New Generation Intel CPU's 'Tiger Lake' Processors

Rendering videos is not real work so it doesn't count. Real work is having your computer boot in 6 seconds instead of 7.

 

Linus isn't too hot on Tiger Lake either - for him, the most convincing arguments to get an Intel notebook are

1.) AMD shortage, which also leads to
2.) The better notebook designs containing Intel, as Dell and co. cant commit to a design with AMD without guarantees of a steady supply.

IMHO, Tiger Lake seems designed to run on small and thins for professionals not doing any computational work (mostly in the service industry). Anyone seeking to use Adobe or AI or whatever benefits better from the better integrated graphics on the go is better off coughing up the extra dough for a laptop with a real graphics card (or a remote solution).

Plenty of students and professionals will benefit from Ryzen, which has more cores (faster compiling, MATLAB, virtualization, content creation in general, etc.). On the other hand the benchmarks Intel touts are either near useless (AI), better served with a GPU (Adobe, gaming), or simply shaving off minuscule amounts of time (Microsoft Office). Saving a couple of seconds on boot and when writing a report isn't much consolation if you burned it all and more on some workload where you would want the multithreaded goodness.

That being said, for 12 year olds getting their first laptop, even if daddy isn't willing to pay for a graphics card, maybe now they get to game, lol.

 

I wouldn't disagree.
Race Car Drivers' biggest test of ability is on a race track, while,
Computer Users' biggest test is on the computing speed journey.
Be it a Vehicle Driver, or a Computer User, they are subjected to be met with hurdles. And from the previous discussions, we cannot deny that Intel users' computing route is much smoother as there are still lots of computing tracks out there tuned for Intel "drivers". Linus comparison of the Tiger Lake sample against the 10th Gen Intel XPS certainly puts a shining aura on top of the testing sample as the storage SSD type and the RAM type is a big win on the "older" INTEL.

In the Tiger Lake introduction video, we're clearly "told" that the Tiger Lake is out to take back the computing throne in the U-Class arena against Ryzen 4000s.
However, Linus benchmarking clip with CineBench 20 gave a slap on Intel's face that even with "slightly" lower spec, AMD's 4800u can still complete it's job faster.
Yes, I won't disagree, Intel's single core performance surpasses Ryzen CPU on certain makes as well as programs. And the comparison choice initiated by Intel(which later got reviewers to pull out the specific make with that mentioned make for comparison, resulted with a big slap) had been little... "unfair" i would say. 4core Intel vs 8core AMD? What is Intel's marketing folks been boozing on? Bootleg Whiskey?

 

Some falls on Aesthetics, Some falls for Performance.
It's like Car and Driver scenario...
Linus strikes on 2 points. However, I wonder is it a miss or a deliberate miss to mention that not all programs or software are designed to work well with Intel system, but not for AMD engine.


Can this ride well on the track design for INTEL?

Let's wait and see...

 

Same core and thread counts.

https://hothardware.com/news/amd-ry...en 3 Pro,3 4300U wields 4 cores and 4 threads.

It'll be interesting to see how INTEL had improved so far with their best against the current..."peers" from another camp. Scout to Scout, Driver to Driver, Sportsman to Sportsman, Competitor to Competitor.
Should AMD loses, no problem. At least they had won a "reputation" that their "little man" can be put up to the same arena against the big Blue Guy.

 

I won't disagree with u on this. When I search for infos on CPU to ACad compatibility, I saw numbers of information regarding ACad doesn't require multicores to process a work(hmmm... I guess this is the same for GPUs as well? Hence, most ACad targeted GPUs can't game? ). However, is drafting the only profession in this world? Nah... I don't think so. There are certainly other lines of professions that may require multicore processing.

 

Most users won't notice any snappy difference between 4.8ghz vs 4.2ghz (especially when you factor in IPC). 30% battery life increase? They can't determine that when comparing different laptops, you're extrapolating 30% from an article that isn't even talking about battery life. You're really just talking about your opinions again. You talk about real world usage but then you mention that rendering isn't something people do, but instead run word and time how long it takes to spell check. We won't see eye to eye on this by any means, that's apparent. I'm glad that Intel has been able to catch up to AMD, in that rights it's a success, but saying that it's better for every user isn't a fact. The platform argument is dependent on the manufacturer of the laptop more so than the CPU itself. AMD has no worse of a platform than Intel in my numerous experiences. You can try to dress up your argument all you want, but it's not going to make it factual.

