I9 9900k Official Reviews from Anandtech, Tomshardware. Add your own links to others !

[Vattila]

Yeah it's a 50% difference in bad code ( cough cough CSGO)

Yeah. Ryzen can be far behind on old and badly written code.

Here is a summary of the 3D Particle Movement benchmark you referenced, including the new version, and focusing solely on Ryzen 2700X vs i9-9900K (both 8C/16T). It is remarkable how far behind Ryzen is on the old version, and equally remarkable how well it catches up on the new optimised version.

In this particular benchmark, Zen 2 does not need to do much to overtake Core on optimised non-AVX code, but needs more to overtake on AVX code, and a lot more to overtake on old and poorly written code.

http://www.anandtech.com/show/13400/intel-9th-gen-core-i9-9900k-i7-9700k-i5-9600k-review

 

[coercitiv]

Well we have a flow of getting obsolete

Not necessarily, people who bought into Skylake and Coffe Lake got a better deal than those who went with Kaby Lake. The Coffee Lake refresh is a mixed bag, on paper they increased core count and clocks which should make it a compelling refresh, but pricing greatly limited perf/$ gains in the old price brackets, while adding another to make room for 8c/16t.

Personally I'm becoming increasingly convinced most of the MSRP increase was dictated by production difficulties, even if they set the price lower it wouldn't have made any difference.

2019 will come with 2 important changes: Intel will increase production capacity and eventually meet demand, AMD will increase pricing pressure (more perf for same cost is likely). It will be the perfect time for another jump in price/perf from Intel, just like the one we got with the original CFL.

 

[LTC8K6]

The 7700K was launched in January 2017. The 8700K was launched in October 2017 and effectively rendered Kaby Lake obsolete.

There seems to be little disadvantage to the 7700K owner so far though. There is as yet no need to upgrade from it as far as I can see. Certainly not for gaming.
There may be some specific examples where more cores are giving enough of an advantage to make a 7700K rig owner think about upgrading.

 

[LTC8K6]

IIRC, lots of people skipped the 7700K because it was not much of a bump over the 6700K.

 

[coercitiv]

IIRC, lots of people skipped the 7700K because it was not much of a bump over the 6700K.

How does that relate to the original point? It didn't age badly because it didn't sell to begin with?!

There seems to be little disadvantage to the 7700K owner so far though. There is as yet no need to upgrade from it as far as I can see. Certainly not for gaming.

A $200 CPU matched it's performance in gaming 12 months later. Becoming obsolete doesn't necessarily mean not being fit for the task, as in not being able to play the games at good frame rates. It can also mean another product does the same for less.

Look at the 8700K, the new gen launch did little if anything in making it less desirable even for people building a new system. This should be a clear warning sign that something is off, and come 2019 my bet is we'll see a correction, just as we saw with Kaby Lake.

 

[PeterScott]

IIRC, lots of people skipped the 7700K because it was not much of a bump over the 6700K.

I think even among enthusiasts, most people don't upgrade every year. The point of each new product isn't to give people with one year old products a reason to upgrade. It's to give people with older products ready to upgrade, a reason to choose their product over a competitors.

A $200 CPU matched it's performance in gaming 12 months later. Becoming obsolete doesn't necessarily mean not being fit for the task, as in not being able to play the games at good frame rates. It can also mean another product does the same for less.

Actually, being obsolete does mean no longer adequate to the task.
https://www.merriam-webster.com/dictionary/obsolete
"no longer in use or no longer useful"

Something newer for a lower price is kind of irrelevant, since the older products are not on sale anymore. Pricing is moot. The issue is does product still do the job you need it to.

 

[coercitiv]

Something newer for a lower price is kind of irrelevant, since the older products are not on sale anymore. Pricing is moot.

The 7700K is still on sale all around the world.

 

[PeterScott]

The 7700K is still on sale all around the world.

"There seems to be little disadvantage to the 7700K owner so far though"

Old stock still on sale is irrelevant, especially in context of 7700K owners.

 

[coercitiv]

Something newer for a lower price is kind of irrelevant

Pricing is moot.

Old stock still on sale is irrelevant

https://en.wikipedia.org/wiki/Hand-waving
"is a pejorative label for attempting to be seen as effective – in word, reasoning, or deed – while actually doing nothing effective or substantial. It is most often applied to debate techniques that involve fallacies, misdirection and the glossing over of details."

 

[PeterScott]

https://en.wikipedia.org/wiki/Hand-waving
"is a pejorative label for attempting to be seen as effective – in word, reasoning, or deed – while actually doing nothing effective or substantial. It is most often applied to debate techniques that involve fallacies, misdirection and the glossing over of details."

You are ignoring the context. Which was to owners.

 

[IntelUser2000]

Actually, being obsolete does mean no longer adequate to the task.
https://www.merriam-webster.com/dictionary/obsolete
"no longer in use or no longer useful"

Something newer for a lower price is kind of irrelevant, since the older products are not on sale anymore. Pricing is moot. The issue is does product still do the job you need it to.

I think we far passed that definition of obsolete in certain sections of the market for computers. What happens is people's need for better is insatiable and our senses keep adjusting to improvements so its never enough. I know 10 years ago I was fine with 480p resolutions, but nowadays it seems 1080p isn't enough.

