Hacker News new | ask | show | jobs
by slantyyz 3524 days ago
This is probably offtopic, but I hate, hate, hate Intel's CPU naming strategy for the Core i line.

An i7 (U,M) in an ultrabook is vastly different from the i7 (MQ, HQ) in a gaming laptop.

A lot of people don't realize that, they just assume that i7 means "fastest intel processor".

It also doesn't help when computer manufacturers say "#th generation Intel processor". Unless you're keeping up with that stuff, how does a consumer buying a computer even make sense of that stuff?

5 comments

What's worse is that even if you know the exact CPU and it's performance intel has a new program to arbitrarily throttle CPUs to fit a particular power/cooling envelope.

So vendors can cheap out on cooling, cripple the cpu, and there's no way to tell until you buy it and check yourself.

Yah, and then consider the m3/m5/m7, x3/x5/x7, Pentium and celeron brands are all still available.

Its the same with the xeon line, where you really have to dig into the details to know what the difference is between a E5-46xx and a E7-48xx. Early on it seemed like intel was going to use E3=1 socket, E5=2 socket, E7>2 socket, but then they started bluring the lines. And then there are all the secondary features like extended availability of a SKU which isn't even broken down on the ARK matrix.

Then of course it changes every month or so when they add a couple new ones and remove a couple old ones. No wonder everyone is confused, half us here are computer geeks, and we need the internet to determine the difference between most of the part numbers. To average users, its complete crap. Worse, since so many of them are similar, you have to look at heaps of benchmarks to determine if some random i7 is actually faster than some random i5/whatever..

Basically, it seems to me that intel has far to many people in product segmentation/marketing and they could use some serious housecleaning. Do we really need 500 (probably worse that that) SKU's varying in one feature or another? They could rip out every other clock rate stepping and no one would really notice, heck they could rip out 1/2 the core count ones too.

And now they have the x5 and x7 atoms (formerly Cherry Trail) with things like the Z series for tablets: http://www.intel.com/content/www/us/en/processors/atom/atom-...
I agree. The intel naming convention has sucked since they switched to the Core iX names a few years back. Why are they still selling Core i5 computers in 2016 when they were available in 2012? It's like they don't want consumers to know there is a new processor out.

No wonder PC sales are so bad these days.

Because i5 and i7 processors are exactly the same silicon. It's just for i7 they turn on hyper-threading.
The important part is the generation, which, to your point, is rarely mentioned. The generation is in the name, e.g. Core i7-7xxxx parts are the newest Kaby Lake, 6xxx are Skylake, etc, but I agree that it's still confusing and requires a visit to Ark to see if it's junk or not, and only people who spend hours reading about this stuff can readily identify the differences and families.
2xxx,3xxx,4xxx,5xxx,6xxx,7xxx I do not see a problem with this naming convention.

Each generation have code name like Caby Lake etc.. but also corresponding number. You know that if computer have i7 6700 it is roughly 5-15% faster than previous generation i7 5700.

Just because Apple tend to hide this detail it is not Intel to blame.

Ok, just grabbing some scores off Passmark [1]

Intel Core i5 6440HQ -- 6390 Intel Core i5 7200U -- 4856 Intel Core i7 6650U -- 4857 Intel Core i7 6700HQ -- 8023

By looking at the numbers in the name alone, there's a wide variability in how fast they are. That's pretty confusing to someone who doesn't keep track of CPUs.

[1] http://www.cpubenchmark.net/cpu_list.php

Yeah if you compare 2 core vs 4 core you will have these results. You keep comparing different models. Sure there low voltage or mobile versions but core tech is the same in all models from given generation.

Intel Core i5-7200U vs.Intel Core i5-6200U difference is what you will expect - 10-15% in most benchmarks

There is some variation but naming convention is better than Qualcomm where 820 vs 821 do not give you any information. AMD naming conventions is also useless when they release the same chip multiple times. nVidia sometimes release new generation of chips where low end models are in fact previous generation with different names. ARM naming convention is useless as well... Intel compared to rest in industry have in fact good naming conventions.

I am not Intel fanboy, they cut huge premium on theirs chips, still have crappy drivers in Linux and release buggy chips.

I get that but I'm speaking in terms of --consumer-- perception. I can't tell you how many people tell me they think an ultrabook with an i7 must be super fast just because the ad says i7.