How are Core iX names like Core i5, i7 related to Haswell, Ivy Bridge?

Solution 1:

Ivy Bridge, Haswell, etc. are code names taken during the development period of a new CPU architecture, which often corresponds to a generation of Core-i processors (and their Pentium and Celeron product lines with specifications even lower than Core i3).

And this is where the first few exceptions arise (smile after me):

• The code name Coffee Lake is used for both 8th generation and 9th generation Core processors.

• Different code names are used for the 1st generation Core processors, including Clarkdale (mid-end Core i3 / i5, as well as Pentium and Celeron of the same generation), Lynnfield (mid-high Core i5 and i7), Bloomfield (high-end Core i7), Gulftown (Core i7 Extreme) and Arrandale (mobile CPU), and potentially more.

  Despite having different code names, their microarchitectures don't differ much.

For this part, I assume we're talking about desktop processors (mobile processors have another whole load to talk about).

As explained perfectly in other answers, Ivy Bridge is the code name for the 3rd generation Intel Core processor family, and you'd expect Core i3-3240, Core i5-3670, Core i7-3770. The first number of the 4-digit "model number" usually is the generation number, with some exceptions:

• The 1st generation Core processor doesn't have a leading 1, so instead of Core i7-1920, you have Core i7-920 instead.
• For extremely high-end processors with their second number going up to 8 or 9, they belong to the "previous generation". For example, Core i7-4820K is a 3rd generation processor and has a code name of Ivy Bridge rather than Haswell.
  • This only applies to processors with their leading number ranging from 3 to 7 (i7-3820 to i9-7980XE)

Within the same generation, all CPUs have an identical microarchitecture, and usually come with related traits, for example CPI (cycles per instruction), on-chip cache arrangement, power consumption, PCIe bus support, and ISAs (with exceptions).

The name Core i3, Core i5, Core i7 are product names for processors targeting different markets, just like Apple's iPhone XS Max, iPhone XS and iPhone XR (all of them are the same generation iPhones with different specs and prices). Within the same generation, they're identified easily by the number of cores and support for hyperthreading (HT). For example, for all Core processors ranging from 1st generation to 7th generation, Core i3 has 2 cores with HT, and Core i5 has 4 cores without HT, while Core i7 has at least 4 cores with HT.

When comparing across generations, things start to get complex:

• All Core i5 processors up to the 7th generation (Kaby Lake) have 4 cores with no HT. However, 8th generation Core i5 increased core count to 6, which is kept with 9th generation Core i5 (still no HT).
• All Core i7 processors up to the 7th generation have 4 cores with HT (8 threads), and the core count increased to 6 with 8th generation Core i7. A drastic change is that 9th generation Core i7 processors have 8 cores without HT.
• Core i9 first appeared with 6th generation Core Extreme processors (the successor of i7-6950X is i9-7960X/7980XE), replacing the top-end i7 product lines.

I think mobile processors can't just be ignored this easily.

Contrary to desktop processors, mobile ones are heavily constrained by TDP, due to considerations of power supply, cooling as well as battery life, and as a consequence mobile processors aren't separated vertically the same way as desktop ones.

The first notable difference is that the first digit of a mobile processor always corresponds to a generation (as of now). That says, even if i7-6950X is a 5th generation desktop processor (Broadwell), i7-6970HQ is still a 6th generation mobile processor (Skylake). One same thing is that the 1st generation doesn't have a leading 1, so there comes i3-330M, not i3-1330M.

The key difference between desktop and mobile is their vertical segment.

• For all processors from 3rd generation up to the 7th generation, Core i7-H and i7-M (trailing letter) have 4 cores with HT
  • 1st generation Arrandale Core i7-M has 2C/4T
  • 2nd generation Sandy Bridge Core i7-M has 2C/4T or 4C/8T, identified by the existence of a trailing Q (Quad-core, this letter is used until Coffee Lake)
• For all processors up to the 5th generation, Core i5-H and i5-M processors have 2 cores with HT. For 6th generation and 7th generation, Core i5 H/M have 4 cores without HT. This is because up to Broadwell, 4C/4T would generate too much heat and eat too much power for a i5-level laptop, but with Skylake this is possible. However to keep the line clear between i5 and i7, Intel decided to remove HT from i5 processors.
• All Core i5 and i7 have "Turbo Boost", while i3 doesn't (the data "Max Turbo Frequency" is missing for all i3).

