Ryzen 9 3900X vs. Core i9-9900K: 36 Game Benchmark

[pcnthuziast]

There is still an excellent chance I will skip a cpu upgrade this go round, but just in case I do, my choices are now solidified. 9700 non k since I have rarely ever oc'ed my cpus. I can't seem to find any info about it's default all core boost speed though, but I'm pretty sure it's close to the k's 4.6 which would be enough for me or as long as it's at least 4.2. I'm waiting for a little more drop in price and basically looking to buy a combo with a board for 400 or less.

Any feedback or suggestions would be great!

Also just to throw it in, I still kinda think AMD makes more sense because boards can last through many gens of chip upgrades and future of gaming will undoubtedly get more threaded and the red chips may mature better than blue for this reason. All that said, even though I'm "poor" I still like to get a new cpu every 24-36 months and likewise for the gpu.

 

[CodeCreator]

Steve, please perform the exact same test but this time using two Ryzen 3900x CPUs and no Intel CPUs. This will allow us not having to read the AMD fan boys incessant crying that their $500 desktop flagship CPU is 8-6% slower in gaming then Intel's $500 desktop flagship CPU. Perhaps you can even give the Ryzen a "red" ribbon award or something.

Most of them are actually happy that it's just 6-8% slower, don't know where you got that from.

 

[Paul1979]

So 6% slower then the Intel cpu whiles the Ryzen cpu is a new architecture with raw driver and bios's.

There is a fair chance that over the next 6 months that these Ryzen cpu's are going to close the gap even more and might even over take Intel's cpu's when it comes to gaming.

With the performance being so close already and the value and performance for other task then gaming being much better, I have to wonder why anyone would go for Intel now.

Ryzen is offering better performance and performance for gaming is close enough to not really matter whiles also being better value and with it being knew, it's a given that performance for Ryzen is only going to get better and that isn't including the next gen of consoles that have Ryzen in them which likely will mean a lot more games being better optimised for Ryzen.

Seriously, don't go for Intel just for a few percent that is likely going to get wiped out over the next few months.

 

[Paul1979]

Gaming benchmarks on 2560x1440 would be most interesting.

I couldn't care less about 1080p gaming because this is not what people buy a top CPU with a top GPU for.

True and the gap in performance at higher resolutions looks smaller and the same if frame rates are in the range of 50fps and 70fps, the gap always looks bigger when having 100fps, 200fps but in reality, most gamers are going to play at higher resolutions and lower frame rates then that and the gap then looks a lot smaller from Ryzen and Intel's cpu's.

 

[codgerface]

Please do the test correctly, you are using 3200 speed memory for the AMD and give the INTEL 3600 SPPED. Give me a break use the Asus Strix motherboards they make boards for both CPU, use the same speed memory too. It already known that the Ryzen 3000 series CPU is memory sensitive the faster the memory the better it works, sure the 9900k has the better clock speed and all core boost. If you guys are so call tech geniuses then compare apple to apples and price to price please. Please stop acting like salesmen for Intel or AMD. I hate these type of articles that are not honest.

OOF. From the article you just commented on:
"Before we jump to the benchmarks, a few obligatory test notes: the 3900X has been tested on the Gigabyte X570 Aorus Xtreme using the latest BIOS revision and there are two test configurations. The first which we’re calling "stock" is the 3900X with the Wraith Spire RGB box cooler and nothing more than the XMP profile loaded in the BIOS using DDR4-3200 CL14 memory. Then we have an overclocked configuration using DDR4-3600 CL16 memory, the Corsair Hydro H115i all-in-one liquid cooler with PBO+AutoOC enabled."

 

[Paul1979]

While I support AMD and their efforts, they are barely competing on 7nm when Intel is still on 12nm. When Intel moves to 10nm, the IPC gap will again widen with Intel maintaining it's lead. AMD needs to go back to the drawing board on IPC. On the bright side, AMD does have a good handle on the multi-processor performance. If game developers were able to better write for multi-processor usage, then it may be all mute and won't matter who you go with.

