

Intel aren't even publicly promising 10nm delivery dates for full power desktop CPU's any longer - instead 14nm CPU's for those tiers of performance are planned out for the next 2 years.

Intel is now pressing forward with 7nm and given that the time of 10nm has passed and 7nm should be taking over in Intel's road map, no one would be surprised if Intel never delivered 10nm desktop CPU's.

I'm not saying that 10nm desktop is not possible, or even that 10nm desktop is not still privately planned, I'm saying that if Intel's 7nm process succeeds concurrent with 10nm progress, and Intel's 7nm process is as good as performance, yield, power, etc as the 10nm process - or 7nm is better in any category - Intel would pick the 7nm process to invest in and grown long term over choosing to proceed with the tattered history of the Intel 10nm process.

Keeping 10nm alive through low power low production products - favorable to yield problems - until either 10nm or 7nm desktop CPU's are viable makes some sense, but it isn't a strong move for Intel.

 

Not quite. Around Kaby to Coffee, due to changes in the cache sizes, there was an appreciable (more than variance) increase in IPC, even if limited. Cache changes will effect IPC, are hard to separate from IPC, and should be noted.

Moreover, I noticed that the Intel graph DOES NOT HAVE an X-axis label on their infographic. THAT is a problem. Take a look for yourself:



So, SPEC 2016 & 2017 were used, Sysmark 2014 SE, WEBXPRT, and CB15 were used in the IPC slide (which doesn't separate them out nor have any label for the X-axis to understand what the increase is relative to in the slide). For the Core Performance slide, they used SPECInt 2006. All of these were internally compiled, which Intel has been caught before using flags in compilers that would NOT be allowed in a commercially released product (not saying that occurred here, just saying it has occurred before).

Have you seen what sort of scores I can reach on first gen Ryzen 1950X when using Linux and the -O3 flag? Night and day versus regular compiling without platform optimizations. Intel specifically claims that they were compiled for the specific platform and that using it on any other platform could materially change the results. That sounds like hyper-optimization.

But, aside from those things, look at the benchmarks used. Are any of those your "go to" on benchmarks to estimate performance? CB15 gets the closest, yet then you run into the same criticism that is hoisted upon AMD.

Let's dive into the impact on mitigations. Start with Sysmark 2014 SE by BAPco.


https://www.anandtech.com/show/1365...ith-spectre-and-meltdown-hardware-mitigations

As you can see, there is a 6% drop in performance before being patched. Now the article shows lower numbers with the overall categories, and it doesn't give us the full breakdown by test, but it should be remembered that the impact is seen more on elevated requests for I/O processing, which is not going to impact consumer programs as much, and which the bulk of Anandtech's testing takes place on.

When you look at the SPEC 2016 and 2017, then a custom SPECInt 2006, and it seems like some gaming is afoot. From the footnotes, Intel specifically claims it may or may not have included all software mitigations at the time of testing. That means, specifically, some of the mitigations ARE in place. Now, there is a question on when the tests were conducted. The public release of patches for Zombieload, etc., didn't arrive until May 11th. That means those may or may not be included. We don't know what all microcode was applied to the platform, what was turned on in OS mitigations, etc. Hell, we don't even know for certain whether it is comparing HT on and off, although I will assume for Intel's sake it was on. We don't know if the Portsmash fix was incorporated either.

So, depending on the state of the vulnerability mitigations, much of that IPC increase truly could be mitigations and clawing back lost performance. We simply do not know.

What we can say for sure is the 40% raw performance number comes from comparing a skylake-u sample to an icelake-U sample, which comes with more advanced boosting and higher clock speeds which is mixed with the IPC gains and with the software mitigation sandbags. Plus, that was ONLY showed with SPECInt 2006, which is curious why they wouldn't do a multi-benchmark analysis for overall performance, instead relying on an old integer benchmark with new CPU optimizations in compiling.

As for the claims on the laptop iGPU, those are NOT accurate. Intel's Ice Lake-U officially supports LPDDR4 3200. They used 3733, then they stuck AMD's APU with DDR4 2400, which means they are using 55% less bandwidth for the comparison.

They also OVERCLOCKED the Ice Lake-U CPU to 25W, meaning the performance shown is not what you get, then tried to justify the 2400MHz as what you will commercially buy, meaning that they are using dual standards for the testing of the device (not surprising with Ryan Shrout). If you wanted to make the fight about the performance of the actual tested component, you have to standardize the other variables. Without that, you are not properly comparing ANYTHING.

