Ryzen 7 2700X vs. Core i7-8700K: 35 Game Benchmark

[Sausagemeat]

Just like the video states there will always be fringe cases when working with high temps.
And the temps of the CPU are not the only thing you should worry about, your mobo too and the components that are around your CPU (like the VRM in some cases).
You are also forgetting something: with time dust and the aging of the thermal paste can and will raise your CPU's temps by a few degrees. And your room's temp will not always be low. You need a good temp buffer when you OC your CPU.

I agree. But running modern Intel CPUs between 70-90 degrees doesn’t really cause any problems. The vast majority of users are running at these temps at stock on the crappy stock cooler. From my experience, CPUs are pretty solid, they don’t fail very easily or very often.

The fact is, faaaq is correct, most users are hitting 5.0ghz with air, on coffee lake without deliding and running at reasonable temps that aren’t going to cause problems. When people say that most can’t get 5.0 ghz or that you have to delid the CPU it’s just a desperate attempt to put misinformation out there about the current coffee lake CPUs.

 

[Puiu]

I agree. But running modern Intel CPUs between 70-90 degrees doesn’t really cause any problems. The vast majority of users are running at these temps at stock on the crappy stock cooler. From my experience, CPUs are pretty solid, they don’t fail very easily or very often.

The fact is, faaaq is correct, most users are hitting 5.0ghz with air, on coffee lake without deliding and running at reasonable temps that aren’t going to cause problems. When people say that most can’t get 5.0 ghz or that you have to delid the CPU it’s just a desperate attempt to put misinformation out there about the current coffee lake CPUs.

I have no idea where you are getting this "most users" thing. The majority of posted stable OC results using air cooling are 4.8-4.9GHz (check forum posts like guru3d, hardocp, techspot, overclock, etc and dedicated 8700k OC reddit posts). This is true for both user posts and OC guides. People have to use good liquid cooling or get extremely lucky and win the lottery to get 5.0GHz.
On a side-note: nobody is saying that 4.9GHz is not a great result for the 8700k on air - it's an easy to OC CPU.

 

[Sausagemeat]

I have no idea where you are getting this "most users" thing. The majority of posted stable OC results using air cooling are 4.8-4.9GHz (check forum posts like guru3d, hardocp, techspot, overclock, etc and dedicated 8700k OC reddit posts). This is true for both user posts and OC guides. People have to use good liquid cooling or get extremely lucky and win the lottery to get 5.0GHz.
On a side-note: nobody is saying that 4.9GHz is not a great result for the 8700k on air - it's an easy to OC CPU.

You are wrong of course. 4.9ghz is a poor result. I’ve just had a scan through he forums, dozens of people hitting 5ghz on air. The “extremely lucky” people are getting 5.2ghz. I also have two mates who have bought the 8700K and both hit 5.0 ghz, one with a watercooler the other with a Noctua NHD15. I guess I have “lottery winning” friends?

We are of course splitting hairs here. But it’s a misleading statement to claim that hitting 5.0 ghz is “extremely lucky” or akin to “lottery winning”. Sounds like you’re attempting to downplay the facts to me. I wonder why? Lol.

Silicon lottery claim that 86% of 8700Ks will hit 5.0 ghz or higher. Since when did 86% of people win the lottery?

https://siliconlottery.com/collections/coffeelake/products/8700k50g?variant=224965885964

Looks to me like you’re in a small minority if you can’t hit 5.0 ghz.

 

[Puiu]

You are wrong of course. 4.9ghz is a poor result. I’ve just had a scan through he forums, dozens of people hitting 5ghz on air. The “extremely lucky” people are getting 5.2ghz. I also have two mates who have bought the 8700K and both hit 5.0 ghz, one with a watercooler the other with a Noctua NHD15. I guess I have “lottery winning” friends?

We are of course splitting hairs here. But it’s a misleading statement to claim that hitting 5.0 ghz is “extremely lucky” or akin to “lottery winning”. Sounds like you’re attempting to downplay the facts to me. I wonder why? Lol.

Silicon lottery claim that 86% of 8700Ks will hit 5.0 ghz or higher. Since when did 86% of people win the lottery?

https://siliconlottery.com/collections/coffeelake/products/8700k50g?variant=224965885964

Looks to me like you’re in a small minority if you can’t hit 5.0 ghz.

