New Generation Intel CPU's 'Tiger Lake' Processors

Intel Introduces 'Tiger Lake' Processors


Intel Teases Tiger Lake And Demos DG1 Dedicated Graphics Card In Action

 

Intel confirms 'Tiger Lake' is the next Intel Core processor you need to care about
https://www.pcworld.com/article/351...el-core-processor-you-need-to-care-about.html

 

Intel's Tiger Lake Laptop CPU Brings Thunderbolt 4, AI Graphics Processing
https://www.pcmag.com/news/372882/intels-tiger-lake-laptop-cpu-brings-thunderbolt-4-ai-graph

 

Intel at CES 2020: 45W 10th Gen Mobile CPUs Soon, Tiger Lake with Xe Graphics Later
https://www.anandtech.com/show/1531...e-cpus-soon-tiger-lake-with-xe-graphics-later

 

Intel Tiger Lake-U 4 Core CPU With 4.3 GHz Boost Clock Tested – 15W Variant Up To 32% Faster Than Ice Lake & 28W Variant Up To 62% Faster
https://wccftech.com/intel-tiger-lake-u-15w-4-core-cpu-performance-benchmarks-specs-leak/

 

I especially like how Intel's recent marketing in desktops is desperately trying to diminish the relevance of synthetic workloads (that they lose at horribly to Ryzen such as Cinebench) in favour of "real world performance" ... and then use entirely synthetic benchmark numbers to support their peak marketing claims of performance benefit with their own CPUs

 

I especially liked the one slide where they have an asterisk about performance tested without security patches

Spoiler

 

I get why nobody wants to look like Intel's friend right now. But this isn't a popularity contest and hopefully, most of us are not giddy school girls either. I buy tools with my $$$$$$$$, not try to make/impress friends.

I'll reserve judgment until we get new hardware to play with. But I am liking the new direction Intel has been heading for the last while (a few years, actually).

These new mobile solutions are looking great right now. From both sides. Let's see the reviews already!

Intel's offers are still more compelling to me. Maybe just because they are still so far ahead of what AMD's best can do right now.

While AMD may be catching up performance-wise after all this time (mobile), what they don't have is the vision that Intel has to actually do anything new and different that Intel seems to be able to pull off as required.

They're still piggybacking on Intel's coattails, IMO. They may even surpass Intel, performance-wise. But productivity (a car race, winning a war, or getting to space) isn't about the one who has the most (expensive/powerful) hardware. It is about the one that has the clarity of vision to get to the goal line with what they do have.

Computers have always been mere tools to me ever since I passed puberty with productivity as my goal.

Raw performance (synthetics) is always trumped by real-world productivity. The platform as a whole matter here. Not merely the NOS that some systems can breathe impressively.

For the weakminded here, I'll repeat it robotically for them: yeah, a few very specific workloads can and will take advantage of the multicore advantage that AMD offers. For the rest of the world, Intel's solutions are now not only more relevant, but they're also cheaper too.

With the shenanigans that AMD has been pulling lately, I can't see me giving them any purchases or 'moral' support. Even if I can thank them for the riches that Intel has begun to offer us too.

 

CB has always been AMD’s trump card since the very first Zen reveals, and SMT is a bigger perf improvement than HT, but Intel “losing horribly” to Ryzen on comparable core/thread count CPUs in CB is news to me.

 

Tigerlake looks pretty good. The Xe GPU seems to be a significant perf/area gain over the Gen 11 despite being on the same process. 20-30% larger GPU(my estimates) that pack 50% more EUs but performing better than that while at the same TDP.

My guess is based on their earlier claims, the biggest gains will be seen in the fanless -Y space. I expect it'll be a start to move -Y and fanless to mainstream just as Haswell moved 35W down to 15W.

The comparison is against 9750H, which is a 6 core part. It's still an impressive 46% gain, beating the theoretical 33% core count difference. A combination of clocks, perf/clock and better SMT scaling advantage. Intel needed to move to 10nm cores 2 years ago.

