Speculation: i9-9900K is Intel's last hurrah in gaming

[TheGiant]

The 3900x does 3.8 ghz@105 watts.

Mark where did you get this number? thanks

 

[Markfw]

Mark where did you get this number? thanks

Well, its 105 watt TDP, and based on my experience with my 2990wx and 2970wx, they hold to their tdp unless OC'ed. That information is in a slide in this thread.

 

[PotatoWithEarsOnSide]

Had my heart set on a 9900KFC.

 

[Mopetar]

Had my heart set on a 9900KFC.

I think that would require Intel's chickens to come home to roost.

 

[ZGR]

I was hoping to see a 9900KFC with an L4 cache. Just because.

 

[aigomorla]

9900BFG edition... with 44 pci-e lanes, so we can load the thing up with SLI + 10gbe + Nvme's til no tomorrow.

Sigh... i really hate Intel and how cheap they are with PCI-E lanes....

 

[Vattila]

Hi all, time to revive this thread. AMD has just revealed the details of Zen 2 and the Ryzen 3000 line-up at E3 — with claims of remarkable progress in gaming — but there is still some time for speculation before the processors are released and the benchmarking embargo is lifted (7 July), confirming or disputing AMD's numbers.

Since I started this thread, the sentiment has shifted markedly. The poll started well in favour of Intel retaining the gaming CPU lead, but it has since shifted to 48%/39% in favour of AMD (as of time of writing). For those among the 39% — do you expect i9-9900K to still be better, or do you expect i9-9900KS (5 GHz all-core boost clock, to be released in Q4) to retake the lead, or do you still expect high-performance 10nm desktop chips from Intel this year (or next)?

And how shall we define "take the lead"? Top of the benchmarks? How many? All? Over 50%? Of which reviews? Or by reviewer's recommendation? Again, whose? Personally, I guess we are looking at AMD becoming a "disputed" champion, with Intel still winning a lot, especially when flexing its frequency muscle, but overall, I sense that reviewers will recommend AMD, in particular for non-overclocked performance, power and value.

Here is PCWorlds's take:

AMD's Ryzen 9 3950X is a 16-core CPU aiming to topple Intel's gaming dominance

"With the debut of AMD’s 16-core Ryzen 9 3950X, Intel’s slim lead in gaming CPUs could disappear."

PS. Personally, I'm astounded by Ryzen 3000s results in CS:GO. I expected AMD would still lose in this title, as they were so far behind, but it seems that architectural changes — cache in particular — have changed the picture completely.

 

[Atari2600]

Being beaten in cinebench by a CPU with 50% more cores is pretty irrelevant for a gaming CPU which is the topic of this here post. And yeah the normal one and the s are going to be very very close since most games are still bottlenecked by graphics.

So basically you are saying better off getting either (i)the cheaper CPU and spending that saved money on a GPU or (ii)paying the same money for a more-core CPU which will more or less game the same but be much more productive away from games.

Oh, and since most games are bottlenecked by graphics, better to get the cooler running CPU that uses ~half the power while pushing out the same framerates.


Thanks for the advice. Most sense I've ever seen in one of your posts.

 

[epsilon84]

Hi all, time to revive this thread. AMD has just revealed the details of Zen 2 and the Ryzen 3000 line-up at E3 — with claims of remarkable progress in gaming — but there is still some time for speculation before the processors are released and the benchmarking embargo is lifted (7 July), confirming or disputing AMD's numbers.

Since I started this thread, the sentiment has shifted markedly. The poll started well in favour of Intel retaining the gaming CPU lead, but it has since shifted to 48%/39% in favour of AMD (as of time of writing). For those among the 39% — do you expect i9-9900K to still be better, or do you expect i9-9900KS (5 GHz base clock, to be released in Q4) to retake the lead, or do anyone of you still expect high-performance 10nm desktop chips from Intel this year (or next)?

