Intel Core i9-9900K benchmarks show it easily outperforming Ryzen 7 2700X and i7-8700K

[fadingfool]

I expect the 9700 to retail for 449-499 and the 9900 for 749-799.

Given the way everything seems to be increasing at the moment I'd add 50-100 to each of those. The Skylake i7 7820X is still retailing at the 449-499 mark - Intel (after the disaster of Kabylake on X299) will be well aware that they could destroy their low end HEDT market if they make the i9 9900k too palatable. AMD have disrupted Intel's product stack and the 9000 series needs to have a sensible product positioning or Intel may find they are competing with themselves in the niches as AMD takes the general purpose computing market share.

 

[Dosahka]

...I use a Samsung 950 PRO (NVMe SSD), 64 GB of DDR4, and GTX 1070 with a 60 Hz full HD monitor (so I don't care about FPS above 60 at this moment, which seems to be the territory where Intel does a difference).

Almost all my friends will be moving to AMD since their Intel platforms are too old and updating to new good Intel is too expensive for them. If me and also colleagues from work (you can check where I work in my profile) are migrating to AMD, there must be some great value there. I'm not preaching AMD, each person can do an assessment and pick accordingly to their features-performance-budget expectations; just be objective and chill out, mate.

I don't know if you are aware of that amd has a chip which controls the PCIe, which can be 10% or more slower compared to directly connected to the CPU (Intel) which means the all of M.2 slot has dedicated PCIe lanes (not shared with PCIe slots). Which is a lot if we thinking about professional work and not gaming - there is still a difference but not so noticeable.
Also this is not the situation with the x470 series AMD mobos, they have at least 1 M.2 slot directly connected to the CPU.

 

[quadibloc]

It doesn't surprise me that Intel's chip is better in per-core performance. After all, Ryzen didn't beat Intel in per-core performance even when it first came out, before Intel responded. Ryzen was still exciting because it offered more cores for the same money - and its cores, once again, at least came close to those from Intel.

Because a computer needs a case, a power supply, and memory, and so on, paying twice as much for the CPU to get 10% or 20% more performance isn't as stupid as it might seem.

But AMD lets you upgrade the CPU because it isn't going to keep changing the socket. (How it's going to change from 8 cores to 16, as rumored, and still do that isn't entirely clear, but then the 32-core Threadripper shows one way.)

 

[meric]

Whistle. T for time out.

You skipped the ThreadRipper 1950.....or did I just speed read this all wrong.
? ? ?
I see a fanboi fight.....SO wassa i9-9900x run when it comes out?

You wouldn't compare a desktop CPU to an HEDT CPU. Different markets.

i9s are also HEDT aren't they? 9700k sounds more like a desktop part

 

[LemmingOverlrd]

3DMark is *NOT* what you want to be testing CPUs on.

Show me some pure Cinebench results and we'll have another talk.

In essence, Intel's 14nm 9900K will outperform AMD's 12nm+ Ryzen 2700X, core-for-core. Intel still has a meager advantage in IPC which it needs to milk a while longer. It'll cost more, because Intel isn't willing to give that up. But buck-for-buck, AMD will be ahead.

 

[AkumaMini]

Have AMD fans been bullying me? I don’t know any AMD fans mate, they are quite difficult to find. I am quite chilled actually. I’m looking forward to reading more entertaining butt-hurt comments like yours; “all my friends and colleagues are migrating to AMD so there” lmao!

If you have done a quick check of "where" me and my colleagues work, as I suggested [because I won't be writing it down explicitly in a comment], you would have a different perspective of whom it is coming from. And I don't know why you mention a boss and a company covering the cost [hint: the company I work for will never provide systems with anything other than Intel, ever]; I thought it was clear I was talking about their PCs for personal use.