 

This is finally here. In the meantime, AMD didn't really rule anything on the notebook side of the fence either.

See:


See:
https://newsroom.intel.com/press-kits/11th-gen-launch/


I didn't know I was talking about RUGS* so long ago (since I've joined NBR, actually). But that is exactly what Intel is talking about. Let's forget about mere BM 'scores' and talk about actual increases in productivity (however you happen to define being 'productive'). See around the 1:05 mark for a few minutes in the video above for the talk about RUGS.

Synthetic BM/'Scores' is an abysmal representation of relative platform benefits. Rather, I have always stated my workflow (which is blueprinted, repeatable, and consistent) is my basis of whether a new platform has any possible merit of being purchased or not, depending on what time savings I see on my critical workload/workflows.

AMD has not been on my radar for decades for the vast majority of my workstations of this core workflow/workloads I profit daily from. The last few years from AMD hitting on some cylinders (for my workflows) haven't changed that either. 'More cores' are not enough. Smarter cores are simply better for real-world workflows that most people use.

On the notebook side, it looks like Tiger Lake is another reason to continue buying Intel this coming holiday season, not that there were many reasons not to before. Nothing has changed here. I just know the people I helped get a new notebook 'now!' will be regretting not waiting as I told them too.

Once again, I thank AMD deeply and give them a COVID-19 safe virtual hug for all the goodness they've inspired in Intel. Even without buying a single AMD product (that I've kept in the long-term), they have proven to be beneficial as I also predicted they would be almost 4 years ago.

What AMD sorely lacks and what has kept me from buying/keeping any of their recent examples is what Intel excels at: breadth of implementation, imagination/guts to give a true benefit to actual users vs. mere marketing yammering ('moar cores' is not more performance) and the connections to make it happen at the level and scale it needs to.

When the Intel desktop platforms are introduced in due time with similar/better than Tiger Lake specs, the competitions' 16/32 cores will be the 3D TV from yesteryear that made people spend $$$$ but benefited them little in return.


*Representative Usage Guides

 

If you believe what is shown on a marketing slide, especially since Principled Technologies is rearing its head again, I have some beachfront property in Arizona you might be interested in buying.

 

This is what I believe in.

Beachfront property, anywhere? Never had any notions for those fleeting pursuits.

 

Why your next Tiger Lake laptop likely won't perform as well as all those Intel slides claim notebookcheck.net

Intel should use real-world retail laptops to represent their Core i7-1165G7 benchmarks instead of in-house kits that will likely never make it onto store shelves. Their initial benchmark scores don't mean very much if the average Core i7-1165G7 can potentially run much slower on final OEM designs...

Main reasons for this...
How Dell cripple performance explained by.........

 

That explains a lot.

 

Yeah. Explains why I'm not gullible.

 

Of course I could be wrong but I feel like Intel’s x86 glory days are slowly coming to an end and I’m not so sure about the so called guts and vision. Intel definitely is recovering from the Skylake stagnancy period to push back against AMD, granted despite this they still had higher sales and increased demand even to the point of leading to shortages. In the long run however with ARM becoming more and more viable I can see various vendors cutting out AMD/Intel and going vertical for certain situations and ARM allowing room for more competition with others coming into the chip business. I can see others following Apple in tow. I mean I’m not saying x86 is going to disappear overnight or anything but Intel is not going to have it as easy as it has till now. I personally would be glad to see x86 slowly get driven out by ARM implementations, obviously it will take time due to all the legacy software and resulting compatibility issues.

A former Intel engineer François Piednoël gave some insight into Apple’s own frustration with Intel in terms of buggines at the time of Skylake’s development and release which he suspects probably accelerated Apple’s in house ARM development efforts . It wasn’t just Apple but other vendors too that had various issues, ie high MS Surface return rates due to Skylake issues. So when someone in this forum talked about getting reliability with Intel vs AMD this isn’t necessarily always the case.

