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

[epsilon84]

I upgraded from a 6700K to a 8700K. BF1 performance improved a bit, but not to the extent that I was 'wowed', also the 8700K was clocked at 5GHz whereas the 6700K was at 4.7GHz which probably helped along with the extra threads.

To be honest the 6700K was more than enough as a gaming CPU, even for BF1. I only upgraded because I got a great bundle deal on the 8700K, and sold the 6700K for a good price.

4C/8T Skylake era CPUs may be 'obsolete' in terms of productivity and also having no upgrade path, plus I'm sure games will catch up eventually, but right now as it stands they are still good gaming CPUs and you gain very little moving up to higher core count CPUs.

 

[Chaotic42]

Turning off hyperthreading on my 4790k made CoD instantly playable, so I'm not really considering a 9900k anymore.

 

[ozzy702]

There are occational CPU spikes to 100% on my 8700k while playing Battlefield games on 64 player servers. Most of the time it's fairly low CPU usage though, DICE has a fantastic engine. Looking forward to playing BFV soon!

 

[beginner99]

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

True but it goes both ways about the unsubstantiated claims. Since it's not repeatable and hence not tested (I only found a very old BF4 test that shows FX 8 core on par with 4770k from a site i have never heard before) it remains personal anecdotes which you can easily challenge correctly so. My gripe was with you replying with single-player benches when talking about multi-player. That is simply extremely deceiving and annyoing.

Then it also matters if you play as bush wookie hiding in bushes and barley seeing any action or well actually playing the game. And if it wasn't obvious we are talking 1080p anything above is GPU limited.

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.

If it can use 4 cores to the max, what happens with background processes and all other apps? Yes the eat into your CPU budget. Of course you could close most of the stuff to limit this but that would be annoying by itself. This is even relevant for Single-player as these benches will never have running background processes. But that is very far from reality. Starting with anti-virus to browser open to plex to torrents and so forth and suddenly having more cores helps even if the engine can only use 4.

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

it's the spikes to 100% that matter and yeah those usually happen when there is a lot of action (explosions -> physics -> more cpu usage) and hence the most inconvenient.

 

[PeterScott]

If it can use 4 cores to the max, what happens with background processes and all other apps? Yes the eat into your CPU budget. Of course you could close most of the stuff to limit this but that would be annoying by itself. This is even relevant for Single-player as these benches will never have running background processes. But that is very far from reality. Starting with anti-virus to browser open to plex to torrents and so forth and suddenly having more cores helps even if the engine can only use 4.

This is basically a red herring. This stuff mostly uses inconsequential amounts of CPU. Like maybe 2% on a modern CPU.

 

[Pandamonia]

True but it goes both ways about the unsubstantiated claims. Since it's not repeatable and hence not tested (I only found a very old BF4 test that shows FX 8 core on par with 4770k from a site i have never heard before) it remains personal anecdotes which you can easily challenge correctly so. My gripe was with you replying with single-player benches when talking about multi-player. That is simply extremely deceiving and annyoing.

Then it also matters if you play as bush wookie hiding in bushes and barley seeing any action or well actually playing the game. And if it wasn't obvious we are talking 1080p anything above is GPU limited.



If it can use 4 cores to the max, what happens with background processes and all other apps? Yes the eat into your CPU budget. Of course you could close most of the stuff to limit this but that would be annoying by itself. This is even relevant for Single-player as these benches will never have running background processes. But that is very far from reality. Starting with anti-virus to browser open to plex to torrents and so forth and suddenly having more cores helps even if the engine can only use 4.



it's the spikes to 100% that matter and yeah those usually happen when there is a lot of action (explosions -> physics -> more cpu usage) and hence the most inconvenient.

I'd argue that 8 threads and 4 cores is enough to handle anything that 8 cores can handle.

I'm betting that bottle neck is on a single thread and no matter what you are using more cores won't fix it. Games are notorious for being coded with 1 main thread and then other threads used for less demanding parts of the game. Even when they claim to be multicore the load is never spread evenly.

