Why does Intel try to deceive people with it's mobile CPU naming schemes?

[frozentundra123456]

I didn't say Java applets were multi core and I bring it up because its something I run into at work all the time specifically because the ones I have issues with specifically cause one "thread" but really one core to run at 100% while it's also testing out how much memory it can take. All of a sudden you get into a scenario where it becomes an io battle to use anything else because one legitimate (more then a couple cycles of processing) workload turns the machine into a horrible slugfest.

This is a legitimate issue on 2 core i7's in my office. More MHz doesn't help. The HT on the non-threadlocked core doesn't help. Under these circumstances now that I realize you are right on the mobile i3's tending to have HT, there wouldn't be a difference to these users whether the CPU was titled an i7 or an i3.

Again I do understand that in stopwatch situations, especially ones where you are compiling code, encoding video, doing photoshop work or other tasks that can take time there would be a great difference in the performance of an i3 and a 2 core i7. But these jobs are so ill suited for a 2 core machine, that it would be a bad purchase for the user.

Maybe you are right, but intuitively it doesnt make sense. Lets assume that the applet takes the entire resources of one core. Are you saying that the applet would not run faster on a 3.5 ghz i7 core than on a 2.4 ghz i3 core? And are you also saying that the remaining other tasks would not run faster on the remaining 3.5 i7 core than on a 2.4 ghz i3 core? Because when you say there is no difference between an i3 and an i7, that is what you are implying. Actually, it seems to me that you may be i/o limited, or that the workload is preventing the faster cpu from maintaining turbo, in which case, I would agree, one might not see much if any difference between an i3 and an i5 or i7.

One thing I can say though, just personal observation of one instance, but my wife recently went from a laptop with a 2.4 ghz i3 to a KL 6500U (2.5 to 3.1), and the difference is *very* evident in day to day usage. So maybe in your particular use case, clockspeed does not matter, but in the way most laptops are used, I think it certainly does matter.

 

[IllogicalGlory]

I was pretty disappointed when I found about this after purchasing an "i5" laptop a few years back. It's weird that you can buy an i7-7660U and it's worse than an i5-7300HQ.

 

[frozentundra123456]

I was pretty disappointed when I found about this after purchasing an "i5" laptop a few years back. It's weird that you can buy an i7-7660U and it's worse than an i5-7300HQ.

You are comparing apples to oranges. The U vs the HQ suffix tells you that the cpus are in a different class. Five minutes on google would have told you this. Which one is "worse" depends on what you are looking for. In terms of performance the U cpu is certainly worse, but if you are looking for power efficiency in a thin and light formfactor the U cpu is the obvious choice. Really this thread in its entirety is basically a loaded question, so I originally planned not to respond at all, but finally gave in, because I did want to express my opinion.

Actually, I have no problem with intel's mobile naming scheme, if the consumer is willing to devote a few minutes of research to the topic. i3 is dual core with hyperthreading, i5 is the same with turbo, i7 U is just a slightly faster i5, and i7 HQ is a true quad core. The only thing that I think is deceptive is naming the abominable (for a laptop) atom cpus as pentium and celerons, which even in mobile used to be big core.

 

[godihatework]

Guys, I think we all know the real question is when did Intel stop beating its wife?

 

[crashtech]

...Your missing the point that a quad core isn't necessarily better than a dual core.

For instance some Cinebench R15 multi thread and single thread comparisons between mobile quad core and mobile dual core

N4200 - Quad core ST 52, MT 162

N3450 - Quad core ST 44, MT 143

i7-7600U - Dual core ST 156, MT 343

In this example the mobile dual core is twice as fast in multi thread and three times faster in single thread and the OP would have the dual core described as i3 and the quad core as i5 so how would this analogy even be useful.

Just as with desktop CPU's the server side also uses 3,5 and 7 with E3, E5 and E7 where the numbers also have nothing to do with number of cores however where 3 means cost effective it would not be unreasonable to see more lower cores in that category but that does not make 3 mean dual core.

To be fair, maybe we could stipulate that we are talking about Core CPUs; Intel doesn't have the brass to use the i(x) moniker on an Atom-based CPU... wait, do they?