Personally I'm becoming increasingly convinced most of the MSRP increase was dictated by production difficulties, even if they set the price lower it wouldn't have made any difference.

Kinda like how RTX GPUs are more expensive(or GPUs in general) because the die size is big and the production costs are high?

Whatever cost increase there was for Nvidia and Intel they more than make up by increased prices and thus ASPs. No, I think its the constant desire to increase stock value by increasing revenue that's at fault. That's not just with those two companies. AMD will likely follow suit, Apple and Samsung does this too. Not only that airline industries, car industries all do this.

I think Intel's Cllient computing revenue is at many-year record highs, if not all-time high. That's despite the oft-quoted saying that "PCs are dead". Yea, they only look at volume though.

 

[LTC8K6]

How does that relate to the original point? It didn't age badly because it didn't sell to begin with?!


A $200 CPU matched it's performance in gaming 12 months later. Becoming obsolete doesn't necessarily mean not being fit for the task, as in not being able to play the games at good frame rates. It can also mean another product does the same for less.

Look at the 8700K, the new gen launch did little if anything in making it less desirable even for people building a new system. This should be a clear warning sign that something is off, and come 2019 my bet is we'll see a correction, just as we saw with Kaby Lake.

Not as many people had 7700K chips to become obsolete, imo.
Which $200 chip defeated a 7700K in gaming 12 months later? The 8400 doesn't seem to defeat it?
And computers don't just game.

 

[coercitiv]

Kinda like how RTX GPUs are more expensive(or GPUs in general) because the die size is big and the production costs are high?

Nvidia's case is a bit more tricky and not directly comparable in terms of cause, since their capacity/yield issues are not clear while competition is a lot weaker in the upper performance segments. However, the end result does show similarities with CFL -> CFL refresh transition, as Pascal cards have remained largely competitive due to RTX price structure.

Whatever cost increase there was for Nvidia and Intel they more than make up by increased prices and thus ASPs. No, I think its the constant desire to increase stock value by increasing revenue that's at fault.

I'm pretty sure Intel would try to drive prices up, after all the new i9 branding is the perfect sign. It's just that at this moment price is dictated by stock, not MSRP.

Not as many people had 7700K chips to become obsolete, imo.
Which $200 chip defeated a 7700K in gaming 12 months later? The 8400 doesn't seem to defeat it?

The 8500/8600 and it did not defeat, it matched. See my original reply addressed to you.

Spoiler: 8600 vs 7700K in productivity, gaming, price/perf

And computers don't just game.

You brought up gaming, and I addressed it in a specific reply. Otherwise the benchmarks I quoted contain both productivity and gaming numbers.

 

[LTC8K6]

8500/8600 only match the 7700K if you aren't going to overclock it...

But the point is that the 6700K and 7700K are not obsolete, imo.

They keep up just fine with the 8000 series CPUs that were released within a year.

 

[beginner99]

The 8500/8600 and it did not defeat, it matched. See my original reply addressed to you.

Thing is all these benches don't include multliplayer especially BF-series. If you play BF multiplayer, a quad-core simply is a bottleneck even a 7700k.

 

[Pandamonia]

Thing is all these benches don't include multliplayer especially BF-series. If you play BF multiplayer, a quad-core simply is a bottleneck even a 7700k.

Except that we are talking about a gpu limited game. 64 player still runs at 80 to 90 fps on a 1080ti.

There might be some frame variance with 6 core but I can tell you now that it is not noticed when playing because frame rate isn't a problem. The game doesn't bog down just because you have 4/8 cpu. It probably can only be seen in a Benchmark.

So buying a cpu based on one game in one format of server with no noticeable performance loss is pointless.

Sent from my SM-N960F using Tapatalk

 

[PeterScott]

Thing is all these benches don't include multliplayer especially BF-series. If you play BF multiplayer, a quad-core simply is a bottleneck even a 7700k.

You have verifiable evidence beyond anecdotes? Because real tests seem to show a 4 core ceiling.
https://www.tomshardware.co.uk/battlefield-1-directx-12-benchmark,review-33864-8.html
https://www.tomshardware.com/reviews/multi-core-cpu-scaling-directx-11,4768-2.html

 

[Pandamonia]

You have verifiable evidence?
https://www.tomshardware.co.uk/battlefield-1-directx-12-benchmark,review-33864-8.html
https://www.tomshardware.com/reviews/multi-core-cpu-scaling-directx-11,4768-2.html

I have never played a game in my life where a 4.5ghz 4/8 cpu has bottlenecked that a true 8 core would not bottleneck.

If physical cores and threads were such a benefit then amd would have wiped the floor in these games ages ago in 64 player battlefield

Sent from my SM-N960F using Tapatalk

 

[beginner99]

You have verifiable evidence beyond anecdotes? Because real tests seem to show a 4 core ceiling.
https://www.tomshardware.co.uk/battlefield-1-directx-12-benchmark,review-33864-8.html
https://www.tomshardware.com/reviews/multi-core-cpu-scaling-directx-11,4768-2.html

Jesus is it that hard? what you link are SINGLE-PLAYER benchmarks while we are talking about MULTIPLAYER, especially 64-player maps. Seriously reading comprehension?