Also the mess with Intel’s “Technology, Systems Architecture and Client Group” and the refusal to deviate from certain set ideas at their node development team and lack of willingness compromise in certain decisions/aspects led to worsening of the node delays, at least they are cleaning house now in that aspect in terms of leadership.

 

Should the presentation had done...in a competitive and gentleman manner,
I would be interested. However, naming, shaming, and "cheating"???


...now i'm a little turned-off...​

 

Remains to be seen if ( if) the Tiger Lake presentation's claims can be matched by others, of course. But saying Intel's 'glory days' are past and not liking a presentation that compares directly against the competition (what else are they supposed to talk about)? Okay... But where is the shaming? Where is the 'cheating'? Apple? Arm? Are we serious? They aren't even in the game.

In that light, AMD's glory days haven't arrived yet (for notebooks) either then. Let alone the outright misrepresentations that 'moar cores' is all most anyone needs to feel like a champion behind the keyboard.

Actual productivity increases are (or should be) a pre-requisite when a new platform is being considered. AMD hasn't delivered that, over what Intel platforms have had available since 2010 or longer, yet.

But yeah, go ahead and bash Intel for not just stepping up but surpassing (sure on paper, for now) expectations. Buy the AMD products that make you feel good. Just don't think the platforms are in any way equivalent, let alone think that AMD is superior by what they can offer us today or soon(ish).

 

I never commented nor even referred to the slides (You’re literally just putting words in my mouth there), I simply pointed the long term aspect of the CPU industry, that x86 alternatives may eventually become more viable in general and a threat to both AMD and Intel. That along with the internal politics and issues within TSACG were my basis for the “glory days are slowly coming to and end” comment.

I buy what suits my needs regardless of brand. “Productivity” differs by use case, what suits you may not fit others. Currently all my computing products except my pfSense/FreeNAS AMD Epyc box are Intel. I buy their ethernet adapters as well because they’re good.

Putting us niche people out of consideration the average home user/gamer or even office worker would probably be better off buying which ever equivalent desktop product from either company is cheaper as they’d probably not notice a difference. On the mobility side Intel still has the edge for “idle” power draw last I saw, unless things changed recently.

 

Honestly I am only interested in their implementation of TME, and if this CPU architecture is hardened against side-channel vulnerabilities in any way. I doubt I'll jump on it given Intel products as of recently run hotter and slower than AMD, and I don't see how Willow Cove which max out at 4 cores can be faster than 8-core Ryzens; to add an insult to injury, they only seem to support LPDDR memory I misread the slides, all OK with memory support. But still, I genuinely hope Intel did something right for a change.

 

@Starlight5 current and future memory are fully supported.

 

@tilleroftheearth you're right, I misread the slides, only CPUs with 9W nominal TDP are limited to LPDDR according to ARK.

 

@Starlight5, I learned something too!

 

it's...no need to wait that long.
Intel had actually pushed out the APUs before the announcements to OEMs.

There's going to be lots of comparisons and benchmarkings.
We'll get to see if the Tiger Lakes can standup to le Renoirs soon!​

 

Here is to hoping Intel can catch up a little in the U series line as the AMD 4800U demolishes it, even beating the 45w H series in many things:

 

Oh, see that's not "real work" so it doesn't count. I'll only provide vague platitudes about what "real work" constitutes, but Intel wins because I say so.

 

Lol I do like Steve @ Gamers Nexus break down of "real work" where he calls Intel out on saying that they are better at real work and not benchmarks by showing benchmarks that don't simulate real work...

 

Let's try to compare apples to apples, huh? 45W H CPU's from 2019 and 2018 need not apply. As I've already stated, let's see what the 2020 model battles witness for us.

Vague platitudes are not coming from my direction. My workloads are real, and I know which platform(s) work with them. Even if I can't share enough of the details here to please everyone.

