10700 vs 5600X

[Zach_01]

If you don't know what you're doing, you can overheat under any condition. Such as auto voltages. A properly tuned 5600X, at STOCK settings, will not impact your gaming experience anywhere as bad as you're making it out to be.

Of course the cooler with the heatpipes will perform better, I have MULTIPLE of them and I use those effectively in numerous rigs. The main problem with the big Wraith is noise under heavier loads. Your link is also with Pinnacle Ridge counterparts...

the cooler comes with TIM pre-applied, it's a slap on solution and it DOES work, even at %100 load:

https://tpucdn.com/review/amd-ryzen-5-5600x/images/cpu-temperature.png

From TPU review of 5600X:

"We used a Noctua NH-U14S to measure the CPU temperature while running Blender. We picked an actual application as that better reflects real-life usage than a stress-testing application like Prime95."

See temperature page...

3600 with its 88W (actual) max power draw can easily run 80~85C depending case airflow. Maybe 5600X with its way better effeciency (76W actual power draw) could run like 5C lower but nowhere near that 59C TPU resulted.

 

[X71200]

Yeah I thought it was stock, but anyways, point stands really.

 

[RandallFlagg]

AMD Ryzen 5 5600X Review

Six Zen 3 cores beating eight Zen 2 Cores? That's exactly what's happening with the Ryzen 5 5600X. AMD's massive IPC gain helped it overcome a two-core deficitm, even in productivity tests. The Ryzen 5 5600X redefines what you really need for a high-end gaming PC.

www.techpowerup.com

According to the above 5600X review, the non-K series 10700 comes out ahead @ 1080p (in most games). Also beats the 10700K and 10900k (stock).

I'm just just trying to wrap my head around how this was made possible. I've always been lead to believe the K-series CPU's are "marginally faster" out of the box + the added benefit of overclocking. My kaby lake 7700K sees pretty wide single threaded perf gains over the non-K 7700.... "so vat da hell iz goin on with 10th Gen?"

Is the 10700 reference in these reviews based on a premium Z-series board, more expensive lower latency memory and overclocking (as I understand some OC is possible with non-K series)? Or was it based on stock configurations with a B/H series board and not-so expensive RAM? Also your know-how thoughts on how the non-K is able to beat intels flagship i9 SKU?

10700 doesn't have Intel's turbo boost 3.0, while the K models do. That may be part of it. As far as the max power levels and so on, within the review of the 10700 TPU breaks it down. They use the green bar for reference in future benchmarks, but you can see the effect of unlocking power limits and the 3% overclock possible on non-K chips.

Really I think the whole thing is explained by chaos theory.

Take a look at the goofy results AnandTech got below the TPU image.

Goofy AT results :

 

[_UV_]

I'm in no rush though - my 7700K is holding up to anything and everything listed in the review charts and multi-threaded inferiority isn't a big deal with my "current" requirements....

My next upgrade will be put to use for 4yrs+

my 2 cents

I7 2600 nonK + W3670 systems here in use allmost 10 years with minor upgrades like SSD, GPU, RAM and that was a breakthrough performance upgrade from Core 2 Q6600 (upgraded from Pentium D). The only thing that frustrating me - speed of Adobe Illustrator saving big files (allmost single core), anything related to encoding video handled by NVENC or CUDA. Year or two earlier i was aiming to upgrade to something 16 cores, now with current tests showing that mainstream applications still use just 8c/16t i think i will finally upgrade to 10850K and it will be fine for next 5 years. It brings for me double single core IPC and about 4 times multicore performance. Yes, 5950x will give about 10x multicore boost, but i have no tasks for it.

In your case performance boost would be between 20 and 70% depending on the task.

 

[wheresmycar]

10700 doesn't have Intel's turbo boost 3.0, while the K models do. That may be part of it. As far as the max power levels and so on, within the review of the 10700 TPU breaks it down. They use the green bar for reference in future benchmarks, but you can see the effect of unlocking power limits and the 3% overclock possible on non-K chips.

Really I think the whole thing is explained by chaos theory.

Take a look at the goofy results AnandTech got below the TPU image.


View attachment 174725

Goofy AT results :

View attachment 174726

OK this ones got me baffled lol Although it's 4K but i totally get it - "random irregularities" or just different types of games, applications, power configurations or system conditions achieving different results. The 2700X beating the 3700X and the 10700 beating the 10700K definitely puts the chaos theory to action.

Ok RandallFlag, i gotto ask - assuming you were upgrading today with no budget concerns and the targeted workload consists of the following

1. Day trading with 3 broker applications running simultaneously (a bunch of other tools/indicators/market feeds/browser windows/etc)

2. Gaming 1440p 144hz

3. Light-weight video rendering/transcoding

....which CPU would you target from the Zen 3 or 10th Gen SKUs? Overclocking isn't a requirement unless the performance returns make a meaningful difference for the intended tasks.

