[SOLVED] i5-10600k vs. Ryzen 7 3700X, OC Performance

Solution
With that being said. which do you think would perform better, 10600k + 2070 Super vs. 3600 + 2080 Super, stock and OC on the 10600k, and PBO on the 3600? 1080p AAA
They're going to be roughly the same.
If we were to compare the 2 cpus on the same gpu, the 10600K would actually have an edge over the 3600.
But the budget from the lower cost of the Ryzen platform allows you to step up to a higher gpu tier. That's going to be a boon for the 3600 against the 10600K.
Depends on what you are doing. The 10600k does retain a performance edge in high refresh gaming if you have a high refresh monitor and high end GPU and want to chase the highest framerates you can. The 3700x will win in productivity tasks that scale across 8+ cores.

The 3700x is the better value option and probably a little more future proof in terms of games once they eventually scale up to 8 cores and 16 threads (not to mention X570 boards do offer PCI-E 4.0 support for high speed SSDs). The 10600k still has the better absolute gaming performance right now if you are trying to chase those high framerates in competitive titles.
 

Phaaze88

Titan
Ambassador
10600K, no contest.
Ryzen 3000 is just flat out not worth overclocking.

For those hung up on overcocking, Blue Team is all they should be looking at.

Ryzen 3000 on the Red Team has no OC potential. Any manual OC you could achieve on Ryzen 3000 is outdone by the following:
The higher level PBO settings + a darn good cooler. Add in memory frequencies of 3600-3800 and tightening the timings, and the gap gets even larger.
 
  • Like
Reactions: King_V and GarrettL

lolermanlols123

Commendable
Mar 2, 2018
88
0
1,640
10600K, no contest.
Ryzen 3000 is just flat out not worth overclocking.

For those hung up on overcocking, Blue Team is all they should be looking at.

Ryzen 3000 on the Red Team has no OC potential. Any manual OC you could achieve on Ryzen 3000 is outdone by the following:
The higher level PBO settings + a darn good cooler. Add in memory frequencies of 3600-3800 and tightening the timings, and the gap gets even larger.
View: https://www.youtube.com/watch?v=sFiKgkRDH4E
, I don't know how valid the results are but this video is the only video that I can find where the 3700X is OC'ed and the two seem very close in performance, what do you think?
 

Phaaze88

Titan
Ambassador
View: https://www.youtube.com/watch?v=sFiKgkRDH4E
, I don't know how valid the results are but this video is the only video that I can find where the 3700X is OC'ed and the two seem very close in performance, what do you think?
The problem with that video - heck, ANY video overclocking Ryzen for that matter, is that they're being treated in the same manner as an Intel chip.
They are not Intel cpus; the traditional all core OC is not as effective on them. Top notch cooling + high ram frequency and low timings will top any all core OC you'll get on Ryzen 3000.
Plus, they don't have the short term repercussions that comes with manually OC'ing Ryzen 3000.

You probably haven't heard about it yet, but others have managed to degrade and destroy their Ryzen 3000 cpus in a matter of months from doing the traditional Intel method of overclocking.
Why does it happen? It has to do with a basic function that the user takes away when manually OC'ing them: current regulation. This is not sustainable for daily use.
Ryzen 3000 normal operation:
Light loads: high voltage, high current, but few cores/threads are active at any time, running at the advertised max boost clock.
Heavy loads: low voltage, low current, most cores/threads are active, running up to the max boost clock depending on power and thermals

Ryzen 3000 manually overclocked:
Light loads: low voltage, high current, few cores/threads active, can no longer run at the advertised max boost; the user has sacrificed single core performance for more multithread.
Heavy loads: low voltage, high current, most cores/threads are active, none or slight multithread performance bump because of the OC.

The parts I've underlined is the problem with manual OC VS the cpu taking care of itself.

Overclocker's delight: Intel.
Everything else: AMD.
 
Last edited:
  • Like
Reactions: King_V

lolermanlols123

Commendable
Mar 2, 2018
88
0
1,640
The problem with that video - heck, ANY video overclocking Ryzen for that matter, is that they're being treated in the same manner as an Intel chip.
They are not Intel cpus; the traditional all core OC is not as effective on them. Top notch cooling + high ram frequency and low timings will top any all core OC you'll get on Ryzen 3000.
Plus, they don't have the short term repercussions that comes with manually OC'ing Ryzen 3000.

You probably haven't heard about it yet, but others have managed to degrade and destroy their Ryzen 3000 cpus in a matter of months from doing the traditional Intel method of overclocking.
Why does it happen? It has to do with a basic function that the user takes away when manually OC'ing them: current regulation. This is not sustainable for daily use.
Ryzen 3000 normal operation:
Light loads: high voltage, high current, but few cores/threads are active at any time, running at the advertised max boost clock.
Heavy loads: low voltage, low current, most cores/threads are active, running up to the max boost clock depending on power and thermals

Ryzen 3000 manually overclocked:
Light loads: low voltage, high current, few cores/threads active, can no longer run at the advertised max boost; the user has sacrificed single core performance for more multithread.
Heavy loads: low voltage, high current, most cores/threads are active, none or slight multithread performance bump because of the OC.