Yeah, I'll stick with the majority hitting 4.9GHz on air cooling. I really don't know what you've been reading but official forums are more trustworthy than you.
FYI the link you've given pretty much confirms what I've said. The 86% number is achieved with delidding and water cooling. They even linked you the coolers used (the QVL link) and the that they use thermal paste underneath the IHS (Thermal Grizzly Conductonaut). You kinda imploded there with that link

 

[Evernessince]

You are wrong of course. 4.9ghz is a poor result. I’ve just had a scan through he forums, dozens of people hitting 5ghz on air. The “extremely lucky” people are getting 5.2ghz. I also have two mates who have bought the 8700K and both hit 5.0 ghz, one with a watercooler the other with a Noctua NHD15. I guess I have “lottery winning” friends?

We are of course splitting hairs here. But it’s a misleading statement to claim that hitting 5.0 ghz is “extremely lucky” or akin to “lottery winning”. Sounds like you’re attempting to downplay the facts to me. I wonder why? Lol.

Silicon lottery claim that 86% of 8700Ks will hit 5.0 ghz or higher. Since when did 86% of people win the lottery?

https://siliconlottery.com/collections/coffeelake/products/8700k50g?variant=224965885964

Looks to me like you’re in a small minority if you can’t hit 5.0 ghz.

I believe you are both right. Silicon Lottery is in the business of binning their chips to their maximum potential at 1.40 - 1.45v. A majority of their i7 8700K CPUs likely reach 5.0 GHz because 1) They are using high voltage 2) They delid 3) They use high end cooler. If you take those factors away you are looking at completely different numbers.

As you said though, a majority of i7 8700Ks can hit 5.0 GHz and that statement is completely correct. What I think Puiu was trying to say is that the average user won't get the same probability as Silicon lottery because they are not Deliding, using high voltage, or using the most expensive cooler they can get their hands on.

 

[Outlander04]

It’s fine mate. The failure rate doesn’t really significantly go up until you hit 100C.

You mean intel lied to us for the last 20 years when they didnt want us running chips above 70 C ?

 

[cliffordcooley]

It’s fine mate. The failure rate doesn’t really significantly go up until you hit 100C.

You mean intel lied to us for the last 20 years when they didnt want us running chips above 70 C ?

Here is what Intel said back during the Core2 days. That is right smack in the middle of your time frame.

https://www.intel.com/content/dam/w...sheets/core2-qx9000-q9000-q8000-datasheet.pdf

In the event of a catastrophic cooling failure, the processor will automatically shut down when the silicon has reached a temperature approximately 20 °C above the maximum TC. Assertion of THERMTRIP# (Thermal Trip) indicates the processor junction temperature has reached a level beyond where permanent silicon damage may occur. Upon assertion of THERMTRIP#, the processor will shut off its internal clocks (thus, halting program execution) in an attempt to reduce the processor junction temperature. To protect the processor, its core voltage (VCC) must be removed following the assertion of THERMTRIP#. Driving of the THERMTRIP# signal is enabled within 10μs of the assertion of PWRGOOD (provided VTT and VCC are asserted) and is disabled on de-assertion of PWRGOOD (if VTT or VCC are not valid, THERMTRIP# may also be disabled). Once activated, THERMTRIP# remains latched until PWRGOOD, VTT or VCC is de-asserted. While the de-assertion of the PWRGOOD, VTT or VCC signal will de-assert THERMTRIP#, if the processor’s junction temperature remains at or above the trip level, THERMTRIP# will again be asserted within 10μs of the assertion of PWRGOOD (provided VTT and VCC are valid).

So no Intel didn't lie, when they wanted you to run at an optimal temperature. They allowed a 20C tolerance and even then they state may cause permanent damage. The 70C temp is optimal max conditions not extreme max conditions. Intel doesn't want anyone taking their CPU to extremes, hint the 20C tolerance. Nothing about that is a lie.

"In the event of a catastrophic cooling failure, the processor will automatically shut down when the silicon has reached a temperature approximately 20 °C above the maximum TC. Assertion of THERMTRIP# (Thermal Trip) indicates the processor junction temperature has reached a level beyond where permanent silicon damage may occur."

 

[Sausagemeat]

Yeah, I'll stick with the majority hitting 4.9GHz on air cooling. I really don't know what you've been reading but official forums are more trustworthy than you.
FYI the link you've given pretty much confirms what I've said. The 86% number is achieved with delidding and water cooling. They even linked you the coolers used (the QVL link) and the that they use thermal paste underneath the IHS (Thermal Grizzly Conductonaut). You kinda imploded there with that link

No 86% is the number of all chips they have had in ever.

Stop spreading misinformation. The majority of users out there will hit 5ghz.

