So now that all the new Ryzens are here how many cores do we actually need in 2020?

[Brazo]

Looking at cinebench single thread scores there is no meaningful difference between a 3700x and a 3950x.

So the only benefits of the ‘better’ Ryzen CPU’s are more cores(?) But can we realistically make use of those cores?

Is eight cores the new ‘four’?
Is having 16 cores only for niche use?
Isn’t eight cores simply overkill for most desktop consumer software?

 

[wazza300]

Core blimey

If software takes advantage of the cores/threads then yes

 

[Brazo]

Core blimey

If software takes advantage of the cores/threads then yes

Of cores (see what I did there) but most software can’t?

 

[Rroff]

Have to say I'm still making do with an i7 4 core / 8 thread for most stuff - though I have an additional i7 laptop which is also 4 core / 8 thread I can offload stuff to - though sometimes means focusing on one application at a time I rarely encounter a situation where I'm needing more than 6-7 CPU threads or so in one application.

I think 6/12 is still overkill for most consumer desktops but nice to have and only a small minority of games benefit from more than 8 CPU threads (most are pretty happy with 2-3 real cores and half a dozen extra SMT threads) definitely the average user isn't needing 8/16 but if and when I do upgrade the minimum step up I plan to do is 10/20.

 

[wazza300]

There must be a reason why chip makers are releasing chips with more cores more threads to the desktop user

Edit: why dont gpu's have multicore/thread chips?

 

[LeMson]

8 will become the new goto when the new consoles are out, though at moment it's prolly sensible to use a 3600x, and upgrade next year with a cheap second hand chip or new when the next ryzen proc's are out

 

[PiKe]

There must be a reason why chip makers are releasing chips with more cores more threads to the desktop user

Edit: why dont gpu's have multicore/thread chips?

Because it doesn't scale very well, just look at SLi.

 

[wazza300]

Because it doesn't scale very well, just look at SLi.

Cant they improve on it though?

 

[Rroff]

Edit: why dont gpu's have multicore/thread chips?

How do you mean? internally GPUs use lots of smaller cores and they are capable of doing something kind of like SMT. If you mean like with Ryzen with multiple packages on the same interposer then that gets more complicated as that can be beneficial to compute like and workstation rendering tasks but hard to scale up for game rendering.

Cant they improve on it though?

The only way to improve on it is to get the application developer to farm out processing their game to individual GPUs in a way that scales best for their game (explicit multi adaptor) - any other approach requires a fundamental reinvention of the GPU.

 

[Deleted member 66701]

Looking at cinebench single thread scores there is no meaningful difference between a 3700x and a 3950x.

So the only benefits of the ‘better’ Ryzen CPU’s are more cores(?) But can we realistically make use of those cores?

Oh yesssir!

Is eight cores the new ‘four’?

No, 16 is the new 8. Just 8 cores is so 2017!

Is having 16 cores only for niche use?

No, it's mainstream. Niche is 24, 32 and 64 cores (sorry Intel, you aren't invited to this party).

Isn’t eight cores simply overkill for most desktop consumer software?

No.

 

[DONOHUE07]

Still rocking a i7 4790, 4 cores and 8 threads, paired with a 1070ti, and I can see this CPU lasting another GPU upgrade, maybe even two...

When I comes to gaming, the CPU advancements are pretty awful

 

[Deleted member 66701]

Still rocking a i7 4970, 4 cores and 8 threads, paired with a 1070ti, and I can see this CPU lasting another GPU upgrade, maybe even two...

What is this mythical 4970 cpu you speak of?

 

[DONOHUE07]

What is this mythical 4970 cpu you speak of?

Woops!

Edited

 

[Rroff]

No

If someone is [actually] doing proper media creation/editing or streaming, etc. sure but for the average consumer, even ones that leave a lot of stuff running, even 4 cores 8 threads is overkill.

 

[b00merang69]

Marketing/E-Peen, Intel tried it with the GHZ race on P4 till they hit throttling (they really did see us on 6-10GHZ CPU's by now), now it is AMD with more cores than needed but hey thy are cheap.

 

[Grim5]

Cant they improve on it though?

yes they can and they will

but they will only do so when monolithic designs fail to provide performance gains which it's still doing.

when one generation to the next of graphics card drops under 10% performance boost then we'll see them push up multi core/chip designs

but as of 2019 nvidia and amd has no issues turning more transistors into more performance using existing architecture designs and without turning the gpu into a nuclear plant which is what was happening with cpus that's why they moved to dual core

 

[Rroff]

yes they can and they will

but they will only do so when monolithic designs fail to provide performance gains which it's still doing.

when one generation to the next of graphics card drops under 10% performance boost then we'll see them push up multi core/chip designs

but as of 2019 nvidia and amd has no issues turning more transistors into more performance using existing architecture designs and without turning the gpu into a nuclear plant which is what was happening with cpus that's why they moved to dual core

The problem is with how a game works you have very discrete frames of data, very small timeslices to work with and you can't just go X chip has the data I need *yoink* even with a fast interface between them as you need to know what the other GPU is doing, whether it has finished processing that data, when you can actually fetch it without causing a slowdown on that GPU as you can't just start reading from memory while the other GPU is active and the data will necessarily remain intact while you do so without the other GPU being part of the process - which inhibits its ability to do its job, etc.

Some big improvements are needed to either dynamically pool hardware resources on the fly or better farm out areas of a scene so that you don't have one GPU package sitting there almost idle rendering the sky while the other is working flat out rendering a more complex part of the scene, etc. but that still doesn't address the situations where the effect one GPU might be processing could depend on data from a previous frame that sits on another GPU or data that the other GPU is currently working on, etc. stalling processing until it is available.

There simply isn't a way to simply make something like Ryzen but for GPUs work with game rendering especially not without a complete ground up reworking of how GPUs work.

 

[Brazo]

There must be a reason why chip makers are releasing chips with more cores more threads to the desktop user

Edit: why dont gpu's have multicore/thread chips?

Because they’ve hit a brick wall in single core speed so the only way to make a ‘better’ cpu is to increase core count but what they don’t tell you is that a 3600x is as good as a 3950x 99% of the time barring of course niche applications, rendering h etc.

 

[TonTom]

I went 12 to be on the safe side.

 

[wazza300]

Whatever happened to graphine cpu's? I thought that was the next stage