CPU/Mobo Comparing CPUs across generations

calvin1719

Mostly harmless.
Adept
I recently had a conversation with a member here about an i7 4770 for 7k, vs a new i3 9100f for a little more than 7k.

These 2 benchmark about the same (passmark), the i7 being a tad higher, but can someone tell me they'd compare CPUs across gens? I'd think that improvements in architecture and efficiency would mean that the i7 wouldn't be comparable to a 9th/10th gen i7, with lower tier CPUs achieving the same performance as previous gen higher tiers.

If not the benchmarks, what else do you look at?
 
Using Passmark as a benchmark to judge CPUs is pretty useless. Use gaming benchmarks for comparison, as it's the most easiest way to know the better one.

However, you must remember that the i7 4770K/non-K exhibits:
  • Slower DDR3 RAM usage.
  • Runs hotter and consumes more power.
  • A framework that is outdated, especially in terms of motherboard features such as M.2 support.
 
This is is how I will judge...

How recent is the release (of course wouldn't opt for recent i3 and ditching some old i7 but investment should be worthy considering how long you will use the used stuff you buying.

Secondly, wattage, how much is it gonna consume.
Cores, threads, L1234 etc. caches
Overall scores.
Game fps.
Finally the price I'm getting it for else any day the latest Ryzen makes sense if you not getting the old proccy for a throwaway price!
 
Last edited:
Use gaming benchmarks for comparison
I don't expect comparative or comparable benchmarks are available for CPUs 5 or 6 gens apart.

A framework that is outdated, especially in terms of motherboard features such as M.2 support.
Yes, you'll definitely run into features that were only introduced with later gens when comparing old CPUs, but my question is mostly about CPU perf itself, all else being equal.

of course wouldn't opt for recent i3 and ditching some old i7
Not sure which one you're advocating for here. Or are you saying if you already have an old i7 don't ditch it for a new i3?

cores, threads, L1234 etch. caches
This is where the architecture improvements come into the picture IMO. What good is more cores and threads if a newer architecture can do more with less cores/threads?

There will of course be a difference in some applications that are highly multi-threaded and can benefit from more threads, but how do you compare that or quantify apart from benchmarks?
 
Not sure which one you're advocating for here. Or are you saying if you already have an old i7 don't ditch it for a new i3?
Of course. I still run my first gen i7 2600 non K and it still kicks ass on games and for my AIO needs!
I was to upgrade to Ryzen just last month but the capital was spent on home reconstructions..
 
Repeat after me :- buying used intel 3rd/4/5/6/7 cpu doesn't make sense anymore unless they are very very very hmm verrry cheap
I3 10100f (lol oof) for 7.5k = i7 7700 in performance
Also old motherboard can die as they have less life as compared to a cpu
 
Repeat after me :- buying used intel 3rd/4/5/6/7 cpu doesn't make sense anymore unless they are very very very hmm verrry cheap
I3 10100f (lol oof) for 7.5k = i7 7700 in performance
Also old motherboard can die as they have less life as compared to a cpu
This was my opinion as well, but isn't the point of the thread. The question here is effectively comparing, quantitatively, CPU performance across generations.
For example, what quantitative basis do you have for your opinion that 10100f is equivalent to an i7 7700?
 
I will be honest with you and i'm sure others are confused by your statement as well. You keep on asking for a quantitative basis and that literally is benchmarks. I must have either misunderstood your question or you haven't phrased it very well. The reason a 10100f is said to be equal to an i7 7700 is due to the fact that the numbers that it gives out are relatively the same and hence it is believed that the 10th gen cpus i3 tiers have caught up to their i7 counterparts from 7th gen.
And don't trust any of those "benchmark" websites, 90% of them are bullshit.
 
Alright, maybe I miscommunicated a bit.

IMO, the only way to compare is benchmarks. In my discussion with another member, I brought up the point that the 9100f/10100f have the same benchmark numbers that the 4770 does, and their response was that the 4770 was still better because some CPUs are just better (paraphrasing, but that's the gist).

That leads to my question, that in general, is there any other measure that can be used to compare other than synthetic benchmarks? Given that fps benchmarks for the same titles and with the same GPU hardware probably won't be available for CPUs released far apart.

And don't trust any of those "benchmark" websites, 90% of them are bullshit.
I agree with this, but could you also expound on why you think so?
For example, I believe userbenchmark.com gives way too much weightage to stuff like age and popularity rather than just performance.
 
The i7 4770K/non-K uses DDR3 RAM. i3 9100F runs on DDR4.
Ah, I thought you meant slower than other DDR3 processors. Obviously the DDR3/DDR4 difference would matter.

Like I said above, IMO comparing the specs themselves does little to put things into perspective, because what is bad or good or great for one architecture may not be so for a newer architecture.