Are Intel's benchmarks handpicked? Sure, as is everyone else's. But they're not pulling them out of their a$$ either. I'm willing to wait for the real comparisons to happen. At least Intel is picking a current competitor's platform to compare against. But yeah; Intel, so I guess no points for that?

 

I mean it clearly shows Intel's current mobile offerings being totally dominated in every workload they tested by up to 2x. I was simply stating Intel better have something up their sleeves. Unless you have a totally single core workload, AMD currently wins (oh and in the 15w market they still win in single core). I'd love to hear what your workload is though... And yes Ice lake came out late 2019 and Zen 2 a couple months after. We can always play the what's next game since Zen 3 is right around the corner too. Intel is clearly shaking in their boots since they mentioned 4800u in their conference more than they mentioned their own product name (which they didn't even remember, lol).

 

No one is shaking in their boots. They are are firing on all their respective cylinders. But there is no doubt that they have different strengths. Even if the nuances are lost on many.

What you fail to mention is that every single time AMD was mentioned, it was at a loss to Intel.

 

Lol... Tiger Lake (Willow cove) doesn't appear to be anything special.
Its IPC has remained the same as Sunny Lake.
The only reason they may have gotten higher performance because its a 4c/8th chip... that leaves Intel with A LOT of room to jack up the clocks way past what AMD can do with double the cores.
Oh and that also (by default) means Intel has more space on the die to put in more GPU cores... and thus far, it seems its only 20% better than enhanced Vega in Renoir.

That's actually not 'groundbreaking' to me... yes its a step forward for Intel (and good on them), but they (and some reviewers) are actually getting a bit ahead of themselves - especially when you consider the premise that Intel cannot make an 8c/16th low power chip on 10nm like AMD could on 7nm (which is only slightly better density-wise than Intel's node).

That said, Zen 3 is about to be released too with general IPC improving by about 15-20% (possibly more given its a completely new uArch)... then there are design improvements which unify the cores to clamp down on latencies, along with a minor clock increase (the new node only allows either 10% increase in efficiency or 7% performance increase)... so if clocks go up, they will probably increase by 10% (to gain 5% performance improvement).

While enhanced Vega won't change from Renoir to Zen 3... games will still benefit (aka FPS will be higher) due to the overall stronger CPU design.

So, come Zen 3 APU's and I think that AMD and Intel might end up matching up in both single-core and graphics performance (with graphics MAYBE going to Intel a bit)... but multi-core performance still goes to AMD (heck, it will still go to AMD with Renoir).

Oh well, we shall wait and see... but we still need to see how Tiger Lake will perform when tested by third parties in a thermally constricted environment (and whether OEM's may give Intel options better TDP range by always using 'unlocked' versions as opposed to configuring them solely to 15W like most 4800U's are)

 

Tiger Lake doesn't have to be special. It only has to be better.

See:
https://edc.intel.com/content/www/u...11th-generation-intel-core-mobile-processors/

So far, it seems like it is. Let's see if it delivers.

 

But that's exactly my point.
Its NOT intricately 'better' uArch than Sunny lake (except perhaps in graphics).
It just gets away with higher clocks and more cores because the chip is limited to 4c/8th, and Intel is incapable of making an 8c/16th version (which would definitely drop the clocks severely for them along with the number of graphic cores).
And we still don't know if it will behave like Intel advertises in a thermally limited environment.

 

Right, we will have to see when real reviewers get their hands on them. Last time Intel paid an independent company to review their product it was quite the shady event. Either way I do enjoy new tech and reading about it.

 

The only point that is important is that 'better' is from what we had before. Tiger Lake so far shows it so.

uArch improvements are not important on their own. Better user experiences are.

I'll take 'old' tech that performs better any day over technologically superior, but performant inferior, products.