 

[Puiu]

No 86% is the number of all chips they have had in ever.

Stop spreading misinformation. The majority of users out there will hit 5ghz.

Sure dude, if you want to believe in fantasy stories then be my guest. There is nowhere on that website that says that the 86% number was achieved without water cooling or even delidding.

"This CPU is guaranteed stable when using the settings below and matching components from our QVL. We go through a rigorous stress test routine to ensure stability for the vast majority of use cases. "
- their own testing methodology specifically states that they are using water coolers.

I'll continue to believe in simple facts that can be seen everywhere in pretty much every tech forum thread and even dedicated 8700k OC reddit threads where people post their results. I think that you didn't read the website before you linked it and just got burned.

 

[Sausagemeat]

Sure dude, if you want to believe in fantasy stories then be my guest. There is nowhere on that website that says that the 86% number was achieved without water cooling or even delidding.

"This CPU is guaranteed stable when using the settings below and matching components from our QVL. We go through a rigorous stress test routine to ensure stability for the vast majority of use cases. "
- their own testing methodology specifically states that they are using water coolers.

I'll continue to believe in simple facts that can be seen everywhere in pretty much every tech forum thread and even dedicated 8700k OC reddit threads where people post their results. I think that you didn't read the website before you linked it and just got burned.

No, the site just claims all CPUs tested, doesn’t say anything about delidding or watercooling, they just say every CPU they have tested, I mean by your logic you think they delid and watercool every single CPU they test, they might do but then again you think they might mention this. Let’s agree to disagree, you can live in a fantasy world and believe most chips won’t hit 5.0 ghz and I will believe the people who actually test large amount of chips alongside the posts of multiple users across multiple forums on the Internet and believe that most chips will hit 5.0 ghz.

Would you like me to spam you with links to forums where most users are claiming they hit 5.0 ghz or higher? Because I’m almost certain you must only be reading the AMD subreddit...

 

[Puiu]

No, the site just claims all CPUs tested, doesn’t say anything about delidding or watercooling, they just say every CPU they have tested, I mean by your logic you think they delid and watercool every single CPU they test, they might do but then again you think they might mention this. Let’s agree to disagree, you can live in a fantasy world and believe most chips won’t hit 5.0 ghz and I will believe the people who actually test large amount of chips alongside the posts of multiple users across multiple forums on the Internet and believe that most chips will hit 5.0 ghz.

Would you like me to spam you with links to forums where most users are claiming they hit 5.0 ghz or higher? Because I’m almost certain you must only be reading the AMD subreddit...

Ok dude, you ca read it however you want. Normally you would show evidence alongside that claim, but you can't. That's the difference between making a statement based on simple facts (like I did) and one where you try to interpret something wrote in just 1 website (although the website clearly states that if you buy you must use the QVL components, or similar ones, to achieve similar results - I can't understand how you can take that single line out of context).

I still believe that you are trying to turn around a situation where you burned yourself by not actually reading properly the website you linked. Many make such mistakes. I find it to be very hilarious

FYI there are over 390 pages in the overclock.net 8700k OC results thread. You are free to check real results there. If you want you can also check the 8700k OC reddit megathread.

 

[Sausagemeat]

Ok dude, you ca read it however you want. Normally you would show evidence alongside that claim, but you can't. That's the difference between making a statement based on simple facts (like I did) and one where you try to interpret something wrote in just 1 website (although the website clearly states that if you buy you must use the QVL components, or similar ones, to achieve similar results - I can't understand how you can take that single line out of context).

I’ve shown more evidence than you have. In fact you’ve shown nothing whatsoever. Just claimed falseley that most people in forums can’t hit 5.0ghz.

My evidence was silicon lottery claiming that 86% of all 8700Ks tested can hit 5.0 ghz or higher. This evidence goes against your false claims and you made a false claim to try and counter it, you claimed that they would delid and watercool all these CPUs before even testing them. Which demonstrates quite a misunderstanding of how binning works. It would make sense to see if a chip can hit 5.0 before delidding otherwise they would be wasting money, time and effort on the delidding process.

Overclockers UK outline that when they bin chips, they test them before delidding as you can see here;

https://www.google.co.uk/amp/s/www....151-processor-oem-cp-642-in.html?template=amp

At the end of the day I am right, most users are hitting 5.0 ghz and I’m the only one who cited any evidence. Get proof that I and silicon lottery are telling fibs or go home..

 

[Puiu]

I’ve shown more evidence than you have. In fact you’ve shown nothing whatsoever. Just claimed falseley that most people in forums can’t hit 5.0ghz.

