At 4.7ghz All core OC you should see better performance out of the 5900x in all but a select few titles. Your RAM configuration and everything is all ideal, I don't see any performance being left on the table.

What are you playing? Have compared your scores to reviews?

 

Glad to see I am not the only one having issues, I upgraded from a 10900k to a 5900x. In call of duty I have lost average 30-40fps. I use to be able to sit at 160-170fps (running against frame limited for nvidia reflex) quite a bit now I am lucky to hit 120-130. I am hoping that someone can give some info, this is my first team red build

 

Upgraded LOL

I have both 5900x and 10900k, and 5900x is a sidegrade in some programs at best.

Frustration upgrade i therm of memoryoverclocking 🤣

 

So are you guys saying every reviewer out there got it wrong. Intel is the way to go PC gaming,good thing Intel has some great sales on there CPU's.

 

No what I am saying is there are quite a few games that are not at all favoring AMD, I mean you tell me how I go from running on the limit of of nvidia relfex to changing to 5900x and godlike and running 120-130. Most reviewers are showing the same games being benchmarked, trust me I would rather be bragging about how awesome the 5900x is, if I wasn't custom hardlined water cooled I would already be back to team blue.

I keep reading that it could potential be a bios that will need to mature some

 

It is impossible for AMD to be faster in every game. Especially after Intel having the fastest CPU for years and every game in the last 15 years were made for Intel.
If you had a specific game in mind when you bought your CPU then that is on you. All I can say I stopped using my Intel computers for PC gaming and went with AMD Ryzen for the last two yeaRS.

Yepper I knew AMD was slower in games but to be honest I lived with it and was not thinking while I was playing a game on my AMD machine that my Intel machine was faster than this AMD thing.

Now I am not all in on AMD,I actually have my finger on the button to buy Intel new machine because some of there CPU's are so cheap. If I do I will not be wondering that my 5Ghz 5800X was faster in some games than my new Intel machine. That would just drive me crazy.

You can say I am full of it or anything but it is what it is.Play the game at 120FPS and be fine with it ,it is not all that bad.

 

Interesting thread. I think we need more data to reach conclusions

What game titles?
Also, like someone suggest, are your results comparable to those obtained independent reviewers? They should be better since it looks like you overclocked your RAM really well.

Maybe it's a BIOS thing from the motherboard...
Or a configuration thing between your ram OC/other cpu settings in the motherboard. Setting everything to default and the RAM to it's XMP profile could bring more clues to this.

How's the CPU and GPU usage in those games, either of them in 100%? (on the case of the CPU, 100% on the cores being used by the game)

 

Hi.

I decided to update my system after two years for something more powerful for gaming. I game in 4k and I know normally I'll be gpu restricted. But still I have seen bechmarks 5900x beating the 10900k with good rams and a beefy gpu. I got the Asus 3090 Strix and I even had a lower 8700k @5GHZ (lower than 10900k).

The sintetic benchmarks are superb comparing to the 8700k @5GHZ. The problems comes with gaming, my main usage.

Gaming and graphics benchmarks are slower with the 5900x than 8700k @5GHZ with the same settings with the RTX 3090.

My build is:
Asus Crosshair viii dark hero (570x).
5900x with Arctic Freezer ii 420.
G.skill trident z 2x8gb @3800mhz 15.15.15.35. Infinity fabrik 1900. 1:1.
Asus Strix 3090 OC.
Samsung 970 pro 1tb.
Ax1200w corsair psu Platinum..

I have tried diferent methods with no success.

1) All cores @4.7ghz.
2) PBO OC. It's get to 5.150 at some points adding 100mhz in bios. Rest of cores around 4.475mhz when benchs.
3) Dinamic OC. It's a mix of both PBO and all cores.

Drivers, bios, windows are up to date.

I have been tweaking the bios up and down for days, reading reviews and forums with no luck. I only use BIOS go oc. Never tried Ryzen Master but I don't think it won't change much.

Any suggestions or ideas to improve performance in games-graphic benches? 5-7pfs less in games is a lot for a supposed to update of the system.

Were you running the 3090 on the previous CPU? From seeing all the reviews this does sound wierd and the 5900x is pretty much tied or better then the 10900k let alone the 8700k.

Have you checked what Pcie lanes are running? Sometimes they get funny with m.2 slots. I also knows it sounds silly but when i upgraded from my 3800x to the 5900x it redid all my games settings and i was getting worse FPS due to RayTracing being enabled for some reason after it.

So are you guys saying every reviewer out there got it wrong. Intel is the way to go PC gaming,good thing Intel has some great sales on there CPU's.

Wow so helpful.🙄

 

Were you running the 3090 on the previous CPU? From seeing all the reviews this does sound wierd and the 5900x is pretty much tied or better then the 10900k let alone the 8700k.

Have you checked what Pcie lanes are running? Sometimes they get funny with m.2 slots. I also knows it sounds silly but when i upgraded from my 3800x to the 5900x it redid all my games settings and i was getting worse FPS due to RayTracing being enabled for some reason after it.





Wow so helpful.🙄

Thanks I thought is was.

 

I'm not into any kind of "team". Amd or intel are both to me. I just buy whatever will give me better performance. The reason of the post is to try to find what it could be wrong with my set up or configuration because the numbers don't match reviews in games. The rest like cinebench, aida memory and many other test says in into the "normal' scores of 5900x.

I have tested port royal and went from 1475 with the 8700k @5GHZ to 13923 with 5900x PBO (the highest score I got from the three different methods of OC).
The other benchmarks into 3dmark gets worst scores in graphics (the ones that include CPU scores I only take count the gpu score). The weird part comes when the cpu scores in those tests are much higher than the 8700k obviously, but the gpus scores are lower.

Games I have tested many. Shadow of tomb raider, assassin creed odyssey, battlefield v, far cry 5, shekiro die twice, horizon zero dawn, death stranding and a few others. All with the same result getting between 5 to 8fps less.

I used the same graphics card with the same overclock. Asus Strix 3090 OC @+90mhz gpu, +600mhz memory.

The bios as it's the new Crosshair Dark Hero or the drivers for the chipset might not be mature now but I follow reviews and forums and they get better performance in most games than the 10900k (I have a 8700k @5GHZ and I get less). So the bios-drivers won't be the issue cause everyone would have them.

There is a very interesting article (I think I can't put the link here because its from another website) where it's shows and explain very well how the 5900x outperforms the 10900k if you have a good ram kit and ampere graphic card. In this study they used ram kit of 3600mhz 1:1 Infinity fabrik and a rtx 3090. Both similar I have. If someone can confirm I can put the link here I'll do. