Core i9-9900K At 5GHz Purportedly 16.5 Percent Faster Than Stock Ryzen 7 2700X

[SgtScream]

Just wait until the 2800 and X variant is released. I'm impressed with how far AMD has come to price to performance and competition with Intel.

 

[Karadjgne]

That'll be interesting. The Ryzens so far have been living upto the dreams imagined when FX was introduced, with software now finally making use of multithreads. With Ryzen making better use of high speed ram vrs Intel, if amd can get equitable IPC, that'd be the edge needed to put them on top. Again.

 

[bobba84]

All the AMD fanboys whingeing about the cost difference... if you can't afford it, stick with your AMD coffee warmers.

 

[Karadjgne]

Keep the fanboy crap to yourself, as the father of 4 kids, spending out that kind of cash for a toy for my own amusement is seriously out of budget. Not only was that a seriously stupid statement, it also shows the exact level of ignorance you enjoy.

 

[cryoburner]

Approx 16.5% faster then the R7 2700x
Costs approx 35-40% more than the R7 2700x

Don't forget that the 2700X comes with a relatively capable stock cooler, while a 9900K won't likely come with a cooler at all, and if you want overclocked performance like that, you'll likely be looking at spending around $100 or more extra on cooling the chip. So, make that around 70% more. And of course, the added cost of a Z-series motherboard just to enable overclocking at all. : P

I would kind of think performance might actually be slightly better than those results though, seeing as the clock rates when overclocked like that should be at least around 20% higher than a stock 2700X's all-core boost, and I assume Intel still has a slight IPC advantage, unless more of that has evaporated due to the vulnerability fixes or something.

Sorry but that's not even close to being true. Intel will have 10nm CPU's in 2019. (which quarter in 19' is not known ATM)

Last I checked, their 10nm chips were not supposed to make their debut until near the end of 2019, which should be evident from these new 14nm CPUs not coming out until the 4th quarter of this year. So, I would not expect them to be out for at least another year. So probably not as long as they were trying to imply, but it's anyone's guess how their first-generation 10nm chips will perform compared to what's currently available. I'm sure they'll be more efficient, but they might not actually be much faster.

 

[Karadjgne]

Can anyone say Broadwell? Lol.

 

[cookiemania66]

It will be top of the hill for about 6 months until AMD releases 7nm Ryzens. For dedicated Intel buyers, there won't be anything significantly better for 2-3 years. Intel is pushing every button they have on this one. Got to wait for something new from the architecture team.

Sorry but that's not even close to being true. Intel will have 10nm CPU's in 2019. (which quarter in 19' is not known ATM)

I only believe that 10nm WILL come in 2019 if i see first reviews or get my hands on one of these chips...
Intel said 10nm might come 2016, then 2017, then 2018 and now 2019... With all these problems they have to introduce 10nm, i believe that they push it to 2020 or maybe even 2077...

 

[bit_user]

With all these problems they have to introduce 10nm, i believe that they push it to 2020 or maybe even 2077...

Before GloFo killed their 7 nm, I'd have had little doubt that Intel would (eventually) deliver. I still think they'll keep forging ahead, perhaps putting their efforts to EUV, which their 10 nm reportedly doesn't use.

It would be rather ironic if TSMC and Samsung basically became the only two competitive silicon manufacturers, after Intel holding an unassailable lead, for so long.

 

[acelaya35]

So unreleased, overclocked processor faster than current gen processor. Must be a slow news day.

 

[Ninjawithagun]

...and yet 40% more expensive than a 2700X... Hmm, $320 vs. $450 for a whopping 16.5% performance increase? Me thinks not. I'll stick with AMD thank you very much ;-)

 

[cookiemania66]

...and yet 40% more expensive than a 2700X... Hmm, $320 vs. $450 for a whopping 16.5% performance increase? Me thinks not. I'll stick with AMD thank you very much ;-)

I will probably keep my 6600k until spring next year and depending on how good 9th Intel is vs 3th amd i will buy the best performance/Money option

 

[eltoro]

I don't get it, why only this benchmark? What's the status with other benchmarks? What about real life apps and games? From what I saw, when reviewing an assortment of applications and games, 2700X can't even reach the level of 8700K.
I was starting to get excited about a value CPU giving Intel's top (and costly CPU) such a competition. Was starting to consider it as an alternative, until I went and read some review and comparisons.
https://www.tomshardware.com/reviews/amd-ryzen-7-2700x-review,5571.html
In light of this new information, I get the feeling that this article is manipulative.

 

[TJ Hooker]

@eltoro this isn't a review, the chip isn't out yet. Some guy (Lau Kin Lam) managed to get his hands on an engineering sample and these are the results they published, Tom's is just reporting on those results. There isn't enough information available to make it possible to do a detailed comparison of 9900K vs 2700X.