Well if we was to listen to Intel, they say their 12nm fabs are better then the 7nm fabs from rivals, in any case, it's a mute point, Ryzen is delivering the goods at a decent price and even for gaming, the performance gap is so small to not matter any more and with that, I see little to no reason to buy a Intel cpu now.

 

[dirtyferret]

Most of them are actually happy that it's just 6-8% slower, don't know where you got that from.

...maybe the fan boy above telling me to shut the f**k up...

 

[Lew Zealand]

So much fanboying.

An extra 200-300 MHz will not help AMD catch Intel at the top end, Ryzen's latency (while improved) will take more than that to overcome and that's not going to happen anytime soon. So if you have a 2080Ti and game at 1440p or lower, and maybe even 4K with lower settings, you'll get more FPS from the 9900K (and 9700K?).

There's a 10-15% 9900K advantage in a few games as long as:

You spent $1200 for your GPU, and either:
You game at 1080p or lower
You game at 1440p at lower visual quality settings

That's useful to a few people out there but the vast majority of gamers are not spending that kind of money for potato settings. Instead they're using less expensive video cards and playing games at reasonable resolutions for their video cards, making them GPU-bound.

As Steve said, cutting down to even a $700 card makes the FPS differences within noise value for the vast majority of games and that's only at 1080p. Using 1440p? No difference between these CPUs. Choose whichever CPU you like.

 

[Brock Savage]

The choice is clear. In Counter-Strike, Ryzen got 534 frames per second, while Intel only got 518. That makes a HUGE difference visually on my 144Hz monitor.

 

[OnCorp]

What patch version were these tests run?

I remind everyone that Intel security vulnerabilities are real and will slow down Intel on Windows by about 15% via patches. AMD is affected by the patches by about 3%.

https://support.microsoft.com/en-us...e-execution-side-channel-vulnerabilities-prot

 

[noxendine28]

Are these tests done with the Intel Spectre and Meltdown patches disabled? Just want to make sure these tests were done as fair as possible without Intel's performance being crippled by a patch I don't even have enabled on my own computer.

 

[texasrattler]

Counter Strike is dead. Let it go. Only people playing that crap are the loser cheaters and 10-12 yr olds who know no better.
No real gamer bothers with that game. They have all moved on to other better games.

Forsaken and all the other cheaters ruined the game a couple years ago. Not to mention it's old as hell.

 

[dirtyferret]

Counter Strike is dead. Let it go. Only people playing that crap are the loser cheaters and 10-12 yr olds who know no better.
No real gamer bothers with that game. They have all moved on to other better games.

Forsaken and all the other cheaters ruined the game a couple years ago. Not to mention it's old as hell.

Tell us about that time the loser cheating 10 yr old pwned you over and over again....don't be afraid to cry.

 

[amstech]

Tell me who buys a 500$ CPU to play at 1080p?

Not many, but plenty of higher tier folks spend $300-$400 on a high end gaming CPU.

At 1440p and above the fps difference is almost nothing.

When comparing the 8/16's, there are plenty of titles they struggle to push over 100 FPS, or run between 70-120fps. If gaming on a 1440p 120hz/144hz/165hz monitor, you will need everything you can get and a 15-25FPS difference is significant, even if your over 60FPS doesn't mean its going to be butter smooth, there is a difference and that's why they make 120hz/144hz/and 165hz gaming monitors with 1MS and Gsync, its not a gimmick.

The choice is clear. In Counter-Strike, Ryzen got 534 frames per second, while Intel only got 518. That makes a HUGE difference visually on my 144Hz monitor.

What if your gaming at 1440p and your LCD supports 120hz, but one CPU gets you 97FPS and the other gets you 118? And that's before the one that gets 118 is overclocked. And according to Eve, 7 or over a half dozen games is cherrypicked....lol what a foolish, biased butthurt thing to say. Sorry bro.

Here are results from Techspots review, a mix of 1080p and 1440p results from the 8/16's.
There's a 15-20FPS difference, and sometimes more, in many games...this is just 7 examples and most importantly, this is Intel @ stock clocks. Add another 10-20FPS when its overclocked.