Please see my above analysis for why 18% IPC may be fabricated. But, with that said, let's dive into changes they made to the uArch that could actually achieve a sizeable IPC uplift, which would make the claim real.



https://wccftech.com/intel-10nm-ice-lake-18-pc-ipc/

Intel has increased the L1 data cache by 50%, they doubled the L2 cache to 512KB, increased microop cache, the Out of Order Execution window has been expanded, etc. Now, for the cache, this can lower cache and memory latencies further, which can boost IPC. The OoO window can allow for the data to not be retired as quickly with the speculative execution mitigations in place, which would save on the stale data problem, helping to bring that back up in performance.

They also increased the width of allocation, expanded the execution ports by 25%, doubled the L1 store bandwidth, added an extra AGU (33% more), and doubled the store data. If you look, much of this can help with the mitigations applied for security vulnerabilities when combined with above features.

But, putting security aside for a moment, all of the above can improve the IPC performance of a chip on certain workloads. It isn't a question of if they improved the IPC, rather it is a question of by how much they increased it.

Intel never sat still while not using the uArch Sunny Cove. They have been refining it. So I'm waiting for better third party reviews to find out the actual amount.

Definitely.

 

I thought the 10nm Ice Lake CPU's were releasing in June? The Dell preview from Liliputing below saying delivery of the i3 (10nm?) variant of the laptop is in the summer, and the other 10nm overview from HardwareCanucks below says that the 10nm Ice Lake CPU isn't available in notebooks until October / November - and they had just visited Intel to get the latest details a week before Computex 2019. So is the 10nm release a staged rollout drawn out over the 2H rather than all at once in June, or are 10nm Ice Lake CPU's all delayed until October / November?

Here's a couple of previews of Ice Lake 10nm:

Intel Ice Lake Processors Explained - 10nm Arrives!
HardwareCanucks
Published on May 27, 2019
Intel is finally ready to talk about their Ice Lake 10th generation Core processors, Sunny Cove CPU architecture and 11th Gen graphics. While they'll only be available on U-series and Y-series notebooks for now, there's a lot of cool technology baked into their designs. Could these processors create the best slim and light notebooks of 2019? Maybe....

With the much lower boost clock maximum from Ice Lake 10nm 4.1ghz vs Whiskey Lake 14nm 4.8ghz, the purported 18% IPC boost only gives about a 6% performance improvement - as seen in the Intel charts in the HardwareCanucks video. Also the iGPU improvements compared to the previous generation iGPU @ 1.8x don't map high enough to catch up to the Ryzen 2200G or 2400G in gaming results.

This Dell laptop comes with a range of CPU's, with the Core i3 10nm Ice Lake $999...no price or delivery for higher spec models is given, maybe those will be in October / November?

First look: Dell XPS 13 2-in-1 with Intel Ice Lake
Liliputing
Published on May 28, 2019
Article: https://liliputing.com/2019/05/dell...1-with-intel-ice-lake-cpu-bigger-display.html

If the previous generation had Y processors, Intel 10nm 9w-15w-28w Ice Lake CPU's might over power the cooling capacity in the higher i7 configuration - maybe even in the i5 configuration, only time will tell. I liked the valiant attempt by Dell to redesign the intakes / exhaust vents to allow more cooling, perhaps it will be enough?

 

That is unclear. The 9900KS is supposed to release by the holidays of this year. The Ice Lake-U is shipping to partners now, but I am betting back to school systems (possibly in limited supply), followed by full supply by the holidays. For Cascade-X, they said fall 2019. That likely will be Sept-Nov 2019, similar to the 7000 series HEDT chips and the 9000 series HEDT chips.

So, that makes me think 3 months minimum for deliverable Intel products announced at computex. Makes the 1 month wait for Zen 2 seem much better, even if just a paper launch (nothing wrong with those, but always nice if you can drop hardware at the same time as the announcement).

But we also haven't heard anything on Comet Lake. That 10-core may be pushed to CES release, unless they are doing the 9900KS WITH the Comet lake release (but then you would think it would have been discussed or mentioned like the other three discussed above).

Intel also may figure that the 16-core is being held back for that, so they are trying to create distance between the 9900KS drop (which is NOT what AMD was hoping to respond to) and the 10-core Comet, hoping that AMD will not wait for their 10-core and will respond to the 8-core so that they can seem to respond to it with the 10-core but tweaked to be more competitive. Who knows?