If you think my comment was "butt-hurt", then you can't realize how emotional you are in both your comments, you even write as if it is something affecting you personally to wish things going bad for "x" company. Even quoting my own words, word by word, is childish; you show to be offended, even if you say you aren't. I only stated reasons and scenarios pro-AMD, to give you a big picture so that you could understand the other side, nothing personal against you or Intel, I'm not even mad my Intel board died on me so quickly. Another hint for you: if I could buy a X99 mobo + an Intel Extreme CPU, brand new for $600 total, more than a year ago... I don't think you realize that the retail prices don't add up (yes, it would have been even more expensive).

Well of course it out performs, your comparing the wrong levels of chips, that's like comparing a GTX 1080ti to a GTX 1070. They should have tested the core i9 against the real equal in AMD, the Threadripper 1950x but the scores would be too similar. Who ever wrote this article is obviously not a tech type person. How about Comparing TR 1950x vs i9-9900k, the i7-9700k is what you would compare to Ryzen 7 2700x.

Why it would be fair, for you, to compare a 8C/16T CPU to a 16C/32T CPU? If you want to show the strength of core-for-core, clock-for-clock the closest thing is Core i9 9900K (8C/16T) to Ryzen 7 2700X (8C/16T). In case you don't know this, according to the leaks: Core i9 9900K is a 8C/16T CPU, while the Core i7 9700K is a 8C/8T CPU (yes, no hyper-threading).

You compare products in the same price range. 9900k will be at least 100$ more expensive than 2700x and probably around 200$ more expensive.

If 9900k is 5% faster but costs 30-50% more, then it really isn't a good deal.

 

[HardReset]

Intel Core i9-9900K benchmarks show it easily outperforming Ryzen 7 2700X and i7-8700K

I saw only one benchmark...

 

[darkphantom]

So let me get this right in my mind
you said "show it easily outperforming "
while in reality the difference is literally less than 5% and not only that it needs to clock 300mhz faster than amd in order to achieve that tremendous amount of perfomance?

is this where you have devolved into? clickbaiting so hard to get traffic instead of actually reporting the news as it is?

 

[darkphantom]

It doesn't surprise me that Intel's chip is better in per-core performance. After all, Ryzen didn't beat Intel in per-core performance even when it first came out, before Intel responded. Ryzen was still exciting because it offered more cores for the same money - and its cores, once again, at least came close to those from Intel.

Because a computer needs a case, a power supply, and memory, and so on, paying twice as much for the CPU to get 10% or 20% more performance isn't as stupid as it might seem.

But AMD lets you upgrade the CPU because it isn't going to keep changing the socket. (How it's going to change from 8 cores to 16, as rumored, and still do that isn't entirely clear, but then the 32-core Threadripper shows one way.)

eh amd lacks speed...clock to clock we already know since the first ryzen came out that they have greater ipc

 

[EEatGDL]

I don't know if you are aware of that amd has a chip which controls the PCIe, which can be 10% or more slower compared to directly connected to the CPU (Intel) which means the all of M.2 slot has dedicated PCIe lanes (not shared with PCIe slots). Which is a lot if we thinking about professional work and not gaming - there is still a difference but not so noticeable.
Also this is not the situation with the x470 series AMD mobos, they have at least 1 M.2 slot directly connected to the CPU.

You got sort of the idea:
1. With both Intel and AMD you can connect PCIe directly to the CPU or the chipset. Only when you have low-range chipset with Intel or you have two M.2 slots with AMD: one of them connect to the chipset, losing performance. When only one M.2 slot is available in a fairly decent motherboard, it is connected to the CPU, being Intel or AMD.
2. Not even my old X99 platform had 4 dedicated lanes for the M.2 slot. They were shared with PCIe_3 slot I think. You either used the PCIe slot or the M.2, not both. A similar situation happens with AMD: you simply can't use one of the PCIe slots if you're using the M.2 connector.
3. At least with the motherboard I'm using, the M.2 slot is being enabled by CPU not chipset.

Lol, so now you’re arguing with me and asserting to me that I’m butt hurt? Desperate stuff! But why exactly? I’m not an AMD or an Intel fan specifically, surely only AMD fans have a reason to be upset as their chips are now facing new, rather collosal looking competiton and have lost their core count advantage.

Nobody cares where you work mate. And nobody cares if they are all switching to AMD. It’s not proper evidence of market share, it’s just anecdotal evidence in a comment written by a rather upset TechSpot member. It means nothing.

I’m actually finding all this rather amusing pal.

I never mentioned my scenarios were explaining the market situation, I even clarified in my second comment that they were scenarios for you to understand another perspective.

You think I'm "upset", but the only reason I'm continuing the thread is because I'm amused. I have direct interaction and certain influence in the [CPU] market, but still no one can know how it will behave or react. You can speculate, do projections, and set expectations; and yet under or over-perform. Best you can do is read the numbers, detect patterns or trends, try to understand the why, and do your next move; but still every move is a gamble.

You could say I'm bluffing, but it turns out I've been here for almost 7 years (my last years of college and the whole time I've been working in the same blue logo, big company). Let me make it easy for you: I work for the "blue side" and we have [pretty good] employee discounts for buying its products, if that is not enough for us, let it be an eye-opener about the red side value. That is the important piece I wanted you to realize and give some weight to a situation that has nothing special besides that (the context, with whom and where the scenario is happening). But OK, be my guest to dismiss everything I've written, whatever you reply I won't follow up beyond this point.

 

[Evernessince]

Lol, so now you’re arguing with me and asserting to me that I’m butt hurt? Desperate stuff! But why exactly? I’m not an AMD or an Intel fan specifically, surely only AMD fans have a reason to be upset as their chips are now facing new, rather collosal looking competiton and have lost their core count advantage.

Nobody cares where you work mate. And nobody cares if they are all switching to AMD. It’s not proper evidence of market share, it’s just anecdotal evidence in a comment written by a rather upset TechSpot member. It means nothing.

I’m actually finding all this rather amusing pal.

In this comment you don't accept his personal perspective but a mere rumor about a new Intel chip qualifies as "colllosal looking competition". When you start believing rumors are real products on the market right now, you've drank too much of the kool-aid. Remember all those GTX 11xx series rumors about when the next gen cards were going to drop? All of them were wrong and now the rumor mill is simply producing more as the previous one's are invalidated. How about we not jump to conclusions about Intel's chip here and not assume anything until the actual product is released.

As I see it, this 8 core chip having similar clocks to the 8700K at the same TDP is a red flag. Just like the rumors that Intel is going to go back to soldering, we've seen the same dang rumor 5 times already and it was false every time.

 

[Jimster480]

So AMDs 8 cores perform closer to Intels 6 cores than Intels 8 cores. Now we have core parity Ryzens poor IPC and clock speeds are being shown up.

Expect price cuts for the Ryzen line up and the Ryzen 2 hype train to get into full locomotion.

The article is wrong since there are plenty of 8 core Ryzens' over 10k score without any 5ghz OC?

 

[Jimster480]

So let me get this right in my mind
you said "show it easily outperforming "
while in reality the difference is literally less than 5% and not only that it needs to clock 300mhz faster than amd in order to achieve that tremendous amount of perfomance?

is this where you have devolved into? clickbaiting so hard to get traffic instead of actually reporting the news as it is?

That is indeed what it seems. Since the i9 isn't anywhere compared to the 2700x so far...
Intel is hanging in there with their extreme clock speeds (and extreme power consumption) that they are manipulating through the reducing of their base clock and also through the media channels by having channels only test workloads that will coincide with their faked TDP numbers.

 

[Jimster480]

Why is it that a 1080Ti and 4.4 GHz Ryzen 7 2700 (https://www.3dmark.com/spy/3771579)
gets the following Time Spy scores then:

CPU Score 10594

Could be a few reasons but it also could be that the Ryzen in your test link is apparently running with 3532MHz memory.

This 9900k test was performed with just 2666 RAM.

An 8700K gains a few hundred timespy points going from average 2666MHz to decent 3200MHz RAM. I haven't tried with anything faster, but I imagine the gains are noticeable on that test having memory running nearly 1GHz faster, especially for Ryzen oft being memory sensitive.

I would find that to be less relevant since I'm sure this i9 will cost $450+, meaning that the cost for 3700mhz ram + 2700 will be less than the i9 while still matching/beating it.

 

[Jimster480]

Its good to see Intel come out with a better performing processor but this will likely cost about 40-50% more than the Ryzen 7 2700x which you can buy now for under $300. Intel's current best i7 chip is $400 and doesn't come with an adequate cooler.

The 8700K is for overclockers. If you want Intel's best i7 at stock then buy the 8700 that comes with a cooler. It's also $50 cheaper than the 8700K....

The cooler is so poor that the 8700 cant even reach boost clock speeds with that cooler.
On top of that it gives up a ton of clockspeed and basically doesn't compete with the 2700x at all, infact its closer to the 1700x which is $189 currently. Meaning that its literally 50% more to have no performance advantage.

AMD has reduced intel to a single-bin company, they have just one chip to compete with... and its the top chip. This is the same thing that happened in the early 2000's with Pentium 4. Only their $1000 extreme edition 4.2ghz chip competed with AMD, while using twice the power and costing 2-5x as much as similar performance AMD chips.
Now in 2018 history repeats itself, as Intel blatantly lies and manipulates their TDP by lowering the base clock and raising the boost clock. They are misrepresenting the power they actually use, as even the 8700k under proper testing is using 160-190w when the boost clocks are in effect.

The rest of the chips that you would buy in the Intel lineup in the past years are all pointless now as the i3 and i5 chips are completely overrun by much cheaper Ryzen processors, and now the i7 wont have hyperthreading.... well other than the top bin chip with an extreme clockspeed; those next gen chips won't compete just as the current gen i7's don't compete other than the 8700k.

 

[Jimster480]

3DMark is *NOT* what you want to be testing CPUs on.

Show me some pure Cinebench results and we'll have another talk.

In essence, Intel's 14nm 9900K will outperform AMD's 12nm+ Ryzen 2700X, core-for-core. Intel still has a meager advantage in IPC which it needs to milk a while longer. It'll cost more, because Intel isn't willing to give that up. But buck-for-buck, AMD will be ahead.

Its not even really about IPC as literally Ryzen and Coffee Lake trade blows in IPC depending on the task at hand. Intel has a maximum clockspeed advantage and this is why they are so desperate to push their chips out that are pushing the clockspeeds to 4.5ghz and beyond (which also uses a ton of power) because truly this is the clockspeed any Intel needs to be able to actually beat a Ryzen chip, since even the laptop chips are running at 3.6ghz+.

 

[foxrox]

I keep seeing this "story" resurface on the web. I'm looking forward to finding out what real production Core i9-9900K results look like when overclocked by people that know what they are doing. These leaks and unconfirmed stories just spark a lot of annoying speculation and bickering about things. It's probably something worth getting excited about (hopefully) at the right time, but right now there's nothing anyone can hang their hat on. While it may well end up spanking a 2700X when overclocked, just the headline itself is enough justification for AMD fanbois to play the Intel fanbois card. Let's hope the final production version is more consistent and less of a silicon lottery with bin quality that isn't all over the board as much as 8700K. I love my 8700K, but there were three samples before my nice current sample that were just junk.

 

[Jimster480]

I keep seeing this "story" resurface on the web. I'm looking forward to finding out what real production Core i9-9900K results look like when overclocked by people that know what they are doing. These leaks and unconfirmed stories just spark a lot of annoying speculation and bickering about things. It's probably something worth getting excited about (hopefully) at the right time, but right now there's nothing anyone can hang their hat on. While it may well end up spanking a 2700X when overclocked, just the headline itself is enough justification for AMD fanbois to play the Intel fanbois card. Let's hope the final production version is more consistent and less of a silicon lottery with bin quality that isn't all over the board as much as 8700K. I love my 8700K, but there were three samples before my nice current sample that were just junk.