 

[RandallFlagg]

OK this ones got me baffled lol Although it's 4K but i totally get it - "random irregularities" or just different types of games, applications, power configurations or system conditions achieving different results. The 2700X beating the 3700X and the 10700 beating the 10700K definitely puts the chaos theory to action.

Ok RandallFlag, i gotto ask - assuming you were upgrading today with no budget concerns and the targeted workload consists of the following

1. Day trading with 3 broker applications running simultaneously (a bunch of other tools/indicators/market feeds/browser windows/etc)

2. Gaming 1440p 144hz

3. Light-weight video rendering/transcoding

....which CPU would you target from the Zen 3 or 10th Gen SKUs? Overclocking isn't a requirement unless the performance returns make a meaningful difference for the intended tasks.

If money is no object I would get a 5950X and pair it with some very high quality DDR4-3600 memory - at least from a pure performance perspective.

I'm not sure if that's really what you were asking though

I think the only weak point in the Zen 3 lineup is in the $250 - $400 range where Intel's 10700 / 10700K and sometimes 10850K (on sale) wind up staring down the 5600X. i.e. I would rather have any of those 3 than a 5600X, even for your use cases. If day trading speed is of #1 importance though, the Zen 3 is winning on web tests as well as java, which means it would probably perform better overall on the types of applications a day trader uses.

Also of note, if stability and such is of great importance, I would also go Intel for that. I'll get mobbed for saying that here, but there's no doubt in my mind that Intel is a more stable platform. It should be by now, after all it's basically just an evolution of a 5 year old platform.

 

[thesmokingman]

lmao more stable... /facepalm

....which CPU would you target from the Zen 3 or 10th Gen SKUs? Overclocking isn't a requirement unless the performance returns make a meaningful difference for the intended tasks.

A 5800x or 5900x (for padding) will more than cover that. A 5950x doesn't make sense on a cost basis unless you can actually make use of those extra cores and you won't be doing that with the workload you described.

 

[Melvis]

Apples to Apples

 

[dir_d]

Apples to Apples

Really weird that she had problems in warzone where the CPU flew in every other review

 

[phanbuey]

The only issue is that the 10600K easily hits 5.1Ghz and 48x ring, while the 5600x doesn't seem to OC very well at all so... Those percentages shrink massively as a tuned 10600K can match a 10900k at stock. At the end of the day you won't be able to tell the difference either way.

 

[RandallFlagg]

The only issue is that the 10600K easily hits 5.1Ghz and 48x ring, while the 5600x doesn't seem to OC very well at all so... Those percentages shrink massively as a tuned 10600K can match a 10900k at stock. At the end of the day you won't be able to tell the difference either way.

The OC capability of the K series Intels is why they dominate user submitted benchmark sites. 10900K's running 5.6 Ghz and 10980XE's at 5.1Ghz with DDR4-4600 dominate the top 100 spots on PCMark's site.

 

[Vya Domus]

I'll get mobbed for saying that here, but there's no doubt in my mind that Intel is a more stable platform

Yes you would because it's untrue.

 

[X71200]

Stability is not %100 relevant to how mature a platform is, at all. He doesn't even talk about the node itself, or how the VRMs react to loads. It's just pure fanboy nonsense. If he did, he would have talked about the actual chip or the boards, etc. You could make some random CPU and it could be more stable than your grossly produced Ryzen, or it could also be a lot less stable. All this is obviously related to what comes out of the wafer, the power quality you're getting from your VRMs, your power supply and such. If anything, the AMD platform actually works at far lower temperatures. There are other bits like the Intel chips being LGA sockets so you could mess that up if you like switch CPUs / re-TIM in every once a while. I've grown to come back to pin grid array after messing up couple X299 boards. The pins on the Zen chips do not as easily get FUBAR. Once my CPU got stuck to the TIM and I pulled it straight out of the socket with the water block without opening the lock. I'm still using it without any issues.

 

[Max(IT)]

If you don't know what you're doing, you can overheat under any condition. Such as auto voltages. A properly tuned 5600X, at STOCK settings, will not impact your gaming experience anywhere as bad as you're making it out to be.

Of course the cooler with the heatpipes will perform better, I have MULTIPLE of them and I use those effectively in numerous rigs. The main problem with the big Wraith is noise under heavier loads. Your link is also with Pinnacle Ridge counterparts...

the cooler comes with TIM pre-applied, it's a slap on solution and it DOES work, even at %100 load:

https://tpucdn.com/review/amd-ryzen-5-5600x/images/cpu-temperature.png

Are you serious ???
The graph you linked is a test using Noctua UH-12S, a far better cooler than the Stealth...

Yeah I thought it was stock, but anyways, point stands really.

No it doesn't.
I showed numbers about how poor the stock cooler is for gaming and rendering.

Yes, if you are mostly doing browsing and office duties, you can keep it. For everything else you need a proper cooler.