 

AMD could drop ThreadRipper 3 in response to the Intel 10 Core CPU and blow away all of Intel HEDT + the 10 Core. I hope AMD drops the 16 core when it's ready, not before.

There are plenty of new goodies what with the R5, R7, R9 already announced. Those Ryzen 3 CPU's + Navi 5000 GPU's would all make nice builds for Xmas.

 

Yes, but I have questions surrounding TR3. Specifically, historically, they had 5 months between Zen and TR1, then 4 months between Zen+ and TR2, but there is a chance of 6+ months between Zen 2 and TR3.

All sorts of rumors and postulates have been proposed, including TR3 skipping directly to Zen 3 chips and dropping with Epyc with Zen 3. Yes it can win in performance, but there is the question of when it will drop.

Either way, Intel is showing they are trying to compete. More interesting is where they set the IPM for what "thermal" limit. From what I read, it reaches the thermal limit, then drops the frequency by 200MHz for the all core speed. This is able to be used with existing 9 series chips. So I'd like to see what people wind up with using that utility (which also bundles other utilities like XTU and the graphics command center).

TR3 will come and be a show of force. AMD may even be waiting for the unveil to troll the Cascade-X release, similar to Intel trying to steal AMD's thunder on their keynote by releasing information on the 9900KS, which won't actually be available until the holidays. If being honest, Intel has done that to EVERY AMD major event since the release of Ryzen. Hell, there is an argument that Intel dropped the Z270 when they did to force a squeeze on BIOS devo resources for the first Ryzen release, all as the Chinese New Year was upon them, knowing that it would cause the MB mfrs to blame AMD, which AMD took it in stride. Couple years later, the MBs are finally up to Intel's quality (some ahead plus also containing PCIe 4.0), the partners are seeming almost brow beat in regards to AMD, looking extremely excited to work with them, but almost like they have been chewed out and screamed at before reaching that point being on stage (like the Microsoft apology from CES during the presentation, I believe it was).

None of the board partners are happy with AMD in regards to their pushing on PCIe 4.0 this year, though. Primarily it is due to signal integrity and actually having to engineer things, especially ahead of Intel's adoption. But Intel pays for a lot of R&D and ecosystem advancements, something AMD's budget doesn't allow for in the same way (see the Athena project information Intel just disclosed). That means that it is more up front cost to them, but it will pay itself off in the long run.

 

Found out a bit more about Ice Lake delivery dates:

"Dell says the new Dell XPS 13 devices will be available starting in July, starting at $1,000. However systems featuring 10th-generation Ice Lake CPUs aren't expected until this holiday season..."
https://www.zdnet.com/article/micro...er-windows-10-edition-home-ultra-to-its-list/

As far as PCIE 4.0 I think that is all that is needed for a while. It will take time to fill that bandwidth. So that means AMD has headed off PCIE 5.0, which while providing more bandwidth probably is overkill for a few years. Getting a jump on Intel and making it harder for Intel to push PCIE 5.0 as a selling point.

It will be fun to see if the new 5000MB/sec PCIE 4.0 M.2 drives make enough of a difference in every day use - a 10x leap from SATA III M.2/2.5" (500MB/sec) to NvME PCIE 4.0 M.2 (5000MB/sec) is as large a leap as from HDD to the first SSD's, so it might be a real selling point for AMD migration.

 

Even on the Huawei leaked roadmap, PCIe 5.0 is ONLY planned for servers. It's my understanding that PCIe 5.0 IS NOT coming to consumer silicon anytime in the near future.

Intel is adopting PCIe 4.0 for the mainstream desktop next year. There is no information on if their HEDT platform will get PCIe 5.0 or not, but there is no information on if AMD's HEDT platform will get PCIe 5.0 or not either. So AMD gets to claim the win on getting to PCIe 4.0 first. Still an open question on who will get DDR5 first for consumers and for servers. From the roadmaps, Intel is expecting Ice Lake-SP to have DDR5 and PCIe 5.0 in the second quarter of next year. From the rumor mill, AMD will be announcing Zen 3 at CES or around there, which will also feature DDR5 and PCIe 5.0. But announcement does NOT mean release. So there are open questions in that regard.

Now, what I would hope is TR3 gets moved from the Ryzen branding to the Epyc branding, then you merge the speed optimized Epyc 1P systems into the Threadripper offerings, all while allowing for RDIMM, UDIMM, and regular ram, full overclocking capabilities, then just separate the MBs into OC/Enthusiast/Gaming and general 1P server boards that may not need all the extra VRM capacity, etc. But that is just me.