Nobody cares about results from professional overclockers.
Honestly we already know that the current Intel architectures upper frequency limit is 5.3ghz in the best scenarios.
This means that a new chip that has 8 cores isn't going to magically be breaking this speed barrier, infact its only going to have more problems hitting this stock speed on a daily basis similar to your 8700k.
Intel hasn't made any breakthroughs, they are just smashing together everything they have in a desperate attempt to 1Up AMD who currently has superior technology, lithography and design's.

The reason that fans of AMD keep posting on these articles, because they are paid advertisements for Intel in general.

 

[Lew Zealand]

Summary:

AMD's IPC is a few percent behind Intel's. Not a big deal.
AMD's top clock speed is more than a few percent behind Intel's. A bigger deal for some.
AMD in the mid ranges has OC options and apparently higher thread count (if no i5 or i7 HT) and will probably have lower prices. A big deal for many if so.
AMD has higher core count to compensate for IPC & clock speed, but the supposed 9900K erases that. A big deal and AMD will compete on price instead at the top end.
Intel's ring bus gives games a few FPS advantage. A big deal for some.
Intel's ring bus with 8 cores *may* not continue to confer that advantage or even be present (mesh bus?). That is still TBD (IMO I'll bet on ring and still slightly better than Infinity Fabric).
Will Intel still make the i7-8700K available for those who want a high clocked 6-core 12-thread option? Also TBD.

Fun times!

 

[hahahanoobs]

i9s are also HEDT aren't they? 9700k sounds more like a desktop part

Not if they are replacing

The cooler is so poor that the 8700 cant even reach boost clock speeds with that cooler.
On top of that it gives up a ton of clockspeed and basically doesn't compete with the 2700x at all, infact its closer to the 1700x which is $189 currently. Meaning that its literally 50% more to have no performance advantage.

AMD has reduced intel to a single-bin company, they have just one chip to compete with... and its the top chip. This is the same thing that happened in the early 2000's with Pentium 4. Only their $1000 extreme edition 4.2ghz chip competed with AMD, while using twice the power and costing 2-5x as much as similar performance AMD chips.
Now in 2018 history repeats itself, as Intel blatantly lies and manipulates their TDP by lowering the base clock and raising the boost clock. They are misrepresenting the power they actually use, as even the 8700k under proper testing is using 160-190w when the boost clocks are in effect.

The rest of the chips that you would buy in the Intel lineup in the past years are all pointless now as the i3 and i5 chips are completely overrun by much cheaper Ryzen processors, and now the i7 wont have hyperthreading.... well other than the top bin chip with an extreme clockspeed; those next gen chips won't compete just as the current gen i7's don't compete other than the 8700k.

Source link for that stock 8700 cooler throttling please.

8700K is still the best gaming CPU out right now. (source: internet).

8700 boost clock is 100MHz slower than 8700K, so....

You're calling a $60b a year company with their own fabs dead in the water?! Are you insane? Intel has been the ONLY company producing full process nodes. That means your AMD chips have been hybrids.

It's funny how AMD fanboys want competition, but only until AMD has something to destroy Intel with.

So sad....

 

[Jimster480]

Not if they are replacing


Source link for that stock 8700 cooler throttling please.

8700K is still the best gaming CPU out right now. (source: internet).

8700 boost clock is 100MHz slower than 8700K, so....

You're calling a $60b a year company with their own fabs dead in the water?! Are you insane? Intel has been the ONLY company producing full process nodes. That means your AMD chips have been hybrids.

It's funny how AMD fanboys want competition, but only until AMD has something to destroy Intel with.

So sad....