My evidence was silicon lottery claiming that 86% of all 8700Ks tested can hit 5.0 ghz or higher. This evidence goes against your false claims and you made a false claim to try and counter it, you claimed that they would delid and watercool all these CPUs before even testing them. Which demonstrates quite a misunderstanding of how binning works. It would make sense to see if a chip can hit 5.0 before delidding otherwise they would be wasting money, time and effort on the delidding process.

Overclockers UK outline that when they bin chips, they test them before delidding as you can see here;

https://www.google.co.uk/amp/s/www.overclockers.co.uk/intel-core-i7-8700k-pre-binned-5.2ghz-coffee-lake-socket-lga1151-processor-oem-cp-642-in.html?template=amp

At the end of the day I am right, most users are hitting 5.0 ghz and I’m the only one who cited any evidence. Get proof that I and silicon lottery are telling fibs or go home..

I don't get it. Do you usually not read what links you post? You again posted something that literally confirms what I said. It was funny the first time, but now it's just sad.

"De-lidded by 8Pack for cooler running temperatures under load
Guaranteed minimum overclock of 5.2Ghz with an AIO 240mm cooler with 1.35-1.425v"

Here dude, since you clearly "have" the time to read, I'll link for you the overclocker.net thread. 4000 posts of pure results: https://www.overclock.net/forum/5-intel-cpus/1639998-i7-8700k-overclock-results-settings.html

yes you are right, in your world the 8700k easily hits 5.0GHz stable using Intel's stock air cooler

On a more serious note: for the love of god read what you post. it will save you from a lot of headaches. I enjoy a good argument, but this just makes me look like a schoolyard bully.

 

[Sausagemeat]

I don't get it. Do you usually not read what links you post? You again posted something that literally confirms what I said. It was funny the first time, but now it's just sad.

here dude since you clearly don't have the time to read I'll link for you the overclocker.net thread. 4000 posts of pure results: https://www.overclock.net/forum/5-intel-cpus/1639998-i7-8700k-overclock-results-settings.html

yes you are right, in your world the 8700k easily hits 5.0GHz stable using Intel's stock air cooler

Claims I didn’t have time to read, most of those posts aren’t results mate! Often they are comments explaining why the users who couldn’t hit 5.0 were going wrong . You didn’t even read your own link! Desperate stuff. Just admit you’re wrong mate.

Oh and FYI, silicon lottery also state that 99.9% of 8700ks tested hit 4.9 or higher. And they do bin before delidding as every company who sell binned CPUs I can find claim they test before binning, it makes zero sense to test after delidding. 8 pack in his 8700k review states that with even with a cheap sub $30 low end air cooler you can see most chips hitting 4.7-4.8 in his 8700k review. And that was at launch. Bins have improved since launch, which is detailed here;

https://www.tomshardware.co.uk/intel-core-i7-8086k-cpu-8086-anniversary,review-34400-3.html

Why are you so desperate to falsely ascert that most users can’t hit 5.0 ghz? In the very link you can provide there is enough support to get users who aren’t hitting 5.0 ghz to 5.0 ghz. And many of the posters on there did manage to get their clock speeds up to 5.0 with a little bit of help!

 

[Puiu]

Claims I didn’t have time to read, most of those posts aren’t results mate! Often they are comments explaining why the users who couldn’t hit 5.0 were going wrong . You didn’t even read your own link! Desperate stuff. Just admit you’re wrong mate.

Oh and FYI, silicon lottery also state that 99.9% of 8700ks tested hit 4.9 or higher. And they do bin before delidding as every company who sell binned CPUs I can find claim they test before binning, it makes zero sense to test after delidding. 8 pack in his 8700k review states that with even with a cheap sub $30 low end air cooler you can see most chips hitting 4.7-4.8 in his 8700k review. And that was at launch. Bins have improved since launch, which is detailed here;

https://www.tomshardware.co.uk/intel-core-i7-8086k-cpu-8086-anniversary,review-34400-3.html

Why are you so desperate to falsely ascert that most users can’t hit 5.0 ghz? In the very link you can provide there is enough support to get users who aren’t hitting 5.0 ghz to 5.0 ghz. And many of the posters on there did manage to get their clock speeds up to 5.0 with a little bit of help!

From funny, to sad, to completely insane.