We will see what happens when zen 3 comes out. Intel caught up to Zen 2+ which is great, but I don't think they are going to maintain their back and forth come October...

 

the evaluation samples that Intel sent out were based on Zephyrus G14 if I'm not wrong。and i wonder that 30% increase in battery life，is it due to the extra battery capacity which by default，is much more fatty than other makes in the market with average of roughly 50whr or smaller？

 

I'm not sure what you mean by evaluation samples sent by Intel. The G14 has a higher power mobile CPU (4800h/4900hs), not a 4800U. And battery life is really dependent on battery size, laptop specs, drivers, etc. I get an easy 10 hours with my G14 with a 76whr battery, not too shabby considering the specs including a 120hz screen...

 

If you watched Linus’ video, it’s based on an MSI chassis.

Linus also made the point that if battery life were notably better, Intel would want that shouted from the rooftops. Instead, he specifically addressed their NDA saying not to discuss that particular topic.


Sent from my iPhone using Tapatalk

 

Ahhh... I got that mixed up!!!
Yeah. Rewatched that vid again, yep, both Dave n Linus mentioned of MSI.

 

Yeah, when we can't attack facts, let's just make them up (or compare to 2-year-old competitors' products). Yawn.

Everything I stated I backed up with a quote from the two articles I linked to (one much more against Intel than the other...). We do not need to agree completely, you just need to keep your opinions and assumptions in check too. I didn't state rendering isn't done. I state most don't do or even need to do it. Every user? I am writing plainly for you guys. Sigh...

No worries, but I would hate to have any recommendations from this bunch when a real upgrade to productivity is needed.

Nobody stated that video rendering isn't 'real work'. But doing that on a mobile system with a 15W/28W budget is not using all the brain cells available either. Booting up your computer faster than before maybe a huge benefit... if your system is booted dozens and hundreds of times a day. Calling it 'real-work' sarcastically though is not helping you state your point.

Is Linus, still relevant? Not on this topic. The bias is obvious from a decade afar.

MATLAB, virtualization, and content creation are not what most people do on a laptop. Small and thin isn't for service industry professionals either. Intel has 5 'wins' to AMD's 3 in the AnandTech article for Simulation and Science comparisons, for example.

A real graphics card or a remote solution (if it proves to be superior, of course) is available on any platform. Contrary to your beliefs, AI is what is making a lot of this superior onboard performance possible. Re-read the links again if you must. Drafting reports in Office may be looked down upon, but usually, you have a handful of webpages open when you're checking facts/etc. Tiger lake is still over 25% faster from a 15W platform vs. the 65W 4750G Pro.

Cinebench 20 is not a 'job'. We know how much AMD loves that one... (when you're a hammer, you see everything as a nail - it's not like that though). On the notebook/mobile front, AMD hasn't won anything for decades/(ever?). They did catch up a little, but then, Tiger Lake. Maybe it will be their turn next time (a few short months away) with Zen 3. I'm not holding my breath though from everything I've read posted by yourselves.

The NDA to not discuss battery life was because this isn't a system that will be shipped. By anyone. Not with fans on MAX 24/7 when the system is on. Linus, if he said that doesn't know what heads or tails is here.

So, demand exceeding supply is now a sin. Making the most wanted notebook processor(s) is now bad? And then they dare to release a version that is 98% of that most wanted CPU and at less than 93% of the $$$ and that is also a evil too? Yeah, slap for staying still, slap for moving (but only if it's Intel, it seems).


Me thinks that the 'retorts' against Intel's worthy new platform are not coming from the objective perspective we should be using?