And how shall we define "take the lead"? Top of the benchmarks? How many? All? Over 50%? Of which reviews? Or by reviewer's recommendation? Again, whose? Personally, I guess we are looking at AMD becoming a "disputed" champion, with Intel still winning a lot, especially when flexing its frequency muscle, but overall, I sense that reviewers will recommend AMD, in particular for non-overclocked performance, power and value.

Here is PCWorlds's take:

AMD's Ryzen 9 3950X is a 16-core CPU aiming to topple Intel's gaming dominance

"With the debut of AMD’s 16-core Ryzen 9 3950X, Intel’s slim lead in gaming CPUs could disappear."

PS. Personally, I'm astounded by Ryzen 3000s results in CS:GO. I expected AMD would still lose in this title, as they were so far behind, but it seems that architectural changes — cache in particular — have changed the picture completely.

I'm sure AMD chose the most CPU bound games (and at 1080P ofc!) for these tests, but let's wait for actual reviews because company made slides are basically a marketing exercise. Intels own 9900K vs 2700X comparison, FWIW:

I think it will be a very close battle, but I could see the 9900KS just edge it just due to brute force clockspeed with all core 5GHz boost. The 'vanilla' 9900K and 3800X should be a very exciting battle indeed, bring on the reviews already!

 

[jpiniero]

The problem IMO isn't now, it's next year and the year after that, especially if Rocket Lake is still good 'ol Skylake cores. Imagine Rocket Lake using 14 nm Skylake cores versus Zen 4 on TSMC 5 nm... heh.

 

[Asterox]

I'm sure AMD chose the most CPU bound games (and at 1080P ofc!) for these tests, but let's wait for actual reviews because company made slides are basically a marketing exercise. Intels own 9900K vs 2700X comparison, FWIW:

I think it will be a very close battle, but I could see the 9900KS just edge it just due to brute force clockspeed with all core 5GHz boost. The 'vanilla' 9900K and 3800X should be a very exciting battle indeed, bring on the reviews already!

Price=performance ratio, or you do not pay 100$ more for 10FPS higher gaming performance.

It should be noted, with i9 9900K or KS version you dont get CPU cooler. In short, blue warrior can expect a severe panic attack=or a drastic decline in sales.

Unnecessary YouTube clip removed.

AT Mod Usandthem

 

[DrMrLordX]

I'm sure AMD chose the most CPU bound games (and at 1080P ofc!)

Of course they did, otherwise the framerates would have hit the GPU bottleneck and everything would look mostly the same. Though if they used a 2080Ti for the dGPU, I think you'd still see separation in 1440p, where Intel's lead wasn't so high to begin with.

I think it will be a very close battle, but I could see the 9900KS just edge it just due to brute force clockspeed with all core 5GHz boost.

All the 3900x owner has to do is turn on PBO2 and win. Or manually overclock. Either way, the 9900KS will have very little headroom - maybe 100-200 MHz all-core - before exotic cooling is required. Put the 3900x in the same power envelope as 9900KS boosted to 5 GHz all-core turbo, and say bye to Intel's gaming performance lead. Except in DotA2 apparently, and maybe one or two other games.

 

[Hans de Vries]

I don't know why I spend time with this but below these numbers are compared with intel's own infamous Principled Technologies rapport just half a year ago...... Have fun with the so-called "reviews"

                     Principled
                    Technologies
                       (Intel)       Intel
                       Nov 2018     June 2019

Grand Theft Auto V       6.5%  ---->  21.0%
Far Cry 5               19.5%  ---->  38.0%
Warhammer Battle         9.6%  ---->  31.0%
Warhammer Campaign      24.2%  ---->  29.0%
Warhammer Skaven        24.1%  ---->  32.0%
War Thunder             16.6%  ---->  49.0%
World of Tanks          15.3%  ---->  37.0%
Fortnite                16.3%  ---->  27.0%
PUBG                     4.9%  ---->  27.0%

 

[jpiniero]

The Intel results I'm sure uses a faster GPU (2080 Ti versus 1080 Ti)

 

[Vattila]

I don't know why I spend time with this but below these numbers are compared with intel's own infamous Principled Technologies rapport just half a year ago

That's a striking observation!