This is straight from the review you link. Did you even read it?

Bear in mind that this is but one slice of action from a long and varied single-player campaign. Moreover, the multi-player experience is much more frenetic, and based on what we’ve seen from Battlefield games in the past, we know it makes thorough use of fast multi-core CPUs.

Except that we are talking about a gpu limited game. 64 player still runs at 80 to 90 fps on a 1080ti.

if you have a >120 hz screen it matters if you get 80 or 120 fps and what matters more are the dips and stutters that usually happen exactly when you don't need them, eg. in action.

 

[IntelUser2000]

I'm pretty sure Intel would try to drive prices up, after all the new i9 branding is the perfect sign. It's just that at this moment price is dictated by stock, not MSRP.]

I'm saying, nowadays every mega corporation does this. It's been driving up stock prices to the roof.

The ARK price for i9-9900K is $499. It should have been called the 9700K, but they saw an opportunity to further move up the stack.

For the RTX, they call it MSRP, but barely any cards reach MSRP, so.

With the iPhones, they just moved up the price. Yes, all 3 product categories are likely more expensive to produce than before, but not nearly as much as the price increase.

I expect volumes to go down a little for all 3 companies, but record profit to continue. Of course majority of the people look at one metric, and never connect the two(Volume, and Revenue).

 

[UsandThem]

Jesus is it that hard? what you link are SINGLE-PLAYER benchmarks while we are talking about MULTIPLAYER, especially 64-player maps. Seriously reading comprehension?

This is straight from the review you link. Did you even read it?

if you have a >120 hz screen it matters if you get 80 or 120 fps and what matters more are the dips and stutters that usually happen exactly when you don't need them, eg. in action.

Before lashing out at others, you quoted this from the article:

...we know it makes thorough use of fast multi-core CPUs.

A 6700k (or 7700k) is a fast multi-core CPU. I've looked all over for your claims, and I haven't been able to see any. Maybe link a legitimate review of this issue for all us who have trouble comprehending data.

If your point is that people who have to stay above 120 FPS because of their monitors, that's a whole different story. I can play BF1 with no issues on my 6700k, although I don't have a high refresh "gaming monitor".

 

[jpiniero]

The 9700K and 9900K is, yes, pretty tough to find. The 8700K is easily available however although the price is pretty average.

Kind of funny, the 8500 is 10 bucks cheaper than the 8400 at Newegg.

 

[Pandamonia]

Jesus is it that hard? what you link are SINGLE-PLAYER benchmarks while we are talking about MULTIPLAYER, especially 64-player maps. Seriously reading comprehension?

This is straight from the review you link. Did you even read it?






if you have a >120 hz screen it matters if you get 80 or 120 fps and what matters more are the dips and stutters that usually happen exactly when you don't need them, eg. in action.

This is rubbish.

I got 144hz and I can't get near these max frames in battlefield. The game is gpu limited and I have never seen a cpu bottleneck or stutter in 64 player.

Worst case is that the average frames may be slightly lower but its not noticable

 

[PeterScott]

Jesus is it that hard? what you link are SINGLE-PLAYER benchmarks while we are talking about MULTIPLAYER, especially 64-player maps. Seriously reading comprehension?

This is straight from the review you link. Did you even read it?

if you have a >120 hz screen it matters if you get 80 or 120 fps and what matters more are the dips and stutters that usually happen exactly when you don't need them, eg. in action.

"we explored creating a multi-player metric more indicative of how Battlefield 4 is still played years after its introduction. After exchanging emails with DICE, though, it became clear that a consistent, reproducible scenario was not in the cards."

There doesn't seem to be any way to benchmark the multiplayer experience. Which is convenient for people who want to make unsubstantiated claims.

But given that the GAME ENGINE seems to top out at 4 cores, really doesn't indicate that Multiplayer will suddenly use more cores. A core limited game engine, is a core limited game engine.

The likelihood is that if someone experiences a lag in Multiplayer with 4 cores, they would have the same lag with more cores.

So again, if you have actual evidence to the contrary, I would like to see it.

 

[Pandamonia]

"we explored creating a multi-player metric more indicative of how Battlefield 4 is still played years after its introduction. After exchanging emails with DICE, though, it became clear that a consistent, reproducible scenario was not in the cards."

There doesn't seem to be any way to benchmark the multiplayer experience. Which is convenient for people who want to make unsubstantiated claims.

But given that the GAME ENGINE seems to top out at 4 cores, really doesn't indicate that Multiplayer will suddenly use more cores. A core limited game engine, is a core limited game engine.

The likelihood is that if someone experiences a lag in Multiplayer with 4 cores, they would have the same lag with more cores.

So again, if you have actual evidence to the contrary, I would like to see it.

When playing this game on 64 player the cpu is not even maxed out. Most cores are 5o to 65% utilised. Not to mention the spare 4 threads with HT.

The only thing that could make a difference is you get 2 spare cores to run the system processes and a full 4 cores to run the game. But even then I'm not sure how much difference it would make