 

My opinions are simply restating what the website you linked to states. They trade blows on numerous tests, they don't mention battery life, and AMD demolished Intel in encoding (and yes Handbrake is something you actually use to do encoding). I don't need to ramble on to state this point. I'm not saying that AMD is better for everyone, I'm just stating that I'm glad Intel caught back up, but doesn't provide a clear choice still. A person buying a laptop should buy what fits their needs best. Simply stating that everyone would benefit from Intel at this point is something that can't be backed up.

 

Simply stating that I am saying 'everyone' is only plain wrong. I haven't stated that ever.

If you concede that point, then it looks like we do agree altogether on this?

 

You're right, you didn't say everyone, you said most users and that AMD is in second place for productivity-based workflows for mobile. I am disagreeing with both of those points.

 

I've no problem with Intel taking back the computing throne.
However, if a fair game is played... same spec SSD with same spec RAM to work together as a team...
Most of the applause will still goes to the 2nd ranked Ryzen, with it's price and bringing down to affordability of a system in whole.

 

You can disagree all you want. The fact remains that at 15W/28W the productivity-based workloads/workflows are not what you think they are.

High core count workloads are also not matched to a 15W/28W platform either if actual performance/productivity is required. But yeah, there are use cases for high core count workloads being done on a mobile 15W/28W platforms - but certainly not for most users.

A fair game is to let me max out two systems in RAM and SSD's that I would be considering and let the differences speak for themselves. I am not buying what I already have. I want to buy better.

The 2nd ranked Ryzen can forever have the applause. I desire the first-place winner instead.

 

Great that we had finally comes to a point where what a "fair game" ground should be...sort of.
Yeah!
Max up the RAM! Let's go 64gb on 2channels... with same clock speeds of course.
Max up the storage! Let's go same make, same reading and writing speeds on both fields.
Let's hv the same chassis with same capacity batts... If tests are to be on how long the new TigerLake systems can last without mains connection.
All these are fair demands. And of course not forgetting the choices of software and programs that are tuned for both players!

No problemo for most here, if INTEL can prove itself, "the throne is Mine! Mine!"

 

You're missing the point. The maximum performance/productivity overall is still the 'best' even over a superior design/node/'raw CPU HP' that still comes out second best when the product is used as it will be delivered.

This isn't mostly a matter of showing up in the prettiest/'techiest' car/platform on the runway. It is mostly a matter of crossing the finish line first, with style and flair. Having the new platform show up in yesterday's old 'clothes' proves nothing except that old style is old.

Rather, let's see if AMD can step up to the latest standards Intel is aiming for. Then we can have a truly 'fair' comparison.

 

@tilleroftheearth

I don't actually disagree that the Intel offering is likely going to shape out as the superior one on a typical ~25W workload (office worker slaving away on a spreadsheet and powerpoint, sales agent on the go, etc.). I do take issue with the trumpeting of intel's RUGs - which are cherry picked in their favour, some of which are of dubious use in the light workload you tout (AI?!) and still (for now) only on paper. I brought up MATLAB because students (STEM in particular) are one large group that likely benefit from the raw horsepower - if not with MATLAB, then with something similar, and who generally can't afford to just buy the best at any cost.

Why is Intel bringing up performance benchmarks of dubious worth when they could be touting power consumption and battery life (they certainly don't have any qualms using their test system for their benchmarks, so that shouldn't stop them)? If the new Intel offering runs cooler and lasts longer than AMD counterparts then that is a way more compelling reason to grab Intel for a typical slim/light user.

Thermals and power consumption aside, the value proposition of AMD is clear - it provides significantly cheaper and generally less power hungry options if you can benefit from >4 cores. Intel has the ecosystem behind it due to being way more established in the mobile market (which might lead to better overall laptop designs, less driver issues and so on), the better integrated graphics (assuming it is cheaper than getting a 4800H or similar + a graphics card), and the higher single core speed which plays into several important tasks.

I'm waiting for Tiger Lake 8 core laptops - if they run similarly or even slightly worse compared to AMD offerings on multicore workloads (compiling in particular) at the same price, I'm probably going for the Intel, for the reasons above. Not expecting it to happen based on their track record (the 10875H is a blazing inferno and generally more expensive than a 4800H), but if it does that's a pleasant surprise

 

If you feel that RUGS are of dubious worth, then this 13"/14" ultra-mobile platform isn't what you should be looking at. These are systems as thin/light as a ThinkPad X1 Carbon...