 

[Zach_01]

The OC capability of the K series Intels is why they dominate user submitted benchmark sites. 10900K's running 5.6 Ghz and 10980XE's at 5.1Ghz with DDR4-4600 dominate the top 100 spots on PCMark's site.

No doubt... but also the top 100 spots for the most power hogs of PCs...
And what about the cooling requirements for a 250~300W power draw of a chip alone?

 

[X71200]

Are you serious ???
The graph you linked is a test using Noctua UH-12S, a far better cooler than the Stealth...

No it doesn't.
I showed numbers about how poor the stock cooler is for gaming and rendering.

Yes, if you are mostly doing browsing and office duties, you can keep it. For everything else you need a proper cooler.

Stop with the nonsense.

https://eteknix-eteknixltd.netdna-ssl.com/wp-content/uploads/2020/11/25-2.jpg

I can't find a test where the stock cooler is used but the thing seems to be even more power friendly than the 3600 due to the improved nm.

Of course the *insert high core count CPU here* will dominate your poor benchmark, Randall. The benchmark you're talking about uses up to whatever core you throw at it.

 

[Max(IT)]

Stop with the nonsense.

https://eteknix-eteknixltd.netdna-ssl.com/wp-content/uploads/2020/11/25-2.jpg

I can't find a test where the stock cooler is used but the thing seems to be even more power friendly than the 3600 due to the improved nm.

Of course the *insert high core count CPU here* will dominate your poor benchmark, Randall. The benchmark you're talking about uses up to whatever core you throw at it.

again you keep posting random test without a stock cooler use.
I DID, and it showed very poor results.

 

[X71200]

again you keep posting random test without a stock cooler use.
I DID, and it showed very poor results.

The test you linked to was Pinnacle Ridge CPUs. Do you even know the difference between it and Vermeer?

 

[Max(IT)]

The test you linked to was Pinnacle Ridge CPUs. Do you even know the difference between it and Vermeer?

the test was about the three AMD coolers using the same CPU. It shows how poor the Stealth is.

 

[X71200]

the test was about the three AMD coolers using the same CPU. It shows how poor the Stealth is.

I did a test on a car with 3 tires from different companies, and got some results. I didn't do the same test on another car and see how well they do there, I don't care because I simply need to make sure my opinion looks right.

 

[Max(IT)]

I did a test on a car with 3 tires from different companies, and got some results. I didn't do the same test on another car and see how well they do there, I don't care because I simply need to make sure my opinion looks right.

you are arguing for the sake of it.
You wrote something completely wrong and you don't want to admit it.

Here you can find a Prism vs Stealth comparison on a 3600.

Enough of this.

The Stealth is GARBAGE. And the Prism is barely acceptable.

stealth vs prism

 

[X71200]

you are arguing for the sake of it.
You wrote something completely wrong and you don't want to admit it.

Here you can find a Prism vs Stealth comparison on a 3600.

Enough of this.

The Stealth is GARBAGE. And the Prism is barely acceptable.

stealth vs prism

Oh really?

https://static.techspot.com/articles-info/1871/bench/Temps-p.webp

Enough of your opiniated nonsense.

 

[Max(IT)]

Oh really?

https://static.techspot.com/articles-info/1871/bench/Temps-p.webp

Enough of your opiniated nonsense.

lol you just confirmed my point: 80° in blender is a clear indication of poor cooling in such a CPU.
I don't go past 70/72° using a Prism (that still is a modest cooler)

 

[X71200]

It's a result that came out of 1 hour of Blender, you're never going to see that temperature under normal everyday gaming, usage whatever.

I've used the Prism and the older Bulldozer cooler which is something in between the non-heatpipe coolers and the Prism, both did well on my 3600. I have never seen "barely enough" temperatures on either.

 

[RandallFlagg]

Intel "K" chips are easily overclocked 10% and many are getting 15-20% OC. I haven't heard of a non-K chip that won't take a 3% BCLK out of the box, I'm sure one exists.

AMD Zen 2 and apparently Zen 3 struggle to get 100Mhz i.e. about the 2-3% you get with a non-K chip on the Intel side. They are at their limits from the factory.

Intel chips are known to run with much higher speed RAM. DDR4-4200 is easy on Intel Z490 platforms. It is not uncommon to get DDR4-4800 on Intel. AMD struggles with speeds over DDR4-3800, some get DDR4-4133.

What this means is that Intel chips / chipsets have a significant engineering margin of error built in. AMD chips by contrast do not.


When you put a system, any system, near its limits it will not be a stable as a system that is far from its limits.

Intel also sets many standards outside of the CPU, for example Thunderbolt. If you have an AMD board, it's quite likely that board is using an Intel chip somewhere.

It's not like this is news to anyone, these things have been noted by AMD fans many times. But like I said the AMD fanboi mob would be here if anyone pointed it out in contrast to Intel, or as a plus for Intel. The emotional nonsense mob aside, what I've presented above are self-evident facts.