Understanding the Economics of 6 vs. 8 CPU cores | Newbie to Gaming Industry

[Marko_BG]

I don't want to go down the rabbit hole "if 6 cores is going or not going to be enough for the next 2 or 3 years". What I would like to know, since I am new to gaming industry is the financial status of it (and if I ****** it up with new PC). I personally went with 6 cores 12 threads - a good value for a medium range CPU and used extra money on better PSU (it could have been better case or cooling or even better GPU). I think case, PSU and cooling is the only thing you "could future proof", but moving on - not the point of topic.

1. Okay, if I assume that all people that say 6 is not enough after 2 years are right and that I will need to upgrade in 2 years, does the economical or financial sheet looks like this?

2020　　　　　　→ 　　　　　　2022　　　　　　　　　　　　→　　　　　　　2024
170$ mid-end　　　　　　170$ mid-end　　　　　　　　　　　　　　　　170$ mid-end

I am assuming a top of the mid end CPU would cost around 170$ and would give performance for the next 2 years at least. In total over the 4 year period, 510 dollars.

1. Okay, if theoretically I did went with 8 cores, would economical or financial sheet look like this?

2020　　　　　　→ 　　　　　　2022　　　　　　　　　　　　→　　　　　　　2024
340$ high-end　　　　　　no upgrade (future-proof)　　　　　　　　　340$ high-end

I am assuming a top of the line high end CPU would hold itself for at least 4 years (i.e. future proofing). In total over the 4 year period, 680 dollars. Also, I assumed the price would be double, since today a 8 core CPU costs double the amount of money as 6 cores.

The way I see it by looking at this is, if you need CPU to do X work today, pay for that CPU no matter the cost. However, future proofing does not bring any benefits if your needs are satisfied with mid-range components vs. high-end components. Not to mention that future proofing brings risk because we are after all just guessing here. Also, your CPU can die after a limited 3 year warranty - so imagine being that unlucky!

Summary! if you need today a CPU for X work (streaming, gaming AAA and multitasking on 2 monitors at the same time for example), yes - go for it you need high end CPU. However, if your needs (X work) are modest (like gaming AAA games with Discord in background on one monitor) then over buying, paying premium to future proof won't bring you much more today, but will cost you more in the long run with additional risk.

As I said, I have not been following market as some of you expert guys out there. I am really new to this and I am trying to understand price per performance and price per value in the long run. I plan gaming for God knows how much and would like to have bigger picture. Thank You for sharing your knowledge.

P.S. Does this logical apply to other PC components?

 

[CuriousTomCat]

It's not just about cores. eight cores from 2017 are slower than six cores from 2020. In three years time, six core CPUs will be much faster than today's six cores. The IPC of these cores improves each year.

Performance = IPC x cores x frequency

I think the tier 5 processors will have 6 cores for a long time.

 

[Cyber-Mav]

. The IPC of these cores improves each year.

?? Skylake to whatever lake is present??

 

[eagleuk]

For anyone who isn't regularly maxing all their cores ( not running workloads ) the only thing they are really after is the better IPC.

Can't see general users needing more then 6-8 cores for a long time.

 

[CuriousTomCat]

?? Skylake to whatever lake is present??

I was referring to AMD processors. Zen2 improved IPC by 15% and then a year later, Zen3 improved IPC by 19%.

The 6 cores in my computer won't be any good in 2 or 3 years but the Ryzen 5000 definitely will be. To answer the threads question, it matters which 6 cores you're refering to.

 

[Cyber-Mav]

The 6 cores in my computer won't be any good in 2 or 3 years

thats a very short lifespan for your 6 core, my 1st gen i7 equivalent xeon still is useful today, lack of usb3 and sata3 were the main reasons to upgrade.

 

[Marko_BG]

I was referring to AMD processors. Zen2 improved IPC by 15% and then a year later, Zen3 improved IPC by 19%.

The 6 cores in my computer won't be any good in 2 or 3 years but the Ryzen 5000 definitely will be. To answer the threads question, it matters which 6 cores you're refering to.

Sorry for later reply. I was as referring to Ryzen 5 3600 or it's competition in the same rank, the i5-10400F.

 

[Marko_BG]

For anyone who isn't regularly maxing all their cores ( not running workloads ) the only thing they are really after is the better IPC.

Can't see general users needing more then 6-8 cores for a long time.

Interesting.

 

[marms]

If you upgrade to the latest mid range cpu in 2 years time you'll probably need a new motherboard and ram so it's going to be a lot more expensive than just a new cpu.

 

[CuriousTomCat]

thats a very short lifespan for your 6 core, my 1st gen i7 equivalent xeon still is useful today, lack of usb3 and sata3 were the main reasons to upgrade.

Yeah but your Xeon dates back to a time when IPC was improving by perhaps 5% per year. These days, it's more like 15% to 20%.

It's not a good thing when a product in the computer industry remains relaxant for 10 years because it means progress has stopped. I hope graphics cards never last 10 years.

 

[clv101]

It's not just about cores.

Absolutely!

 

[thebandit]

Lol what mine field of a question op. Cores definitely isn't the right question, some kind of overall performance might be a better metric (FLOPs?) and this relates to cores/threads, IPC, clock frequency, memory bandwidth etc etc.

I understand what you are trying to figure out, essentially over time what's the cheapest way to maintain mid range performance. There are probably too many variables here to ever really know this.

If you assume each year a mid range priced CPU increased performance by 10% (ie spending £150 each year would buy you an additional 10%, ignoring inflation) each 2 years it would look like this,
year 0, 100% (spend £150)
year 1, 110%
year 2, 121% (spend £150)
year 3, 133%
year 4, 146% (spend £150)

After 4 years the cycle starts again when you buy a new high end CPU, (of course in reality you would probably need a new motherboard and RAM along the way)
So essentially, you would need to get a CPU with 33% more performance on day 1 for double the cost to "break even". This might not be hard from a raw FLOPs perspective, just buy a higher core count CPU.

Its way more complicated than that of course, it entirely depends on the workload and the other components this involves. Gaming would not scale like the idealised example above, the game engine will never be perfectly threaded, ie able to balance load across all the cores, so your day 1 high end CPU would need a much higher per core performance than you are likely able to get (CPU generations generally have the same IPC and only differentiate themselves on cores and frequency). This is ignoring the GPU and resolution you game at !

I doubt anyone can actually answer your question, having just thought about it, I'm unsure wha would cost less over time. Maybe the high end to start as you would definitely not need new MB and RAM along the way, but then I doubt that would scale as well as buying new along the way in gaming workloads, forgetting resolution and GPU........ hmmmm, I have no idea

 

[Cyber-Mav]

this discussion goes all the way back to when people were saying get a dual core e6300 or a quad core q6600.
q6600 won that.