MATLAB minimum specs seem to point to the Intel, no? Especially with the superior memory subsystem Intel has. If that is wrong, do you have a link where MATLAB is compared between Intel and AMD platforms? In any case, MATLAB students (and their mobile STEM requirements) are not a large group in any sense. The users who use MATLAB and similar for more than 'grades' don't depend to run their work on mobile systems, I'm sure.

See:
https://www.mathworks.com/support/r...will be discontinued in an upcoming release.

A four-core processor with the combined power on tap of the Tiger Lake platform isn't just for a 'slim/light' user. You may just not know what other uses are available, when AI enters the picture (pun intended).

I too would like to see an eight-core Tiger Lake notebook in a bigger chassis to see how my own productivity would increase.

However, what Intel promised and what many independent reviewers have already clearly reported is that Intel has delivered.

AMD has long stopped being the value leader. They are just as greedy as the next corporation too.

 

Quick googling turns up multiple tutorials on how to multicore MATLAB (and several benchmarks that show Intel generally wins slightly per core due to clock speed, but more cores always wins). I am still young enough to remember what it's like being a university student - most of us (especially international students) only had a laptop in the dorm as a daily driver, and being able to run calculations, test code, and all that performantly matters a great deal, even at the undergraduate level. Having the 8 core laptop on hand saves you the unpleasant choice of cycling to the lab to use a grimy communal student PC or watching paint dry while your potato machine chugs slowly along.

Granted I would almost always recommend a student in that situation get a beefier (#doyouevenliftbro) laptop, but if the 4800U provides a cheap, light 8 core mobile platform with only slightly worse performance than the 4800H, then it is certainly a very attractive option.

 

okie... Intel certainly had gained a slap on AMD thru
hardware canucks

However... What the reviewer had failed to disclose,
was how "incapable" on the competitors "hidden strengths"
(i'm still bugging onto fair play specs)​

 

Adobe's results are due to literally no support for AMD graphics from the Adobe side, and its interesting that MATLAB runs way faster on 11th gen.


https://www.techspot.com/review/2096-amd-ryzen-4800u/ shows 4800U and 10th gen in a dead heat with the 4900HS/10980HK (an 8 core processor) about 15% faster - which probably implies a mostly lightly threaded benchmark with no major systemic advantage to either processor. Basically doubling to tripling the performance in 11th gen is weird to say the least, especially because MATLAB only claims support for NVIDIA gpus.

Beyond that, nothing we didn't already know from Intel's presentation - faster in office due to custom support, significantly better integrated graphics, significantly worse multicore muscle.

 

Most will agree with this... honest take on INTEL's improvement "impression"!
Customization! And that's the kind of road layout to INTEL Nascar where AMD F1 may have little control.

 

Here's my shout out to those social media reviewers out there:

We can't ask you to reject INTEL's provision of sample for your review. Certainly, their key goal is to get recognition of improvements from reviewers first, then from consumers. However, was the game played fairly, i have my doubts, so as others who may have such thoughts. In INTEL's launch presentation, their "slides of scores" were clearly indicating that this NEW INTEL is out to take on the best U-Class CPU from AMD Ryzen 4000 series. So... should there in your planned review, if you're doing comparison of this "evaluation sample" supplied by INTEL to other AMD makes with same size chassis or screen, please see that you're doing the comparison in a fair manner as i believe, RAM type of same size can be different, as well as storage type where reading, writing, paging difference in speeds(or clockrates) can eventually end with different benchmarking results!

 

A 'built-in' benchmark function is worth less than the programming required (once again). But doubling or tripling the 11Gen performance is not weird, it was expected... and, with the 'quick google search' I did (see below), even if it's from 2019, MATLAB likes Intel for at least some aspects (STEM not found), nor for the video above from @Ed. Yang which also shows huge wins for Intel on the MATLAB run at 8:37 into the video which places the Intel 2.92x faster @ 35W.

Nothing 'interesting' about it. AVX512 support is simply huge for any modern workflow. Just like Intel has indicated.

