Intel Marketing Needs to Work on their Naming

[JrClocker]

If they would have named the new chips to match thread count like i2-, i4-, and i8- instead of i3-, i5-, and i7-...the world would have less confusion.

Now, if they next 2 digits after the "-" where frequency...that would be great! (Or any of the digits ... geese)

Hey, let's have the last 2 digits actually mean something logical too!

At least we know what the "K" at the end means...but who knows for the rest of the alphabet?

 

[EarthDog]

Can't say I was ever confused about their naming convention honestly...would your suggestion make it easier though? Perhaps.

So their line would look like this... i2xx, i4xx, i8xx, i12xx, i16xx.. Or do you go with 'real' cores? i2, i4, i6, i8....? But then there isn't a mention of HT so that could serve to confuse on that front... then again so would the i8/i12/i16 conventions...

I'm good where it is.

 

[Robert17]

At least they haven't added additional marketing adjectives like 'Super Hyper Ultra Demon Fury Killer Obtrusive Giga Mega Chip' to any of their chips.

 

[Nebulous]

At least they haven't added additional marketing adjectives like 'Super Hyper Ultra Demon Fury Killer Obtrusive Giga Mega Chip' to any of their chips.

 

[mackerel]

The current system isn't perfect, but it can be understood. At least on desktop, i3 = 2 cores with HT, i5 = 4 cores no HT, i7 = 4+ cores with HT. Then a generational number, then positional number and modifier suffixes. My only gripe with the current system is that the -E versions are given a generation number one up from what you expect e.g. Haswell = 4, Haswell-E = 5, Broadwell = 5. Maybe they should have used i9-...

3, 5, and 7 are common marketing numbers, particularly if you want to break things into 3 general ranges. Low, middle, high. BMW series is one that springs to mind and I'm sure they're not alone.

If you were to be more descriptive, I think model numbers could get long and harder to comprehend. You need to distinguish between, for example, 2 cores 4 threads, and 4 actual cores. Even clock isn't simple. Is it the base clock? Or the max turbo? How do you do generations?

So while i7-6700k isn't perfect, I think I'd prefer it over something like i4-8T40B42T6GOC if you were to try to be more descriptive. Imagine trying not to order the wrong part that might differ by one digit in that.

 

[JrClocker]

So while i7-6700k isn't perfect, I think I'd prefer it over something like i4-8T40B42T6GOC if you were to try to be more descriptive. Imagine trying not to order the wrong part that might differ by one digit in that.

Have you ordered any memory recently?



Wait...is that i7 a Haswell or Skylake...oh drat...it's Ungubukoo Island. When did they come out with the Ungubukoo series? What's the difference between an i7 Ungubukoo and DooDab series chips?

I don't know...just mainly a rant. At least on my car the model numbers make sense. While I do have to have a "secret decoder ring", the ring doesn't need a special cypher upgrade and patch every time a new product is launched.


Shrug - Something like:

ia-ccCttTbbBggGyx

a = Base model number of chip (a separate number/letter for Haswell, Skylake, etc.
cc = Number of cores
tt = Number of threads
bb = Base frequency
gg = Embedded GPU type (or 00 for none)
y = Distinguishing feature of product series
x = K = Unlocked Multiplier (or blank)

So, my 5820K would be:
iH-06C12T33B00G1K

Haswell, 6 cores, 12 threads, 3300 MHz base clock, no GPU, 1=28 PCIe lanes, K=Unlocked Multiplier


The 6700K would be:
iS-04T08T40B05G0K

Skylake, 4 cores, 8 threads, 4000 MHz base clock, 05 series GPU, 0=No specific features, K=Unlocked Multiplier


Xeon E3-1231V3 would be:
iH-04C08T34B00GS

Haswell, 4 cores, 8 threads, 3400 MHz base clock, no GPU, S=Server Grade, blank=Locked multiplier


Xeon E3-1230V3 would be:
iH-04C08T33B00GS

Haswell, 4 cores, 8 threads, 3300 MHz base clock, no GPU, S=Server Grade, blank=Locked multiplier


i7-4790 would be:
iH-04C08T36B46G0

Haswell, 4 cores, 8 threads, 3600 MHz base clock, 46 series GPU, 0=no special feature, blank=Locked multiplier



Oh...so you mean the only difference between the i7-4790 and the E3-1231V3 is the core frequency, embedded GPU, and one is server grade?



I don't know...sorry for the thread...more of a rant I guess.

 

[mackerel]

Yes I have bought ram recently, and even today I was shopping for more, although decided not to buy just yet. For shopping, I can get by from the description, although there are subtle variations that can only be seen from closer examination of the part e.g. G.Skill Trident Z reviewed often ends in GTZ, but all I see on sale are GTZB which has slightly slacker timings.

Going back to CPUs, I look at i7-6700k and know roughly what that is. i7, so probably 4C8T, 6xxx so probably Skylake (unless it is a Broadwell-E which isn't out yet), and 700 is a fairly big number so it is quite high up. If I look at iS-04T08T40B05G0K, while the info might be encoded in there, it just looks like a random string without decoding it. Suppose you consider the i7-6700 (non-K): iS-04T08T34B05G0. Not so easy to compare the lot in case there might be a digit or two other difference in there somewhere. Also there's a problem with using the generation as a letter: you can't use the same letter without confusion. Already we run into that with Sandy Bridge and Skylake. Two letters then? SB and S respectively?

How about we have both name schemes? The current scheme we call a model number, and the more descriptive scheme as a generic part number (but distinct from sku).

Right now, if someone mentions a desktop Intel CPU, I can guess most parts as long as I don't need exact clocks. I would still have to go an google it to see what it is exactly. I do struggle with low end and Xeons though, but as I'm not likely to buy either it is no loss.

 

[wagex]

bigger number newer chip, pretty simple.
6700k is newer than 2500k
6>2
i think you are way over thinking it.

 

[JrClocker]

Bigger number means newer chip is the best logic I have heard so far...thank you!

6700K is bigger than 5820K, so 6700K is newer...check...but is it better? What's the difference? That's the confusion to folks that don't follow the technology like we do.

 

[EarthDog]

The people I know look at clockspeed, not the name... well 'this one' is x.x GHz.. while 'this one' is x.x GHz...what's the difference?

AMD named their cpus back in the A64 days after intel performance and clocks...A64 2800+ was = pm 2.8ghz....etc...

 

[wingman99]

The naming scheme is fine for intel and AMD. I remember when intel only had Pentium 4 3.2GHz that was easy, everyone complained when they changed the naming scheme.