 

[Ratman6161]

At work we had a group of users who insisted they needed i7s in their laptops. I explains the the systems theynwantednwere i7u's and they wouldn't notice the difference with the i5u's. They got the i5's but I was verh tempted to buy a few of these: https://www.amazon.com/Original-Intel-Inside-Sticker-888/dp/B0189GR42S

 

[IllogicalGlory]

You are comparing apples to oranges. The U vs the HQ suffix tells you that the cpus are in a different class. Five minutes on google would have told you this. Which one is "worse" depends on what you are looking for. In terms of performance the U cpu is certainly worse, but if you are looking for power efficiency in a thin and light formfactor the U cpu is the obvious choice. Really this thread in its entirety is basically a loaded question, so I originally planned not to respond at all, but finally gave in, because I did want to express my opinion.

Actually, I have no problem with intel's mobile naming scheme, if the consumer is willing to devote a few minutes of research to the topic. i3 is dual core with hyperthreading, i5 is the same with turbo, i7 U is just a slightly faster i5, and i7 HQ is a true quad core. The only thing that I think is deceptive is naming the abominable (for a laptop) atom cpus as pentium and celerons, which even in mobile used to be big core.

I am not comparing apples to oranges, I'm comparing a 7th-gen mobile i7 to a 7th-gen mobile i5. That's what Intel is telling me they are; why would I even consider that the i7 might have lower performance? Why would the trailing letter(s) really demonstrate the hierarchy? I don't know if it's deceptive, or if it's lazy or what, but it makes very little sense.

I'd posit that it's more on the deceptive side than anything; Intel knows the power of the i7 brand. People want i7s in their machines, and Intel is only too happy to provide rebranded 15W i3s and hope they make the sale before we realize. They should all be called i3-7XXXU. Don't try to sell people i5 and i7 performance in laptops that don't and aren't meant to have it. AMD isn't doing much better either, calling some of their 4/8 mobile CPUs R7-2XXXU, when they should just be R5-2XXXU; that would make things so much more transparent.

It's also interesting that Intel aren't calling any of their Kaby Lake refresh CPUs "i3". They're happy to brand 2/4 CPUs as i7, but seem hesitant to brand 4/8 CPUs as i3. If it's simply their special naming scheme for low power SKUs, you'd think it wouldn't change even if the core count does.

I do however understand that it might be the OEMs insisting that they do this; AMD pulled similar stuff with their ridiculous R9 370/R9 360 OEM desktop designs as well, but whether it's OEMs or Intel/AMD, someone is trying to deceive you. The naming scheme doesn't make sense.

 

[Topweasel]

FWIW that's my general experience as a third party entity with average laptop buyers too. I remember one guy telling me he just bought the latest laptop and how great it was. Asked what processor and he didn't know so looked in the properties and it was an older generation CPU that had been superseded for nearly a year already. Maybe I shouldn't have said anything as he was genuinely ecstatic with it until that point. :/




Your missing the point that a quad core isn't necessarily better than a dual core.

For instance some Cinebench R15 multi thread and single thread comparisons between mobile quad core and mobile dual core

N4200 - Quad core ST 52, MT 162

N3450 - Quad core ST 44, MT 143

i7-7600U - Dual core ST 156, MT 343

In this example the mobile dual core is twice as fast in multi thread and three times faster in single thread and the OP would have the dual core described as i3 and the quad core as i5 so how would this analogy even be useful.

Just as with desktop CPU's the server side also uses 3,5 and 7 with E3, E5 and E7 where the numbers also have nothing to do with number of cores however where 3 means cost effective it would not be unreasonable to see more lower cores in that category but that does not make 3 mean dual core.

Wow that is pretty ingenious. Comparing a 4 core Atom setup to a Dual core Kaby Lake. Though that said for the same reasons as I have noted My 4 core S3 is generally more responsive on a general basis than my 2 core i3. But take a slow 4 core and a really fast Dual core from the same arch and gen and the results won't be so lopsided. But out right "speed" isn't the only and honestly is a pretty small reason for why the modeling scheme of the mobile products is a little annoying. That said its really pointless. Sure it can be as a annoying as a BMW to figure out what you are getting by model. But there is some method and reasoning behind it and it's not like Intel hides what is what.

 

[Dufus]

If people are going to base their mobile buying decision on number of cores then that's what they might just end up with.