See:
https://www.reddit.com/r/intel/comments/celdu8/update_intel_vs_amd_in_matlab_benchmarks_july_2019/

Some interesting comments from the above below.

What the above quotes say to me is that AMD even with multicore available isn't an appropriate choice if that multicore platform is AMD. This may have changed from 2019, but somehow, I doubt it.

Multicore on its own isn't a savior, even if it has its place.

Future improvements won't rely solely on merely supercharged V8's or even V12's... See the future that is in front of your eyes.

Within a given power envelope, each manufacturer makes the choices and tradeoffs they see fit. Intel's choices give the best jump in performance for a mobile platform once again. I'll be patiently waiting for their 8C/16T 45W+ take on a worthy mobile workstation chassis (preferably a ThinkPad) very soon.

 

2020a introduced an avx patch for amd that increases performance up to 4x on some tasks, so the source above is out of date (although it doesn't explain the astronomical jump cross generation)

Do agree that the prospect of a proper Intel offering at the higher power envelopes for beefy chassis is exciting though, although we'll have to wait at least half a year to see how the Ryzen 5000 mobile turns out.

 

Here's my shout out to those social media reviewers out there:

Stop polluting the internet with nonsense drivel and give us examples of real-world workflows/workloads that people use every day.

 

Intel is manipulating scores by paying companies to use their compilers and forcing AMD CPUs into non optimised code paths that done even take advantage of the whole instruction set, and you are drinking the cool-aid and spewing BS like you can't see past Intel marketing, just proves that Intel marketing still pays off in the end..
Have you even used a Ryzen 4800H based laptop?

My Asus TUF something with a i7-8750H needed to go for warranty and I got a new Asus TUF something with the Ryzen 4800H, both running VMs from the SSD, its not even close, Siemens TIA portal opens up in half the time on the Ryzen based laptop, stop being naive and using cherry picked benchmarks.

If anything Intel using 35W for a quad core CPU in 2020 is a travesty, but you still think that the CPU spiking above 35W to deliver 4.xGhz all core is some mean feat....

 

How unfair/unscrupulous Intel might be doesn't factor into purchase decisions. What does matter though are price, power draw, performance, and broad compatibility (by bribing companies or otherwise). That being said, the intel 8 core mobile chips available right now are terrible compared to the 4800H in competent designs - Intel runs hotter, costs significantly more, and does not offer any tangible performance advantage in multicore workloads, while I have not observed any driver issues with the 4800H or even the 3950X in the Apex 15.

I can't see myself ever using the 1185G7 (using anything below 8 cores is quite literally wasting money), and found the Intel presentation mostly marketing nonsense or irrelevant to my personal workflow even if Tiger Lake did already come with 8 cores. I also suspect workflows that benefit from >4 cores are more common than Intel would have you believe (chrome + word + excel + custom work software + conferencing + ... adds up quick). However, this 4 core chip aims to compete is in the ultra slim/light niche. There is a market for Tiger Lake (specifically, the current 4 core chips) if it can provide smooth experience at 15W on light (bare minimum 2 office programs + chrome + company software + windows bloatware) usage - we'll have to wait and see real laptops with the processors though, and either way, I'm not the target audience for these so /shrug

I'm not terribly optimistic for the next generation of intel mobile 8-cores - half expecting Intel to throw their hands in the air and delay Tiger Lake 8c just like 7nm, not to mention Ryzen has a lot more room to maximize potential of their 7nm architecture - but if they do surprise me with at least 90-95% performance at similar price/thermals to a Ryzen 5000, my next laptop for moving round the office might very well be an Intel, if I deem certain features (the higher single core clock for the odd task that just can't be parallelized, intel VT, better integrated graphics, etc.) worth the performance drop.

 

And do you receive a cheque from AMD to state this for them?

How about showing us a few payments made to these companies that Intel is so much in control of?

The old 'Kool-Aid' argument is really getting old here. There is nothing based on 'just' marketing here. Independent reviewers have verified Intel's claims, really.

Have you even read about Tiger Lake? What is this nonsense for i7-8750H and Ryzen 4800H about? Whose Kool-Aid have you been drinking and how much are they paying you, exactly?