The parts I've underlined is the problem with manual OC VS the cpu taking care of itself.

Overclocker's delight: Intel.
Everything else: AMD.
With that being said. which do you think would perform better, 10600k + 2070 Super vs. 3600 + 2080 Super, stock and OC on the 10600k, and PBO on the 3600? 1080p AAA
 

Phaaze88

Titan
Ambassador
With that being said. which do you think would perform better, 10600k + 2070 Super vs. 3600 + 2080 Super, stock and OC on the 10600k, and PBO on the 3600? 1080p AAA
They're going to be roughly the same.
If we were to compare the 2 cpus on the same gpu, the 10600K would actually have an edge over the 3600.
But the budget from the lower cost of the Ryzen platform allows you to step up to a higher gpu tier. That's going to be a boon for the 3600 against the 10600K.
 
Solution

gtarayan

Distinguished
Mar 2, 2011
207
40
18,740
10600K is the better buy for highest fps

In absolute terms, this CPU does have an edge over AMD in fps. The problem with it is that it is attached to the 400-series chipset. I suppose prioritizing those extra 30 frames per second in World of Tanks over AMD's current diverse CPU portfolio spanning 3000 and soon to be announced 4000 Ryzen cpus appeals to some. To me, it makes no sense considering that the fps surplus constitutes a 10% increase in best case scenarios, and that increase does not "cure" an otherwise subpar performance. In other words, 250 frames per second in World of Tanks achieved on R7 3700x gives one a good experience.

Are there scenarios where an AMD CPU cannot saturate a high refresh rate monitor's capability, and that extra 10% becomes the real gravy? In those limited circumstances that priority becomes actually viable.

When it comes to 10600k, one thing I would agree with - if one somehow finds himself/herself in an unfortunate position that they must have a PC with a Comet Lake inside (no pun intended), 10600k is the best available option. Buying this CPU with a forward view of an upgrade path requires that your choice of motherboard allows for a 350W headroom for i7 and i9 cpus, and when the time for that upgrade comes, the additional costs for a massive cooling solution (and possibly better power supply) is likely to encourage one to abandon the platform all-together.

Intel is stretching this poor Skylake 14nm platform to the point there is no where else to go. I just hope the next architecture will be spectacular.
 
With that being said. which do you think would perform better, 10600k + 2070 Super vs. 3600 + 2080 Super, stock and OC on the 10600k, and PBO on the 3600? 1080p AAA
This kind of depends on the game, and how demanding it is on the graphics hardware. If one is playing something like an esports title at high frame rates, where the graphics card is limited to however many frames the CPU can process, then the 10600K's bit better gaming performance is likely to make more of difference (though how useful those extra frames will be at such high frame rates is debatable). If, on the other hand, one is playing a graphically demanding new release with the settings cranked up to ultra, then the graphics card might make more of a difference.

Personally, I think a 2080 SUPER is overkill for 1080p in today's games though, and the roughly $200 extra it costs over a 2070 SUPER amounts to around a 40% higher price, while only offering up to around 15% more performance in demanding games. You are seeing diminishing returns in terms of how much more performance you will get by paying more once you get to that point, especially at 1080p resolution, where the CPU will be limiting performance much of the time. So, it might be better to put that money toward other parts of the system, or just hold onto it to put toward a more significant graphics card upgrade some years down the line.
 

Zerk2012

Titan
Ambassador
Having used an x570 with PCIe 4.0, I wouldn't have it any other way. The storage options are very fast and not as restricted as 3.0.

Will be interesting to see what Nvidia does with 4.0.
Their going to do nothing with 4.0 it's only the amount of data that can be transfered.

For the NVME speed for most users it also means nothing unless you doing a ton of data transfers.
 
  • Like
Reactions: Phaaze88

Phaaze88

Titan
Ambassador
@GarrettL
For people who only have SSDs in their PC, then sure, there is some benefit. I see you have only SSDs in your system(y)
For people who have HDDs in the mix - the value of these for mass storage stomps SSDs - PCIE 4.0 loses face.

Also, "if it isn't broke..."
Only 1 - ONE mainstream gpu comes close to oversaturating PCIE 3.0, and that's the 2080Ti.
Yes, it oversaturates PCIE 3.0 in x8 mode by a few hundred MB/s - makes 2080Ti even more crap for SLI.
There's still the 15,760 MB/s of x16 mode. That's not getting filled up anytime soon. Nvidia has no reason to change it just yet, other than 'because people are telling them to do it'...
Even though it's been shown that there's no benefit of doing so with gpus - except with trying to SLI 2080Ti level gpus... and no one knows how much the new cards will need, and it's not like they're going to suddenly jump up a few thousand MB/s.

@King_V
I believe people are misunderstanding what overclocking is.
I see it as a separate hobby from PC building, just like custom liquid cooling. It is not a feature.
Neither is necessary, and both require the user to invest time and money into learning the ropes behind either. It is a trial and error process.
You can't just do it once and be done with it; one needs to be ready to troubleshoot should problems arise, and so on.
 
  • Like
Reactions: King_V

TRENDING THREADS