But, if that is correct on not having Ice Lake-U shipping until holiday, it really cuts against the recent Intel statements on 10nm.

 

Well, I didn't want to say it, but that's what I am expecting for Threadripper 3 + Epyc will both get DDR5 + PCIE 5.0, but I didn't want to get anyone's hopes up, we still don't know for sure, but it makes sense that Epyc would do this for sure and ThreadRipper feeds off of Epyc implementations - at least so far, time will tell.

 

Having seen both AMD and Intel's presentations, Intel has nothing to worry about if it delivers even half of what they showed.

Intel has an actual vision and it is one of their greatest assets. They stress real-world results as I do. They haven't wavered too far from that ideal, from what I can see. And they bring actual innovation too, that one can use today, and is usable by the masses.

While others seem to be weighing multiple options back and forth and coming back to how superior AMD will be one day, Intel just keeps chugging along as the best option for those masses. Just look at what they've introduced, AMD doesn't have anything close to it.

The often-delayed Intel 10nm products that are shipping now make anything mobile from AMD seem antique already (as they were already behind what Intel has already been offering for a while now). Intel seems to have made a very wise decision to not introduce something on a larger scale that wasn't actually 'more', two years ago.

More cores are just more cores. Yawn. Intel points us to how their tech can actually enhance what we're doing today.

In the end, Intel and AMD are both in great places today. Regardless of all the analysis being done around the web always bashing one and praising the other.

For my $$$$$$$ though? AMD still needs to go through a few more iterations of their recent successes to be an actual option for me to consider. 2019 isn't that year and from what I can see today, neither will 2020.

Lenovo + Intel + Project Athena is the next 'digital' notebook I'll be considering in the near future. Assuming I can get it in a ThinkPad.

The actual hardware, node used or PCIe interface inside a platform is immaterial if it provides a great experience.

 

Stop with the PR BS from Intel. You are factually incorrect.

Highest performance only matters to a SMALL group that are willing to pay ANY price for it, and Intel sure as hell is bilking/milking them for it! That is why in the DIY market, AMD has flipped the script in desktop and according to those that publish their figures, AMD is nearly 2/3rds of all new sales. Intel makes some of that back in margins, but that was before Zen 2.

Meanwhile, stop pretending that Intel had a choice on 10nm. They did NOT. EVERYONE KNOWS THEY DID NOT! That is why Intel's Comet Lake-S is 14nm, and why Rocket Lake-S, although not marked as 10nm or 14nm (except on mobile where it was marked as 14nm and 10nm iGP), it shows constraints. Further, finding out volume laptops with Ice Lake-U will not be present likely until this Thanksgiving/Christmas shoots the notion Intel REALLY was going to flood mobile with these chips. Because of Intel's shortage, mobile started adopting ARM chips and put out more AMD lines, including moving AMD devices into the mid-range and high-end offerings. That is a threat that cannot be ignored after AMD decimated Intel's HEDT sales with Zen 1, then took a chunk out of desktop with Zen+. Intel's shortage not only hurt them, but everyone in the chain for laptops, including the touchpad, displays, wifi, etc. This is why those companies WERE FORCED to consider AMD for laptops, because Intel couldn't deliver. And now looking at Ice Lake-U quad cores and the expected volume delivery date, it seems they were right to look to other than Intel!

Then you ignore the change in IPC, meanwhile ignoring that on desktop and HEDT, there is no Sunny Cove or IPC change, and the suggestion of Sunny Cove being in Comet lake is an unsubstantiated rumor, meaning that the 10-core Intel has likely will NOT have the IPC increases that Ice Lake-U does. Further, Ice Lake will be battling Zen 3, which will have another IPC uplift, not Zen 2, and for servers doesn't land until next year. That sounds like a whole lot of incremental, not revolutionary, changes from Intel.

No one cares about YOUR money. Market has spoken, something you ignore TOO much! Instead, you seem to have your head firmly planted in the sand, not examining the changing surroundings, something the Art of War should have taught you.

Project Athena IS A BIG NOTHING BURGER. It is Intel partnering to help develop a standard for the industry to try to squeeze AMD into complying with their standards, or rather squeeze AMD back out of the market share they captured on mobile. Anyone not blind can see it is pure strategy, not innovation. They did the same thing by giving the patent for an interconnect to DARPA, trying to prevent AMD's offer of NoC innovation for chiplets by creating a routing data standard. If the industry adopted the interconnect from Intel, it would have set back AMD's efforts in chiplets and given Intel an advantage. Funny enough, Intel and Nvidia lost the contract for Frontier and Shasta. Instead, they got a DoE supercomputer contract acting as a subsidy after going and begging Trump for socialism to get that subsidy. Hilarious! Talk about innovation!