Intel doesn't have a 35W QC CPU here. It's a 15W/28W chip that can use up to 50W or so, just like the 4800U Ryzen does.

What is a mean feat is that this little 4C mobile processor, "The i7-1185G7 is at the heels of the desktop i9-10900K, trailing only by a few percentage points.".

Let's put this in perspective. A 28W chip from late 2020 approaches what a mid 2020 125W chip with a higher base and boost clock and 2.5x the Cores can accomplish a few short months ago.

Companies use compilers for the hardware they use and run. AMD needs to get their $@$@ together and make comparable compilers for their hardware.

But yeah, cry Kool-Aid. So much cooler looking than declaring Intel is stepping up in a big way here.

 

Yes I am glad that they are stepping up to play catch up for sure. Competition is a good thing for everyone.

 

If their 4C's can catch up so well with the competition's 8C's... when their 8C lands it will be "AMD who?" once again.

 

And if that happens my next systems will be Intel, but I'll see how it plays out.

 

I checked with a more expert colleague in that domain (as personally, I don't use MATLAB or machine learning stuff). Apparently, Tiger Lake comes with a new AVX-512 instruction https://en.wikipedia.org/wiki/Tiger_Lake_(microprocessor) that vastly speeds up certain calculations. That would explain the three times speed up across generations, assuming the MATLAB benchmarks were cherrypicked correctly.

Not the expert so no idea how valuable that instruction is in general, what its cost to implement is or what its performance gain is on software optimized to make use of it. Linus Torvalds isn't too hot on AVX-512 in general:
https://linuxreviews.org/Linus_Torvalds:_AVX512_Is_"A_Hot_Mess"_"I_hope_AVX512_dies_a_painful_death"

 

https://www.notebookcheck.net/Intel-Core-i7-1185G7-Processor-Benchmarks-and-Specs.467757.0.html

Sure, its on the heels of a 8C CPU, CB R15 (866 score) and CB R20(2214 score) put it on par with i7-7700HQ, sure, it performs better on some AVX bench made for the CPU..

With 8 cores it will perform like a 4800H:
https://www.notebookcheck.net/AMD-Ryzen-7-4800H-Laptop-Processor-Benchmarks-and-Specs.449893.0.html

CB R20 Score: 4298
CB R15 Score: 1850

Sure, single core is better than AMD, its not a secret, 482 on the AMD vs 580 on the Intel.

Doesn't seem like anything special TBH and its just another incremental update of Skylake with more TDP liberties during X amount of time to complete the typical benchs and look good, then its PL1/2 limit town just like we already know.

There wont be any receipts of Intel screwing AMD, because thats ilegal, but Intel has done so for years..
Read about the Matlab performance before and after it being patched:
https://se.mathworks.com/matlabcentral/answers/330889-does-matlab-perform-well-on-amd-ryzen
Oh look how AMD magically performs better when its CPU is allowed to also run AVX instructions....

 

Ryzen's latency is much lower and I can really feel that in opening apps or any intensive apps on Windows/Linux LiveCD etc..
My experience is with Ryzen 3200U 2c/4t and even with low end SK Hynix NVMe SSD (TLC) it can boot Windows super quick without staring at oem logo for 10 secs ran AIDA64 AVX stress test for 3 hrs w/o thermal throttling and EDP current limit. The improvement is much higher if you install AMD Chipset driver and Ryzen Balanced Power or 1usmus Power plan.
Any intensive apps such as CATIA, Maya, SketchUp (small projects work best on Intel for higher Single Core clock speeds), VMs w/ w10 1903 home/pro 2 instances with hyper-v paravitualisation on vbox.

 

Tiger lake overall actually looks like pretty decent upgrade and if the rumored Tiger Lake H comes out it will be some more good competition. AMD on the other hand seems to have improved its idle power draw and their U chips are still pretty compelling, I hope they keep up the pressure.

As for AVX512 annoying thing is that the instructions that comprise it are pretty fragmented across SKUs. However if applications take advantage of available functions like that to boost performance there's nothing wrong.