Intel have done some things that I don't agree with like when they came out with the 'm' line and compared them to the 'U' by crippling it to the same package power limit and claiming how much faster the 'm' was than the 'U'. Nearly everybody else thought it was great news though. :/

 

[DominionSeraph]

Intel's mobile customers are OEM's. Please post evidence they are confused as to what they're getting.
Can't say I've seen an OEM have difficulty with Intel, but I saw Dell advertising a system with an AMD FX-8800P as having 8 cores.

 

[ElFenix]

That was my understanding. Intel leveraging the psychology of luxury brands from a different product domain.

Does this mean intel is going to split up the i3 so they can sell an ugly 4 door as a "coupe"?

 

[crashtech]

Intel's mobile customers are OEM's. Please post evidence they are confused as to what they're getting.
Can't say I've seen an OEM have difficulty with Intel, but I saw Dell advertising a system with an AMD FX-8800P as having 8 cores.

I'm a little confused about your post, I for one am imagining consumers walking into Best Buy and choosing something partially based on the "i3," "i5," or "i7" sticker. I don't think the gist of the OP has anything to do with OEMs at all.

 

[DominionSeraph]

Intel's mobile customers are OEM's. Please post evidence they are confused as to what they're getting.

I'm a little confused about your post, I for one am imagining consumers walking into Best Buy and choosing something partially based on the "i3," "i5," or "i7" sticker. I don't think the gist of the OP has anything to do with OEMs at all.

Intel isn't selling you the mobile processor, they sell it to the OEMs. What's hard to understand that you're not buying from them?
If the OEMs are buying i7's thinking they're quads they can take it up with Intel. You don't enter into that transaction.

 

[crashtech]

Intel isn't selling you the mobile processor, they sell it to the OEMs. What's hard to understand that you're not buying from them?
If the OEMs are buying i7's thinking they're quads they can take it up with Intel. You don't enter into that transaction.

So you are asserting that the onus is on the OEMs to ensure Intel's CPU naming conventions are not misleading?

 

[IllogicalGlory]

Intel isn't selling you the mobile processor, they sell it to the OEMs. What's hard to understand that you're not buying from them?
If the OEMs are buying i7's thinking they're quads they can take it up with Intel. You don't enter into that transaction.

That's not exactly it. The OEMs of course know what they're buying, but it's in their interest (the OEMs, not Intel) that customers think they're buying something great when they buy an i7, even if it's a repackaged i3. The deception works mostly in the OEM's favor, enabled, perhaps by necessity, from Intel and others. If Intel named all of their U-series "i3-7XXXU", OEMs would likely just buy the lowest numbered ones because your average consumer doesn't care about the "XXX" and the OEMs can get them for cheaper. Call it "i7" instead and people start to think they're getting a big upgrade and the OEMs can sell more expensive products that people will pay for.

Intel isn't the only one that does this either. For example, OEM only SKUs in the Rx 300 series of graphics cards from AMD include "R9 360" (non OEM counterpart: R7 360) and "R9 370" (non OEM counterpart: R7 370); they bumped up the number meaninglessly from R7 to R9 to make it more attractive to OEMs, who find it easier to sell "R9 graphics" than "R7 graphics".

 

[DominionSeraph]

So you are asserting that the onus is on the OEMs to ensure Intel's CPU naming conventions are not misleading?

Who says the OEMs find them misleading? Do you really think they're buying millions of these not knowing what they are?

What you should be complaining about are the laptop model names since that's what you buy. What's a "ThinkPad T61"? Sounds like a headrest for a tank.

 

[crashtech]

Who says the OEMs find them misleading? Do you really think they're buying millions of these not knowing what they are?

What you should be complaining about are the laptop model names since that's what you buy. What's a "ThinkPad T61"? Sounds like a headrest for a tank.

Hmm, I think we are just talking past each other at this point. Intel's CPU names are used as selling points to consumers. Of course OEMs know what they are putting in the machines they're building. They may be complicit, but the point that Intel's CPU naming is inconsistent and perhaps even misleading to consumers remains.

 

[slashy16]

I never realized how evil Intel was. Has anyone in this thread sacked up and decided to sue Intel hurting your feelings?

Trolling and Threadcrapping are not allowed
Markfw
Anandtech Moderator

 

[frozentrundra]

I am not comparing apples to oranges, I'm comparing a 7th-gen mobile i7 to a 7th-gen mobile i5. That's what Intel is telling me they are; why would I even consider that the i7 might have lower performance? Why would the trailing letter(s) really demonstrate the hierarchy? I don't know if it's deceptive, or if it's lazy or what, but it makes very little sense.