And you saying those are immaterial is a farce. If you need the extra bandwidth, it is NOT immaterial. PCIe 5.0 is necessary to scale out our networking infrastructure in ways few could dream of before now.

Now, I sing both sides praises and attack both sides regularly. And when you are not pushing this pish posh PR nonsense, and I'm not responding to it, I clearly state both the benefits and drawbacks to each. What I do not stand for, usually, is you making these nonsense posts!

 

Highest performance matters to anyone using a 'smart' watch from a 3-year-old to anyone else that uses tech for work or pleasure. The price they pay is the price they deem appropriate for the level of performance delivered. Not on your definition of being 'bilked'.

Yeah, everyone has a choice and even Intel too with releasing 10nm, or not, before they deemed it was ready. I didn't say they didn't have setbacks. Many examples of AMD having bigger setbacks and yet they still launched an inferior product.

You keep counting IPC in the processor as the holy grail, well, it's not. It is just one small part, of which Intel platforms have plenty more of, including their newest Project Athena. You can dismiss it if you want, but it is exactly what I've been looking for and for a very long time too.

Wow, how silly of them! Ditching 'scores' that mean absolutely zilch to me and many others and actually focusing on how their platforms will actually advance productivity instead.

I have never stated anything like what you attribute to me below. Quite the opposite. I always recommend going with the latest platforms and then max out those platforms with as much CPU+RAM as they can take and you can afford.

I am not paid by Intel, nor do I benefit by calmly and objectively stating my views. The one who is factually incorrect here is not me.

And stating that AMD is nearly 2/3rds of all new sales is actually very humorous if you didn't seem to believe it so much.

I don't see you singing Intel's praises when they're rightfully deserved (and right now, they are). But you sure get annoyed when someone else points them out.

 

I expect we all will see the fruit from the 2/3rd of all new sales early next year

 

Yes, and that is why we ONLY have flagship cell phones and that the majority of volume doesn't come from lower in the product stack! /s

You are absolutely absurd, as Intel lost the IPC against Zen 2 with any products other than a quad core Ice Lake-U. That sounds super impressive Intel!

I WOULD NOT BUY AN INTEL ICE LAKE-U QUAD CORE EVER! IT IS NOT WHAT I'VE BEEN LOOKING FOR AND FOR A VERY LONG TIME!!!

Literally, if I wanted something that performs but is in a mobile device or tablet, I'd go ARM, not Intel. If I was buying a laptop to replace my P770ZM, I'd get either another Intel DTR or an AMD DTR (waiting to find support for Zen 2 chips and either AMD allowing Nvidia cards to be used or putting out a decent GPU to pair with it). That means that NO, I AM NOT IN THE MARKET FOR ICE LAKE-U EVER! You literally are absurd, trying to tell me what I want. You should know better by now! Fillip!

So, since you don't understand that IPC*frequency (or number of cycles completed in a given period) gives you an estimate of performance, and that IPC plays as big a role as frequency, I have nothing more to say to you on that. It shows a lack of understanding of computing. So stick to what you know.

Also, this talk of platforms and productivity is PR speak, not regular person speak. Especially when you try to downplay the importance of IPC in performance before using those buzzwords.

And recommending the latest platform and maxing out on the most expensive chip is BAD ADVICE for many people. It ignores what the various use cases for chips lower in the stack are and why they exist! It shows a lack of nuance and recommending spending more on products without the benefit to many general consumers. Even IT people don't only buy (edit: only) Intel's 28-core Xeons. No, you buy for your need and in your budget. So in other words you give very bad advice.

On the 2/3rds of sales, I clearly stated the DIY market (not OEM builds), and clearly stated the source as Mind Factory in Germany, the largest online retailer in the EU, as well as looking at the partner link data from GN and others. Notice how your stripping my statement of a very integral piece of information, and of the sources, was to make the statement false and to mislead other readers. That is something you do regularly. So, as to shilling, it makes you suspect.

And you sure ignore many of my statements then. Just because I sing their praise and spank their butt in the same statements does not mean I don't sing their praises. In fact I've recently given tons of compliments to their uArch engineers, even as I crap on the process issues they are suffering. But that also doesn't mean, even with that respect, that I do not have qualms with their offerings, strategies, etc.