Maybe I'm wrong but after the lawsuits and anti trust cases (some mentioned below) I doubt Intel will be doing anything like that blatant in terms of exclusion rebates at least so soon. Working with devs to optimize compilers or software to their instruction sets, I don't see much wrong in that. Intel with its larger marketshare would have an advantage there with devs. Thats the thing about being an underdog not just an Intel vs AMD thing, even with decent offerings sometimes it can take a long time for all the pieces to fall into place in terms of recognition and getting vendors put your parts into a higher end offerings and you have far less room to make mistakes.



@senso was over a decade ago right where Intel paid Dell not to use AMD chips by giving rebates worth $4.3. billion. Intel did that with other vendors too. If I recall Dell specifically got in trouble not for the rebates but they way they showed it as income to investors:

https://money.cnn.com/2010/07/23/technology/dell_intel/index.htm

The JFTC (Japan in 2005) and the European Commission (2009) rulings went against Intel for rebates and anti competitive behavior.

 

Linus has his head in the sand. So out of touch with what is happening today.

The 'benchmarks' aspects I can agree with, but these weren't just benchmarks. These were useful tools that anyone can (and will) use. Even on a 15W and lower platform (the AI/AVX512 enabled RUGS).

You and AMD wish an 8C Tiger Lake will perform like a year old, last-gen Zen 2 AMD model.

If that is all you're expecting, then you're right. Nothing special. That isn't what I see coming down the road.

Performing better still doesn't mean the best. You stick with AMD (and keep making excuses for them too while you're at it). I'll stick with actual performance/productivity instead.

Of course, there are receipts. It's called playing ball (i.e. 'business' with the big boys).

We'll see, soon enough. When Tiger Lake 8C/16T is revealed (and reviewed).


Here's a mission for you guys. Go look at all the Samsung 980 Pro Reviews. See the nice synthetic ' scores'? Now find reviews that show how in ( more) real-world workflows, even the now 'ancient' and budget Adata SPG SX8200 Pro is still faster and more responsive than a drive that in mere 'scores', easily outclasses it. At least in some workflows. And with PCIe 3.0, not PCIe 4.0 'tech' either.

Now, do you get it? 'Scores' mean squat. Completing an entire workflow/workload(s) faster is the significant métier (always).

That is what Intel is going on about (and has been for quite some time). That is what I was stating over a decade ago. 'Scores' don't mean squat to me (they never have - because they don't predict what will happen to my workflow). A significant increase in productivity is the essence.

Otherwise, 'tech' is just for the kiddies who argue about whose ('score') is bigger.

 

And thats what I told you, an Intel laptop vs a Ryzen laptop, both with the same 512GB WD SN550, the Ryzen laptop does the same in half the time, from cold boot of a VM to having TIA Portal V15.1 running its 2 minutes on the Ryzen and 4 minutes on the Intel, the exact same VM copied over, both with 4 cores to the VM, both with 8GB of RAM allocated to the VM, to me seeing that was a real eye opener, the whole VM feels way more snappy on the Ryzen laptop, by a long shot, and I dont run bloatware on my laptops..

I dont know you workflow, but you do, so, give a Ryzen based system a try, and you might be surprised, just dont go comparing an HDD based Ryzen laptop with 4GB of RAM vs a 128GB 4TB RAID0 Intel one, put them toe for toe, and go with an open decision, because in your mind Intel already won and you wont even give anything else a try.

"Playing ball" in EU is called bribery, and in the US is called lobbying and its used to stuff lots of pockets that happy to screw over others if that puts money into a new house, I dont condone such behaviours..

Anandtech syntetic scores show that the 980 Pro is an average SSD, not sure what reviews did you read:
https://www.anandtech.com/show/16087/the-samsung-980-pro-pcie-4-ssd-review

Both SK Hynix gold and Kingston KC2500 and faster on the AT Storage bench Heavy test, I never look to those nice UP TO crazy speeds on the box, unless you call those numbers a review..
https://www.anandtech.com/show/16087/the-samsung-980-pro-pcie-4-ssd-review/4

 

I do try everything available, every few months. Still no AMD platform for me (yet).