Hitman 2
9900K = 89/119
3700X = 83/111

World War Z
9900K = 123/151
3700X = 111/135

Far Cry New Dawn
9900K = 96/123
3700X = 88/112

The Division
9900K = 108/172
3700X = 107/158

Shadows Of The Tomb Raider
9900K = 89/123
3700X = 72/102

Battlefield 5
9900K = 125/168
3700X = 107/155

Total War: Three Kingdoms
9900K = 107/128
3700X = 106/123

Seriously, don't go for Intel just for a few percent that is likely going to get wiped out over the next few months.

The ONLY AMD cpu that's 6% slower is the $500 3900X, and that's AMD's flagship against Intels non-flagship.
A good comparision would be the $350 3700X against the $350 9700K, and the difference is 10-20% in gaming, before your overclock the 9700K to 5.2-5.3GHz.

 

[SPAM1]

What the hell?! Guys, if anyone knows what's admins contact, could you PM me or here? Seems like he has some trouble with me deleting all my posts and banning me for spam...

 

[DyneCorp]

Gaming benchmarks on 2560x1440 would be most interesting.

I couldn't care less about 1080p gaming because this is not what people buy a top CPU with a top GPU for.

Agreed. People forget that CPU "Gaming" benchmarks are optimized specifically for the purpose of comparing cpu capabilities. I pity all the poor souls who opted for i5s and i7s thinking they'll get "superior" performance when in reality they'll only get equivalent or even worse performance because either they'll game at 1440p+ or use a mid-range card...or both. Hell, I've seen the zen 2 R5s actually outperform the current i5s in most modern titles despite the benchmark conditions. People are sacrificing cores for nothing.

Uh, you realize that anything above 1080p creates a GPU bottleneck? So going above i5's until Zen 2 (R5 3600) was pointless unless reaching above 60Hz @ 1080p.

Additionally, the R5 2600 (when overclocked and paired with 3200MHz B-Die) barely edges out against an i5-8400 that's utilizing 2666. Furthermore, the i5-8400 still outpaces all Ryzen 7 2000 series processors in gaming.

Until Zen+ dropped in price, it was a tough call and the i3-8100 (a year ago sitting at $100 and outperforming all Zen products in gaming) and the i5-8400 (hitting below $180) were a tough choice when the R5 2600 was sitting $200+ and the 2600X was $220+.

Don't be an elitist. You people have terrible short term memory when it comes to how the market was just a year ago.

 

[CBTex]

Gaming benchmarks on 2560x1440 would be most interesting.

I couldn't care less about 1080p gaming because this is not what people buy a top CPU with a top GPU for.

Agreed. I know they test 1080p because it puts more emphasis on the CPU, but who is spending $500+ for a cpu and $1200+ for a gpu to play at 1080p.

 

[PetrolHead]

Interesting results. I was left wondering about a few things, though:

1. CPU utilization in games. As we saw with PUBG in 2017 (even though they've likely made some progress since then), poor optimization of the game engine can see the CPU bottlenecking performance while at the same time having very low average utilization. Which game engines relied heavily on loading one or only a few cores fully and in which games did the CPU utilization look very different between the 9900K and 3900X (taking into account the difference in core count, of course) or vice versa?

(1.1 Some of the games - Insurgency and Deus Ex at least - barely managed a steady 60 FPS 1% lows with less than impressive average FPS figures with an RTX 2080 Ti and even with an overclocked 9900K. At 1080p. What is going on with these games?)

2. PBO + Auto OC. Gamers Nexus saw in practice no benefit in enabling PBO + Auto OC. However, here we see a difference that is probably not only due to the higher RAM speed. Does GN have a bad 3900X on their hands or could this be caused by for example a buggy BIOS?

So a $475 8/16 chip is 6% faster overall then a $500 12/24 chip.
Pretty sad, and a pretty unfair review IMO, but I guess if talking price, then yes.