 

[jpiniero]

https://www.notebookcheck.net/Schen...D-Clevo-P775TM1-G-Laptop-Review.350636.0.html

Review of a "Laptop" with the 9900K. R15 score of 209/1865 is quite good although it can only sustain around the 1765-1770 range. Battery life is of course abysmal.

 

[Pandamonia]

Anyone who remembers buying a conroe or Haswell desktop chip for £240 I7 with massive clock or performance gains knows just how rediculous this market has become.

It would be interesting to see 5ghz sandybridge perform against a post spectre coffee lake chip.

The sad part is that sandy could well and truly be faster!

 

[coercitiv]

It would be interesting to see 5ghz post spectre sandybridge perform against a post spectre coffee lake chip.

The sad part is that sandy could well and truly be faster!

Fixed that for you, and no it wouldn't. In fact performance penalty from security patches would be even higher on the Sandy Bridge system.

Straight from Microsoft's mouth:

• With Windows 10 on newer silicon (2016-era PCs with Skylake, Kabylake or newer CPU), benchmarks show single-digit slowdowns, but we don’t expect most users to notice a change because these percentages are reflected in milliseconds.
• With Windows 10 on older silicon (2015-era PCs with Haswell or older CPU), some benchmarks show more significant slowdowns, and we expect that some users will notice a decrease in system performance.
• With Windows 8 and Windows 7 on older silicon (2015-era PCs with Haswell or older CPU), we expect most users to notice a decrease in system performance.

 

[Pandamonia]

Fixed that for you, and no it wouldn't. In fact performance penalty from security patches would be even higher on the Sandy Bridge system.

Straight from Microsoft's mouth:

It didn't need fixing.

My point was that a pre spectre sandy could be faster than a post spectre coffee lake. Which shows how poor performance has gotten over time. Whilst also getting more expensive.

 

[coercitiv]

My point was that a pre spectre sandy could be faster than a post spectre coffee lake.

No, it would not, unless you somehow imagine Spectre patches brought down Skylake performance by more than 22% in CB15, by more than 20% in HandBrake, by more than 40% in PovRay.

I consistently criticized Intel for their pricing strategy and stalling their core count in the consumer space to make the most of that pricing strategy, but you're entertaining a performance myth that has no basis in reality.

 

[epsilon84]

Anyone who remembers buying a conroe or Haswell desktop chip for £240 I7 with massive clock or performance gains knows just how rediculous this market has become.

It would be interesting to see 5ghz sandybridge perform against a post spectre coffee lake chip.

The sad part is that sandy could well and truly be faster!

Conroe was a revelation but Haswell was hardly a huge leap over Ivy Bridge, or even Sandy Bridge for that matter.

In fact Haswell has basically the same clockspeed ceilings as IVB and SDB.

All my SDB to SKL era chips all overclocked to 4.6 - 4.7GHz, it wasnt until my 8700K that I was able to crack 5.0GHz.

The rest of your post is pure FUD and has already been debunked so I won't comment further on that.

 

[Pandamonia]

No, it would not, unless you somehow imagine Spectre patches brought down Skylake performance by more than 22% in CB15, by more than 20% in HandBrake, by more than 40% in PovRay.

I consistently criticized Intel for their pricing strategy and stalling their core count in the consumer space to make the most of that pricing strategy, but you're entertaining a performance myth that has no basis in reality.

I had not checked any benchmarks. But I remember that sandy hitting 5ghz and we are only now at 5ghz again but have suffered a drop in performance due to security flaws.

Also I play games on my PC so I don't read benchmarks for anything else. So my post was about 5ghz sandy for gaming how far we have come since then which doesn't feel very far.

 

[coercitiv]

Also I play games on my PC so I don't read benchmarks for anything else. So my post was about 5ghz sandy for gaming how far we have come since then which doesn't feel very far.