 

But you don't know that they perform better. You also still seem to think that 10th gen is better than Zen 2 when it comes to mobile, and it's clearly not. I may be surprised, but I doubt it. The one area I think Intel will be better this go around will be the adobe suite, but that's something they've always been strong at... You tout user experience, but never say how it's better btw. I've only owned every Intel generation from the Pentium M till the 10th gen and Zen 1 through 2 and can say may user experience has been wonderful with both. The only thing that's less fun on AMD is overclocking, and that's not a typical "user experience."

 

Coming Soon... In a "flipper" near u!
https://www.lenovo.com/sg/en/laptops/yoga/yoga-c-series/Yoga-9-14ITL5/p/88YGC901455

 

I never said that we do know it's better, we're waiting for verification on that. Just what has been indicated so far, even if it's only from Intel (I don't doubt, nor believe, before I test myself. I just use that input as something to compare my testing with). User experience is not just the RUGS that is better. It is every aspect of ownership. Glad your experience is equal between Intel and AMD, mine is not.

I state it again because memory around here isn't what it used to be. Incompatibility between programs I use/depend on. Driver issues. Software (interaction) issues. Platforms needed to be replaced within a year for CPU issues (no overclocking applied). Search the posts here. I'm not looking them up nor repeating them again.

Will AMD get better? Sure, eventually, I hope. Does AMD beat Intel on some performance metrics? Sure, but having the fastest car doesn't mean anything when it spends most days in the garage getting fixed.

Let's forget the past and look at the near future. Athena 2 and Intel Evo Platform(s). That is progress for me. What is AMD doing in this regard? zzz...

For an objective comparison, look at the Surface Book, Pro, Laptop, and (I'm hoping) Duo lines. The hardware gets out of the way and just lets you do. IR face recognition, fast(est) startup/resume speeds, Pen input, precision touch displays, good enough battery life (today). That is the user experience for me. None of those have the best hardware specs. Yet they surpass other more potent hardware because of what they let you do, and how they let you do it.

What gets me out of working and back to my friends/family faster is what I consider a win.

Certainly not a nominally superior hardware platform, but with obstacles to that final execution (the sentence above).

See:
https://www.tomshardware.com/news/a...specifications-pricing-benchmarks-all-we-know

Even with best hoped for and deliverable IPC/performance updates on Zen3, AMD is still behind Intel with the info given by both sides so far.

Zen 2? Don't blink, but if you did, it doesn't matter anyway.

 

"AMD says the 400-series BIOS updates come as a one-way path: You can flash your motherboard to the new firmware, but you will not be able to flash back to previous BIOS versions that support pre-Zen 3 processors.

AMD will also require an as-yet-undefined process that verifies you have purchased a Zen 3 processor. AMD says that verifying a customer has purchased a Zen 3 processor is designed to prevent irreversible BIOS flashing issues, but AMD's reasoning behind the forward-only flash requirement is unclear."

¯\_(ツ)_/¯

 

So you can't actually tell me how the user experience is better outside of making generalities. In the hundreds of systems I've had over the past 6 years I haven't had an issue with any of them except one XPS where the MB failed. And RUGS is just what Intel happens to find that they have a lead in (regardless of how obscure the software is, or meaningless the number is such as faster in word) so they advertise it as being better... The examples you give of laptop features are a function of the laptop, not the CPU. It's clear you are biased and won't be swayed, but that's ok, you do you. I'll remain skeptical until these products get out into the market. Until then I'll stick with AMD because I happen to get a 3.18947x better user experience based on my randomly generated number typing experience on my desktop this morning...

 

I have spoken (okay, typed), I've shown where to look, but nobody wants to hear or see what is right there in front of them. These are not generalities, have you ever used a Surface product for an extended period, fully?

Like I said, your experience with having no issues in the past 6 years is great. For you. Not my experience. And not for many Intel systems too. But for not even one single AMD platform that I could potentially incorporate into my workflows either. For many reasons, performance being almost secondary. Stability/reliability and compatibility being king when a new platform is being evaluated and the goal isn't to have 'fun', but to be productive.