I'd posit that it's more on the deceptive side than anything; Intel knows the power of the i7 brand. People want i7s in their machines, and Intel is only too happy to provide rebranded 15W i3s and hope they make the sale before we realize. They should all be called i3-7XXXU. Don't try to sell people i5 and i7 performance in laptops that don't and aren't meant to have it. AMD isn't doing much better either, calling some of their 4/8 mobile CPUs R7-2XXXU, when they should just be R5-2XXXU; that would make things so much more transparent.

It's also interesting that Intel aren't calling any of their Kaby Lake refresh CPUs "i3". They're happy to brand 2/4 CPUs as i7, but seem hesitant to brand 4/8 CPUs as i3. If it's simply their special naming scheme for low power SKUs, you'd think it wouldn't change even if the core count does.

I do however understand that it might be the OEMs insisting that they do this; AMD pulled similar stuff with their ridiculous R9 370/R9 360 OEM desktop designs as well, but whether it's OEMs or Intel/AMD, someone is trying to deceive you. The naming scheme doesn't make sense.

Those kaby lake refresh cpus are all quad cores. Quad mobile big core cpus have always been i7. They actually offer a new level of multi threaded performance for mobile. It would make no sense at all to call them i3, or even i5 for that matter.

 

[IllogicalGlory]

Those kaby lake refresh cpus are all quad cores. Quad mobile big core cpus have always been i7. They actually offer a new level of multi threaded performance for mobile. It would make no sense at all to call them i3, or even i5 for that matter.

I don't disagree with you, but I think that line of reasoning also implies that 2C/4T shouldn't be called i5 or i7. That was the point I was trying to make in bringing them up.

 

[Dufus]

Quad mobile big core cpus have always been i7.

I have already posted and shown that is not the case. You could also add E3-15xxM series too such as E3-1535M v5
which are mobile quad core "3" series.

People have got to get over these poorly preconceived ideas, well they don't have to but it would be nice IMHO if they did.

 

[coercitiv]

According to the reasoning expressed in this thread, Intel is currently marketing 6 core CPUs as 4 core since they are using the same i7 branding. Just imagine the poor customer going back home and discovering his CPU has 2 more cores than what says on the label.

How can it be deceit when the same branding also includes core count advances?

 

[svenge]

The only problem I see is people naively trying to extrapolate so-called naming "rules" that have historically applied solely to Intel's mainstream desktop CPU lines and blindly assign them to laptops. It's not like Intel is hiding the physical core counts on any given CPU, as they're prominently displayed everywhere on ARK, even in the highly-abbreviated "generational" info pages.

A basic rule of thumb is that you shouldn't be able to complain after the fact about any pre-made computer's CPU specs if you can't even be bothered to look into its most basic details first. If you can't see beyond mere marketing names (e.g. Core i3, Ryzen 5) when making decisions involving hundreds if not thousands of dollars worth of product, the hard truth is that you're the problem, not the manufacturer.

On a side note, I don't see anyone complaining about Ryzen 5 model numbering being confusing despite it covering both 4c/8t and 6c/12t models while its Ryzen 3 (4c/4t) and Ryzen 7 (8c/16t) brethren are more straight-forward in terms of core/thread count...

 

[John Carmack]

Now according to the latest leaks posted on AT, a mobile 6c/12t CPU can be i7 or i9 even if they both have the same cache and TDP?

 

[Topweasel]

The point was never about us. I don't agree with the theory that just because someone doesn't know what don't know a company should be able pretty on their lack of knowledge. It's that superiority complex that makes a lot of IT people so hard to deal with.

As for AMD they probably shouldn't get a free pass. But they have the benefit of lack of expectations. R7 is fully featured, R5 is disabled cores with SMT, R3 is disabled cores without SMT. They also aren't selling a $80 CPU as a $300 one.

I know why Intel does it and as they continue to grow the number of QC NB i7's, including some 15w ones. Hopefully they will pull farther away from the DC i7's. But it's the existence of both the DC i7's and QC i7's and the fact that the DC i7 is the exact same chip as the NB and DT i3 that is the issue.