A recreation of Principled Technologies' benchmark suite for i9-9900K would be interesting to see when the Ryzen 3000 series is released. CS:GO was a significant outlier in that paper (vs 2700X it was +50% before, and still +48% after, they corrected their testing methodology). With the extraordinary gains Ryzen 3000 is seeing in CS:GO, the overall results of this suite, vs i9-9900K, with all the Windows mitigations and scheduler enhancements since, would be quite telling.

 

[rbk123]

The problem IMO isn't now, it's next year and the year after that, especially if Rocket Lake is still good 'ol Skylake cores. Imagine Rocket Lake using 14 nm Skylake cores versus Zen 4 on TSMC 5 nm... heh.

- Intel's problem is getting a new platform
- AMD's problem is getting to 5ghz

 

[inf64]

- Intel's problem is getting a new platform
- AMD's problem is getting to 5ghz

Intel's problem is getting to 4Ghz on 10nm and having horrible yields on it.

 

[Markfw]

- Intel's problem is getting a new platform
- AMD's problem is getting to 5ghz

And from all the rumors, AMD has made it to 5 ghz on the series 3000 CPU's. We will know for sure on July 7th

 

[Vattila]

And from all the rumors, AMD has made it to 5 ghz on the series 3000 CPU's.

Good point. The two big open questions are: (1) Are AMD's gaming performance claims fair and representative? (2) How overclockable is the Ryzen 3000 series?

If matching or beating i9-9900K only requires toggling an overclocking setting in BIOS or the Ryzen Master utility, then it will be hard for Intel's status as champion not to be in dispute — especially if i9-9900K has higher power use and/or cost. For i9-9900K to really stretch its legs, it has to go well beyond its rated TDP, as thoroughly explored by reviewers, and discussed elsewhere on this forum.

 

[thigobr]

I don't see 5GHz as an important mark as many seems to do... IPC can easily compensate small clock differences (4.7GHz to 5GHz is just ~6%)

 

[maddie]

I don't see 5GHz as an important mark as many seems to do... IPC can easily compensate small clock differences (4.7GHz to 5GHz is just ~6%)

It's a psychological issue, 5GHz that is. Hell, if we used base 9 we would be 5GHZ+ already with Ryzen 3000.

 

[JDG1980]

It's a psychological issue, 5GHz that is. Hell, if we used base 9 we would be 5GHZ+ already with Ryzen 3000.

Back in 2013, AMD released a Piledriver CPU (the FX-9590) with a stock turbo clock of 5GHz. No one cared - because it sucked. Back when Intel was pushing the Pentium 4, they were the first to hit 2GHz - but AMD's Socket A Athlon chips were still better.

Not many people really obsess that much over clock rates.

 

[Markfw]

Back in 2013, AMD released a Piledriver CPU (the FX-9590) with a stock turbo clock of 5GHz. No one cared - because it sucked. Back when Intel was pushing the Pentium 4, they were the first to hit 2GHz - but AMD's Socket A Athlon chips were still better.

Not many people really obsess that much over clock rates.

Yup, I was only responding to someone that apparantly did care, and said they could not get there.

 

[Vattila]

Yeah, in the context of the thread poll, the question is how much frequency is required for Ryzen 3000 to take the lead in gaming. And, on the other hand, how much it will have to lose, for reviewers to still recommend Intel for "pure gaming".

 

[epsilon84]

I don't see 5GHz as an important mark as many seems to do... IPC can easily compensate small clock differences (4.7GHz to 5GHz is just ~6%)

Ay you're right, it's more that 5GHz 'sounds' cool.

Can I tell the difference between my 8700K at 5GHz or 4.7GHz? Not one bit, but I run it at 5GHz because... why not?