I don't condone such behavior either, but there is little I can do about it, so I don't (and I don't waste my time worrying about it either).

I gave you a link to the reviews. AnandTech's review wasn't out yet when I posted.

Btw, I consider the AT Storage bench 'synthetic' too. That is not how workloads fly by...

See
https://pcper.com/2020/09/samsung-980-pro-pci-express-4-0-ssd-review/#ftoc-heading-6

Even when they acknowledge it, they still award it a gold.

 

I spent the better part of an hour reading the entire thread (starting from https://www.realworldtech.com/forum/?threadid=193189&curpostid=193190). Most of the specifics flew right over my head, but the consensus is that implementing AVX512 comes with significant costs.

Most obviously, it occupies transistors that could be dedicated to improving 'regular' performance. It also appears to increase power cost as well due to the larger instruction width - which may extend to the rest of the system as a frequency drop (various sources from a quick google).

The fragmented distribution of support for AVX* across the Intel spectrum also disincentivizes regular developers (rather than Matlab and co. which target a more niche market, or companies like Adobe that work closely with Intel and their army of programmers) from using compilers that build with these instructions. Also, at least right now, unless a developer writes the code by hand, compilers are not able to optimize(vectorize) code except in the most trivial of cases.

I did not see anyone challenge this consensus over the entire thread, and assuming one trusts the 'experts', it boils down to whether these AVX/AI stuff leads to tangible improvements on anything outside HPC. I'm still in the skeptics' camp regarding its general value right now, but (as Linus does mention in another post) maybe in a few years every Intel CPU has AVX-512 support and it becomes mainstream enough (i.e. the regular compilers build with it as a default).

I should mention that a common opinion floated by the programmers went along the lines of "If I find my code is slow because I'm doing a stupid thing, I don't vectorize the stupid thing; I remove the stupid thing", so no guarantees that just because MSVC and gcc and clang etc. all compile with AVX-512 by default, Intel CPUs will magically become vastly faster overnight

 

That thread? From July 2020? Obsolete information.

 

Surprise, surprise. AMD's (partial) answer to Tiger Lake. (Fighting paper launches with paper launches).

See:
https://hothardware.com/news/amd-ryzen-7-5700u-zen-3-cpu-benchmark-debut

So now, (actually for a long time...) AMD is doing what Intel has been always castrated for. But fans won't care, because, AMD.

Not only are thier true colors showing (they love their 'fans' as deep as thier pockets are, but no more), but there is barely any difference from the 4800U either. But it is a 57ooU!

An educated guess here is that AMD is swimming in Zen2 cores and needs to dump them on the unsuspecting public asap. Business as usual, nothing to see here folks, carry on.

 

No problemo... Jarrod will give u a closer look in a few days time on how
the new INTEL TIGER LAKE fares with it's targeted opponent which was frequently mentioned
during launch presentation...

...even though the storage type is not clearly indicated, but the RAM sure does matches!​

 

Probably yes, Probably no.
Yes to those whom may hv not been following on AMD's development of their CPUs. Any new makes of laptops in the market be fitted with those Renoir refreshed CPUs will still sell... only be taken by those buyers whom are new to the Ryzen performance yet, no idea that those "odd numbered" CPUs they chosen are actually "redressed" 4500u or 4700u.

Definitely a big NO to those whom have been following on Ryzen development. If i wanna get a new Ryzen, i would go for those "even numbered" 5000series makes for sure!

 

On a second thought...

https://www.notebookcheck.net/New-l...ctors-could-launch-in-late-2021.495042.0.html
...I shall wait for the Rembrandt instead as my laptop warranty hv a yr plus journey to go!

 

Lenovo ThinkPad X1 Fold hands-on




The world’s first foldable PC is now available to order from Lenovo

https://www.theverge.com/2020/9/29/...ble-pc-nano-thinkbook-5g-tiger-lake-intel-evo

 

It's not really that budget friendly...
https://www.lenovo.com/sg/en/laptops/thinkpad/thinkpad-x1/X1-Fold-G1/p/20RKCTO1WWENSG0
...@ USD$2.5k approx