These chips are direct competitors because of their pricing. Also, the CPUs represent different CPU architectures, so simple metrics like core count and clock speed do not tell the full story. Not that they need to. After all, the actual performance is what matters, not how it is achieved.

Besides, I think it's pretty safe to say that most games do not really benefit from more than 8 cores / 16 threads. This was also reflected in the 3900X and 3700X review, where the 3700X provided in practice the same gaming results as the 3900X (both stock). So, if you consider the 3700X and say that it's 1-2% slower on average than the 3900X - just to be on the safe side - you're probably pretty close to the truth about what a 8-core vs 8-core CPU comparison would have looked like.

 

[penn919]

What the hell?! Guys, if anyone knows what's admins contact, could you PM me or here? Seems like he has some trouble with me deleting all my posts and banning me for spam...

Maybe your username has something to do with it....just a thought.

 

[DyneCorp]

While I support AMD and their efforts, they are barely competing on 7nm when Intel is still on 12nm. When Intel moves to 10nm, the IPC gap will again widen with Intel maintaining it's lead. AMD needs to go back to the drawing board on IPC. On the bright side, AMD does have a good handle on the multi-processor performance. If game developers were able to better write for multi-processor usage, then it may be all mute and won't matter who you go with.

AMD needs to go back to the drawing board when concerning IPC? Uh, what? Do you even understand how complicated microprocessor technology is?

Owner of an Intel 8th gen Intel processor here. Your comment omits several details and is half-truth.

Games already do develop for multi-threaded processors and that's why the i5-9600K has been rendered obsolete. Processors with multi-threading handily outperform processors without.

Single core IPC is almost tied. Optimization prevents better results in gaming and AMD has completely rendered all Intel SKUs worthless. Not just my opinion, but the opinion of nearly every reviewer. AMD doesn't need to lead in IPC.

 

[Steven007]

Really looking for 1440p benchmarks. Top end purchasers realistically don’t play games at outdated 1080p anymore. With those incredible frame rates either processor is fine, but the Ryzen part destroys in all other non gaming areas. Therefore it is the best value for the money IMHO. I mean, who only games?!?!

 

[PetrolHead]

Until Zen+ dropped in price, it was a tough call and the i3-8100 (a year ago sitting at $100 and outperforming all Zen products in gaming) and the i5-8400 (hitting below $180) were a tough choice when the R5 2600 was sitting $200+ and the 2600X was $220+.

The i3-8100 was not outperforming all Zen products in gaming. It was on par with first gen Ryzen CPUs (stock settings), sometimes matching CPUs like the R5 1600 in games that did not need more than 4 threads and with the caveat that you had to pair it with overclocked RAM, which meant you needed a Z370 motherboard.

 

[SPAM1]

What the hell?! Guys, if anyone knows what's admins contact, could you PM me or here? Seems like he has some trouble with me deleting all my posts and banning me for spam...

Maybe your username has something to do with it....just a thought.

rofl...yeah, for sure. however, my previous nick said nothing about that, still they banned me for...5 comments in 7 months?

 

[DyneCorp]

The i3-8100 was not outperforming all Zen products in gaming.

Depending on the game, yes it could; quite a feat considering how much more expensive the Ryzen competition was at the time and when the 8100 wasn't outperforming, it was at least on par.

the caveat that you had to pair it with overclocked RAM, which meant you needed a Z370 motherboard

You didn't have to pair it with overclocked RAM to get those results and I'm talking after the H and B boards were released but before the Zen+ launch.

Please re-read my original comment. Again, let me make this clear: pre-Zen+ (Ryzen 2000) but after H and B boards were released.

At 1440p or below the most powerful GPU, much of this is moot.

 

[ErikSollie]

There is one very simple reason for this, and that is that you are using the wrong reasoning. You compare multiple cores/threats while talking about software that is almost purely depending on single core speed. And then you see that 5.2 GHz is faster then 4.7 GHz, you are correct. The Intels on average have a higher sinlge-core-speed, so games go faster. What is remarkable is that the ZEN2 tech lowers the Gap so much.