Also, I don't necessarily want to restrict this discussion to just these 2 chips.
 
The board point made above is very true, lifespan will be short. But just my experience, since my board died, I have moved from the i7 3770 to a 5600x using the same 1080Ti and I cant see ANY difference to the variety of games (CP2077, Forza 4, SOTTR) at 1080p.
 
U can compare bench marks, fps counters , whateva u want whichever way u want i3 10100f is equal to older gens i7(7 and below)
U guys don't understand a very simple thing , intel has a basic architecture for its cpu it's based on a 'ring-bus' config and it has made it's cpu on similar layout design ever since 2nd gen(Sandy bridge) hence very meagre ipc gains for sucessive gens has been seen ever since , you pick any intel cpu across generation keeping same core/thread count and same clock(say 3 ghz) they will perform in about 10% difference even though they have decreased node width to 14nm since 2nd gen
On the other hand pick a kaveri/bulldozer and compare it to zen architecture ull understand what I'm talking bout
Also I wud like to debunk a very popular myth , they difference between DDR3 and ddr4(single channel) is not enough to be noticeable on a gaming setup provided the game runs atleast 60fps
Alright, maybe I miscommunicated a bit.

IMO, the only way to compare is benchmarks. In my discussion with another member, I brought up the point that the 9100f/10100f have the same benchmark numbers that the 4770 does, and their response was that the 4770 was still better because some CPUs are just better (paraphrasing, but that's the gist).

That leads to my question, that in general, is there any other measure that can be used to compare other than synthetic benchmarks? Given that fps benchmarks for the same titles and with the same GPU hardware probably won't be available for CPUs released far apart.


I agree with this, but could you also expound on why you think so?
For example, I believe userbenchmark.com gives way too much weightage to stuff like age and popularity rather than just performance.
You can just give it a rest by saying that i3 loloof is better than 4770 that's it , he doesn't have a reason to say why 4770 is better because he doesn't have one so he just proves his point by saying that some cpu are just better.
The board point made above is very true, lifespan will be short. But just my experience, since my board died, I have moved from the i7 3770 to a 5600x using the same 1080Ti and I cant see ANY difference to the variety of games (CP2077, Forza 4, SOTTR) at 1080p.
That's because you are still using 1080ti, upgrade to something better and u will get better noticable differences
Alright, maybe I miscommunicated a bit.

IMO, the only way to compare is benchmarks. In my discussion with another member, I brought up the point that the 9100f/10100f have the same benchmark numbers that the 4770 does, and their response was that the 4770 was still better because some CPUs are just better (paraphrasing, but that's the gist).

That leads to my question, that in general, is there any other measure that can be used to compare other than synthetic benchmarks? Given that fps benchmarks for the same titles and with the same GPU hardware probably won't be available for CPUs released far apart.


I agree with this, but could you also expound on why you think so?
For example, I believe userbenchmark.com gives way too much weightage to stuff like age and popularity rather than just performance.
We'll stop trying to find different ways to compare cpu and start by comparing them on the basis of your intented usage . So if u are gaming compare on fps counters, if u want productivity compare Synth bench , if it's electeicity compare TDP, if it's some workload that benefits from new instructions sets included between generation consider that and that'll be it for you to find your answer
 
Last edited:
We'll stop trying to find different ways to compare cpu and start by comparing them on the basis of your intented usage . So if u are gaming compare on fps counters, if u want productivity compare Synth bench , if it's electeicity compare TDP, if it's some workload that benefits from new instructions sets included between generation consider that and that'll be it for you to find your answer
isn't that going to be apples to oranges comparison; cause if you see Apple has its own set of merits and demerits against Orange and vice versa.
 
Buddy way out of context man, he initially asked for *different* ways to compare cpu , he asked for it so here I make sure he gets a fruit basket with them all apples oranges and banana and all of em
But then I pointed that it's better compare apples to apples so I told him to compare them on his intented usage
So by that I actually told him to narrow it down to his personal use.
See I am confused here
isn't that going to be apples to oranges comparison; cause if you see Apple has its own set of merits and demerits against Orange and vice versa.

Or maybe I just dint understand ur question alltogether
 
But then I pointed that it's better compare apples to apples so I told him to compare them on his intented usage
So by that I actually told him to narrow it down to his personal use.
thats what I meant to be comparison of apples to oranges.

Many people buy hardware that will work in tandem with the existing hardware they have. Like if anyone has 3rd/4th gen motherboard, couple of DDR3 RAMs then this comparison makes no sense as the most suitable option would be to use the 3rd or 4th gen processor and upgrade cost will be minimal.

If you invoke the comment stating 3rd gen/4th gen processors and compare it with 10th gen or upcoming 11th gen on price point - it makes no sense as the person has to buy whole lot of hardware and overall cost of upgrade to go high anytime.
 
Back
Top