The functions I expressed are not a function of the laptop, they are the function of the platform that Intel and Microsoft pioneered. I can't do the same level/quality of work on anything else (even on an otherwise great, in a void, ThinkPad version).

I am not biased. I paid my $$$$$ and returned many Surface products until they worked as they were advertised. This is not a 'rah! rah! I love ' name of corporation goes here' chant from me. It never has been.

I buy, I test, I keep the best. But only when that new 'best' is better overall than anything I am already using.

I'll repeat this once more for your benefit. I don't buy and keep anything unless it is beneficial over what I already have.

The bias insinuated is not coming from me. Sticking to defending AMD even when I provide something for you to chew on, is.

 

Saying "Surface product" is a generality since there are lots of them. Surface Book? Laptop? Tablet?

You may not say those exact words, but that doesn't mean you're not biased. Anyone looking through your post history will see that. Denial ain't just a river in Egypt.

 

No, there are not lots of them. Surface Book. Surface Pro. Surface Duo. I've already stated this (repeating everything is really tiring).

There is no river called Denial in Egypt. The bias is still being shown on the AMD front.

 

Does SPECRate perf. look modest?

More spoiler alerts. I think they didn't test plundervolt and newer June 2020 microcode. If there are more exploits, investors will invest their money in Apple ARM, AMD, Power9 etc....

So Intel is saying CB scores @ HWBot is not real world performance and more gimmicks they performed to get Leaderboards before Ryzens arrived.

Intel based jokebooks are really expensive here in India whilst in comparison to similar Ryzen based models.
Ryzen models are crippled by OEMs say soldered DIMMs maxing to 16GB or even less while same Intel can do 32/48/64GB with replaceable DIMMs, SSD. Even screens on AMD Ryzen model are not calibrated and sometimes awful.

 

Sometimes, ' expensive', is, where the money should be placed. The issues you state about Ryzen models make them the true jokebooks at any (discounted) price.

I want the hardware to disappear and I've been saying that from the start (when I joined this forum). Intel, Microsoft, and Lenovo do it the best and at an entirely different level than any other platform and/or manufacturer does, still.

What Intel is saying about CB scores is that that is not the only metric that is important anymore.

Just like a 600HP 'muscle car' doesn't surpass a 443HP Porche, for example. Even if it can get 'there', faster, in some narrowly specified scenario.

And to be blunt, Intel is showing exactly what is faster (real-world workloads). If your workloads don't match, go ahead, and ignore the new platforms.

To me, CB scores are nice but irrelevant. I'm sure I'm not the only one whose workloads don't depend on this mere ' number' to validate the platforms I choose to run them on.

 

I don't use AVX-512 ML/AI extensions nor any Workstation class workloads.
I just run few db instances, few VMs, 20-100 tabs on firefox/Chrome, 2-4 instances of VSCode.
I believe CB running in 3-5 cycles can really push the hardware and you get overall idea about thermals and performance when its heavily loaded eventhough its not gonna happen every time.

 

We agree 100% there. CB (and other software) is used to stress the system on all new platforms for the first 36 to 48 hours or more before I even consider using it with any of my workflows. But I don't care what the numbers are. And consequently, the Surface and ThinkPad preferences I have. Other manufacturers just don't get how to make a system that performs (day in and day out), stays cool, and keeps working without issues like those products can.

When the hardware is proven stable, running the workloads through them, and seeing the total time taken is all the proof I need of the worthiness of the system(s) I test. If they are indeed faster than what I'm presently using, then a simple cost analysis determines how many systems will be upgraded, now.

 

Thinkpads do come in the AMD flavor, with the same features as their intel counterparts. Again it's a function of the laptop and not the CPU...

 

As does the Surface Laptop 3. Heck, the only way to buy a 15-inch version is with an AMD processor.