Intel is a garbage company that hasn't ever innovated... they always look at the other side, copy what they do (AMD) and then market the hell out of it while also being anti-competitive vs AMD and others.
Their Atom is a great example of what happens when Intel tries to innovate, they couldn't even develop their own architecture without falling back on copying AMD (I bet you didn't know that all modern Atoms past time they were based on P3 are a Clustered multi-threading "CMT" Design very similar to bulldozer but with macro ops instead of normal op decoding). And Atom was so bad that it literally needs a 1ghz advantage to beat Jaguar, but Intel decided to give away 7 Billion dollars worth of Atoms (Contra revenue) to prevent AMD from gaining marketshare with Jaguar. But thankfully AMD landed the contract for Xb1 & PS4 and used Jaguar for that, so the architecture became quite a bit more mainstream.
Then to make matters worse, Atom chips have a fatal flaw that makes their clocks break after ~2 years of service, causing them not to boot and basically to become doorstops. Most Atom's (and Celeron/Pentium N) are affected by this bug, even though its very hard for the average person to find where Intel has marked these bugs, since they aren't public about it (you won't easily find it around the net without searching the exact right keywords) and they are still selling all these Atom boards and laptops today w/ Celeron & Pentium N that are flawed and will break.

As far as thermal throttling:

http://www.tomshardware.com/answers/id-3599511/8700-stock-cooler-question.html
^here is one such link talking about throttling but literally....
https://www.google.com/search?q=8700+stock+cooler+throttle

And oh look... plenty of threads scattered about with people having throttling issues under actual load!
Amazing what you can find if you actually put something into google.

The 8700 has a base clock speed way under that of the 8700k and when you are actually playing games, the chip will not stay at turbo speed the entire time. https://www.gamersnexus.net/guides/3077-explaining-coffee-lake-turbo-8700k-8600k
^Read about how the turbo actually works.

Therefore the performance of the 8700 in real world tasks isn't close to the 8700k, only in things like benchmarks where the turbo can outlast the benchmark (depending on the benchmark).

 

[foxrox]

Nobody cares about results from professional overclockers.
Honestly we already know that the current Intel architectures upper frequency limit is 5.3ghz in the best scenarios.
This means that a new chip that has 8 cores isn't going to magically be breaking this speed barrier, infact its only going to have more problems hitting this stock speed on a daily basis similar to your 8700k.
Intel hasn't made any breakthroughs, they are just smashing together everything they have in a desperate attempt to 1Up AMD who currently has superior technology, lithography and design's.

The reason that fans of AMD keep posting on these articles, because they are paid advertisements for Intel in general.

Your first mistake is the assumption that everyone shares your view on overclocking. There are plenty of people that care. The people that do not are just one group. A nicely binned sample will overclock very well. I run my 8700K @ 5.2GHz 24/7, 1.325V no problem. I bench it @ 5.5GHz no problem. If I couldn't do that with my CPU then I would have no interest in owning a computer. That is why I do not buy AMD products. Not because they are bad, they just cannot meet my expectations with what they have to offer at this time.

 

[cliffordcooley]

Intel is a garbage company that hasn't ever innovated...

No point in reading further. You lost all credibility with your very first sentence.

 

[skline00]

Will be interesting to see what AMD has in store with the Ryzen 2.

Good job for Intel if the 9900k performs this well.

 

[Jimster480]

No point in reading further. You lost all credibility with your very first sentence.

Actually just the truth. Please name a time that they have innovated successfully in any of their markets.
And please don't reference lithography because that doesn't count.

If you take a look at most actually useful modern CPU technologies and concept's, they come from AMD, IBM, Qualcomm, ARM.
Intel has historically been a follower in terms of technology and a leader in marketing and anti-competitive business tactics.

Your first mistake is the assumption that everyone shares your view on overclocking. There are plenty of people that care. The people that do not are just one group. A nicely binned sample will overclock very well. I run my 8700K @ 5.2GHz 24/7, 1.325V no problem. I bench it @ 5.5GHz no problem. If I couldn't do that with my CPU then I would have no interest in owning a computer. That is why I do not buy AMD products. Not because they are bad, they just cannot meet my expectations with what they have in offer at this time.

Sure, I mean some people do go to the absolute extreme in terms of overclocking on a daily basis... but outside of hobbyists who overclock and benchmark for fun, this isn't applicable to the rest of the market. Since people like yourself literally have a hobby of overclocking parts and are only interested in that specifically (as you said yourself).
So you are right in that "somebody" cares about these things, but not many people.
I was speaking in terms of general market characteristics only.