First off the link you posted just makes a reference to Silicon Lottery. Tom's article even states that they delid the CPUs to allow for better OC results.
The line you refused to read (you need to read more dude):
"Silicon Lottery compiles statistics about the samples it modifies and shares them publicly"

Second you say that you read the thousands of posts of people OCing their CPUs, in just a few hours, and you discovered that the majority are "fake". I can only facepalm at what you said.

DUN DUN DUN! It's exactly what I was saying to you.

In the end all you managed to accomplish is send 3 links with results for CPUs using water cooling and/or are delided all the while saying that anything else is fake. Are you just trolling me and I'm too blind to see it? O_o

This is just a waste of my time now. You self-destructed more times than some cartoon characters and you keep one-upping yourself with each post.

 

[MetXallica]

I have the 8700K@ 5GHz and the 1080 ti@2Ghz, best machine I've ever built hands down. Keeps my 1440p g-sync monitor humming above 100fps in pretty much all games max settings. No regrets, however 9900K is tempting...

 

[hyperspaced]

You didn't have to go through the trouble of testing 35 games. Three games at most would have sufficed.

2-line conclusion:
Intel i7 8700k for games.
AMD Ryzen 2700X for productivity and anything else.

 

[pcnthuziast]

6700k is gonna last me another year or 2 me thinks. My gtx 1080 never fails to hit 100% usage so I reckon even a beefier card wouldn't be bottle necked (gaming at 2k/4k).

 

[daDaffmeistah]

The 720p benchmarks are practical in the sense that they aren't as GPU-bottleneck and will give a clearer indication of each CPU's high watermark, but also almost no one plays in 720p anymore.
Some 4K benchmarks would be very nice, I play on a 1080 Ti SLI setup and figuring out which CPU works best with that would give an indication of which platform is more future-proof as resolutions increase so while this test probably helps a lot of folks, it does nothing for me unfortunately.

 

[sthet81]

I was hoping to see 4K numbers in this benchmark. Not sure who will play games at 720P when you have either the i7 8700k or the Ryzen 2700x and your gpu is the GTX 1080Ti.

 

[CaptainTom]

Most people do not overclock their cpus, nor do they want to have to buy an additional cooler. For that reason the 2700X is the stupidly obvious choice for the overwhelming majority of gamers. You save money, and the excellent boosting algorithm in the 2700X makes it effectively equal in performance.

Only gamers on 165Hz+ displays need consider an intel cpu. Oh, and I guess that apparently large segment of gamers who buy $350 processors to play in 720p lol.

I couldn't disagree more with the conclusion.

 

[CaptainTom]

I was hoping to see 4K numbers in this benchmark. Not sure who will play games at 720P when you have either the i7 8700k or the Ryzen 2700x and your gpu is the GTX 1080Ti.

Indeed. I get the impression most Tech "journalists" build custom watercooled PC's so they can game in 720p.

 

[CaptainTom]

What I noticed not just here but in a couple of CPU benchmarks is that AMD tends to catch up on Intel the more you increase resolution/quality. It's almost as if it's not working at it's max if not pushed to the limits.

If you read carefully you would know at higher resolutions GPU becomes bottleneck.

That can't be the entire explanation because the 2700X wins in minimum frames at 1440p multiple times. That's not margin of error.

If you game in 1440p or higher, it seems the 2700X can handle the extra data better.

 

[GreenNova343]

You didn't have to go through the trouble of testing 35 games. Three games at most would have sufficed.

2-line conclusion:
Intel i7 8700k for games.
AMD Ryzen 2700X for productivity and anything else.

Probably more like this:
-- i7-8700K is the absolute fastest, but 99% of people will be satisfied with the 2700X for gaming, especially when gaming at 1440p or higher resolutions and/or with 144Hz monitors
-- 2700X is the absolute best for productivity & other non-gaming situations, but 99% of people will be satisfied with the i5-8400

 

[Rocky12345]

Yup need some proof!

No proof required honestly. I got an 8700k and a 1080ti and even this setup struggles to get 144hz on AAA games. His GPU gets about 75 to 80fps on witcher 3 at uber ultra low

I know cause I had it

The 1060 must really suck then ( I do not own one) because on my Sapphire TRI-X R9 390x 8GB@1236MHz card on a mix of highest settings and shadows turned down 1 setting @1080p I get 85-115FPS avg in Witcher 3. If I was to lower it to ugh 720p I am pretty sure it would go up by a fairly big amount my CPU is at 5.1GHz so it would have headroom left in it my aging 390x is most likely the bottleneck in my system for higher FPS right now but it soon could be replaced by a 1080 or 1080Ti since the prices have really come down for them.