Fair enough, gaming, but here's the kicker: Spectre & Meltdown patches hardly affect gaming performance, we're talking about 1-2% FPS drops if any at all.

Meanwhile Skylake provided on average around 9% higher min FPS over SB before Spectre hit. That's with both CPUs clocked at 4Ghz, a GTX 980Ti, and a non-optimal RAM speed for the 6700K. Move on towards a modern GPU, use lower latency RAM on that Skylake CPU, and even with the Spectre patch in place it has every chance of pushing past that 9% advantage, let alone drop behind Sandy.

 

[epsilon84]

I had not checked any benchmarks. But I remember that sandy hitting 5ghz and we are only now at 5ghz again but have suffered a drop in performance due to security flaws.

Also I play games on my PC so I don't read benchmarks for anything else. So my post was about 5ghz sandy for gaming how far we have come since then which doesn't feel very far.

Gaming performance is barely affected by Spectre, I believe the differences before and after patching is between 0 - 3% max.

As for SB vs CFL clock for clock:
https://www.techspot.com/amp/review/1546-intel-2nd-gen-core-i7-vs-8th-gen/page5.html

+36% and +37% in avg/min fps is not really in the same ballpark in my books.

 

[LTC8K6]

Is it just me, or is the 9900K listing gone from Amazon?

 

[DrMrLordX]

Is it just me, or is the 9900K listing gone from Amazon?

Google found it for me, but the CPU is listed as "currently unavailable" with no price.

 

[Pandamonia]

Gaming performance is barely affected by Spectre, I believe the differences before and after patching is between 0 - 3% max.

As for SB vs CFL clock for clock:
https://www.techspot.com/amp/review/1546-intel-2nd-gen-core-i7-vs-8th-gen/page5.html

+36% and +37% in avg/min fps is not really in the same ballpark in my books.

I'm guessing the accumulation of improvements on the cpu and ram system has added up nicely then. If they can keep 5ghz on the 10nm chips that might be the one to buy in 2020 then.

 

[LTC8K6]

Google found it for me, but the CPU is listed as "currently unavailable" with no price.

Hmmm...when I search at Amazon for Intel 9900K it does not come up.

 

[DrMrLordX]

https://www.google.com/url?sa=t&rct...p/B005404P9I&usg=AOvVaw2sRdIH7d1weV_S8OtU_1kz

or

https://www.amazon.com/Intel-i9-9900K-Desktop-Processor-Unlocked/dp/B005404P9I

Sorry for the long URL, but that is the only way I could find it.

 

[LTC8K6]

Looks like Amazon customers are being told December for the 9900K.

 

[DrMrLordX]

Wow. Ouch.

That's like . . . a 4-5 month market window for the 9900k.

 

[LTC8K6]

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=2ahUKEwipx8aY37XeAhVOq1MKHV2zA2EQFjAAegQIBhAB&url=https://www.amazon.com/Intel-i9-9900K-Desktop-Processor-Unlocked/dp/B005404P9I&usg=AOvVaw2sRdIH7d1weV_S8OtU_1kz

or

https://www.amazon.com/Intel-i9-9900K-Desktop-Processor-Unlocked/dp/B005404P9I

Sorry for the long URL, but that is the only way I could find it.

Yes, I can get there that way, but a normal search used to bring it up, now it doesn't.

 

[PingSpike]

The 8700K was a soggy cardboard launch and Intel wasn't even whining about 14nm shortage before that one. The giant price is a pretty good sign that some one was going to make money off the limited stock situation so Intel felt it might as well be them.

 

[Despoiler]

Battlefield can and does max out 4 core 7700k in certain games at certain times. The below have BF in it. Level 1 Techs also did a comparison that showed a repeatable GTA5 CPU spike that Ryzen took better than the 7700k. Also, BF will use more than 4 cores. Anyone claiming it's a 4 core engine clearly has never played it or never done any research.

https://www.youtube.com/watch?v=yHtSS_kQ9CA

https://www.youtube.com/watch?v=BXVIPo_qbc4