These "standard voltage" processors don't go below 35 watts with their TDP, and there are other product lines for needs of specifically low powers. They aren't directly comparable with standard processors (targeting low power & battery endurance vs. targeting performance).

• For processors ending -U, their TDP are either 15 W or 28 W, and always comes with HT (Intel Core processors, not Pentium or Celeron).
  • For processors up to 7th generation, all have 2 cores and 4 threads
  • For 8th generation Core i5 and i7, they have 4 cores and 8 threads (yeah, evolution of technology has made this possible)
• For processors ending in -Y, their TDP don't exceed 8 W so they can fit in thinnest tablets.

And honestly, I don't see many practical differences between i5-U and i7-U, except for a minimal difference in clock frequency. More important differences aren't identified by i5/i7, however, like integrated graphics (HD Graphics / Iris / Iris Pro, all of which can appear on both i5 and i7).

To sum up, Core-iX are different vertical segments of Intel's CPU product line, while code names are chronological iterations of CPU architecture.

Solution 2:

"Haswell"/"Ivy Bridge" and so on are codenames for different generations of processors. They are probably not that useful unless you're talking very broadly of a family/generation of processors. Intel tends to use codenames during the development process of a microarchitecture, but the retail packaging and such rarely reflects that.

An easier thing to look at is how intel grades processors by generation - with the "modern" core family going from generation 1-9 so far, with the "core i 3/5/7" nomenclature. On mainstream chips, the modern core iX YYYY S nonmenclature has been a thing since the second generation core chips, with the first generation using a 3 digit code. There's a parallel set of models for HEDT, though your HEDT core i7 or 9 is a VERY different beast from your enthusiast level processor.

Intel wants things to be simple so within a generation, a core i3 is worse than a core i5 and so on.

that said you probably need to look at the entire model number for a better idea,

For example a 3770K is a third generation - so Ivy Bridge, Core I7 with an unlocked multiplier. My 9900K is a 9th generation (first digit) core i9, and a 9900KF is an unlocked version of the same chip with no integrated graphics. A U as a suffix is a ULV laptop processor and so on.

So in general, the 4 digit code is "Generation/Line/2 digits for model"

For more fun, the model numbering system is vaguely inconsistant between lines, so you might want to check intel's guide to that.

Intel's feature differenciation between models seems arbitrary - with some higher end models lacking VT-D when lower end models did, some random models having VT-X and some not having them and so on. You cannot rely on a model/line to work out the features you'd have. For example, while core i5s traditionally had hyperthreading, most of the 8th and 9th generation processors do not, other than the 9th generation 9900.

Solution 3:

What I can't seem to relate is how these are related to Intel's release of CPU code-named Haswell, Ivy Bridge?

• Ivy Bridge is the 3rd Intel Core generation
• Haswell is the 4th Intel Core generation

Haswell and Ivy Bridge are microarchitectures. Intel uses them to differentiate between different versions of their Intel Core product line.

Intel Core is a line of mid- to high-end consumer, workstation, and enthusiast central processing units (CPU) marketed by Intel Corporation. These processors displaced the existing mid- to high-end Pentium processors of the time, moving the Pentium to the entry level, and bumping the Celeron series of processors to the low end. Identical or more capable versions of Core processors are also sold as Xeon processors for the server and workstation markets.

Source: Intel Core

3rd generation Intel Core products had a part number i3/i5/i7 3xxx and 4th generation Intel Core products had a part number i3/i5/i7 4xxx

Corei5 and Corei7 are differentiated by hyper-threading.

There are other differences between those two product lines other than hyper-threading.

Solution 4:

Basically, "Core i3," "Core i5," "Core i7," etc. are used to differentiate how powerful/featureful a given product is within a given market segment and generation. These are more marketing terms than technical ones and the differentiating factors between what is marketed as i3, i5, i7, or i9 can be different between different generations or between chips targeted at different market segments within the same generation. Often, these are differentiating between different SKUs of the same chip, which have different feature sets enabled, different default clock rates, different maximum clock rates, etc.

On the other hand, Sandy Bridge, Ivy Bridge, Haswell, Skylake, Kaby Lake, Coffee Lake, etc. are successive generations of microarchitectures. Essentially, these are different generations of core designs, which newer generations generally being better than older ones within the same market segment. For example, a mobile Kaby Lake i7 (such as the i7-7500U) will generally be more powerful than a mobile Skylake i7 (such as the i7-6500U,) though not necessarily more powerful than a desktop Skylake i7 (such as the i7-6700K.)