 

[Karadjgne]

Historically, most Intel engineering samples are also chopped versions, many times missing quite a lot of important stuff like instruction sets etc. You'd have to get ahold of a full, final sample/prototype to make any real meaningful dissection. Comparing a Ryzen engineering sample without the infinity fabric enabled (for instance) to an i7-8700k would be a disaster. Might as well compare to an FX. Entirely possible there's enough not implemented to really affect the %, and that once the retail cpus are out, that that 16.9% difference is really 36.9%

 

[TJ Hooker]

@Karadjgne in this case I think it's reasonable to assume that these results are going to be fairly representative of the final product, just based on the results for the 8700k. Everything we know about the 9900K suggests it's the same architecture as the 8700k but with 2 extra cores, so we'd expect a 9900k @ 5 GHz to score (8/6)*(5/4.9) = 36% higher than a 8700k @ 4.9 GHz. These results have it scoring 31% higher. That's fairly close, especially given that you aren't actually going to get perfect scaling from frequency and core count.

 

[Karadjgne]

And according to the smartest ppl in the world, all the math and physics you can apply, bumble bees cannot fly. The architecture might be the same, but there's really no telling what's missing or might be implemented in designs closer to production finals. For all anyone knows Intel could be working on a way to make floating point more effective, or make lower voltages work more efficiently on the same process as the 8700k. For all we know, that engineering sample was a first run test subject using last gen tech to test a theory.

Simply don't know and won't know anything remotely concrete until Intel actually shows up with a production sample for testing.

 

[bit_user]

And according to the smartest ppl in the world, all the math and physics you can apply, bumble bees cannot fly.

Using hundred-year-old understanding of aerodynamics, sure. Are you using hundred-year-old tech in your PC? I don't think so. That's a bad analogy for pretty much anything except just how much we've learned about science, since then.

https://en.wikipedia.org/wiki/Bumblebee#Misconception_about_flight

I do think you have a reasonable point. You want to reserve judgment until the benchmarks. That's fine. Others are going to read (possibly too much) into these leaks, and that's their choice. I think we all need to accept that nothing can be definitively proven, until review samples get into competent hands.

 

[wifiburger]

Faster you say ? Well apply some meltdown/ spectre bios and windows updates !

 

[agello24]

not only has intel lost employees and consumers. they have lost their minds. However with loyal idiots, if intel builds it, they will cum!

 

[l.r.gardner]

The 9900K will be WAY faster than anything AMD has (AM4 socket). Of course it will be! More performance per clock, with higher clocks, with the same number of threads. What's confusing here.

 

[Shumok]

The 9900K will be WAY faster than anything AMD has (AM4 socket). Of course it will be! More performance per clock, with higher clocks, with the same number of threads. What's confusing here.

Way faster in some games benchmarks, but is there a game where the 2700X isn't already smooth as butter? Must be a lot of diminishing returns once you've reached 'smooth as butter'. Not sure it's worth the extra $200, and Ryzen 3rd gen might just beat it in every way for less money in 6 months.

 

[bit_user]

The 9900K will be WAY faster than anything AMD has (AM4 socket). Of course it will be! More performance per clock, with higher clocks, with the same number of threads. What's confusing here.

Compared with their 12 nm Ryzen 2000 series, yes. But at a significantly higher price.

Now, I think AMD will launch 7 nm Ryzen ahead of Intel's 10 nm mainstream CPUs. If they do, then you might see AMD finally take the crown. At least, for a little while.

 

[quilciri]

The 9900K will be WAY faster than anything AMD has (AM4 socket). Of course it will be! More performance per clock, with higher clocks, with the same number of threads. What's confusing here.

More performance per clock? Are you implying that the cannon lake refresh is going to have over 15% better IPC, or are you just using the term "per clock" incorrectly?

...because AMD has a big per clock advantage on Intel right now.

https://digiworthy.com/2018/06/14/amd-ryzen-intel-coffee-lake-ipc/

It's gonna be even more interesting when AMD skips straight to 7nm while Intel is on 10.

 

[mlee 2500]

The 9900K will be WAY faster than anything AMD has (AM4 socket). Of course it will be! More performance per clock, with higher clocks, with the same number of threads. What's confusing here.

More performance per clock? Are you implying that the cannon lake refresh is going to have over 15% better IPC, or are you just using the term "per clock" incorrectly?

...because AMD has a big per clock advantage on Intel right now.

https://digiworthy.com/2018/06/14/amd-ryzen-intel-coffee-lake-ipc/

It's gonna be even more interesting when AMD skips straight to 7nm while Intel is on 10.

I would observe here that in the real world users notice application performance, not instructions per cycle or how tiny the process node is or isn't.

In the gaming world today, the edge happens to go to the silicon running at a higher frequency, period, and that's unlikely to change until games start to take better advantage of more then a few cores.

In other applications the math is different. I still use 128 thread chips from SUN Microsystems running at a paltry 1.4GHz for massively multi-user server instances where concurrency is king (in fact, I think of them often when AMD fan boys act like AMD is doing something new or different...and that's no knock against AMD, I think what they've done recently is awesome for everyone except Intel).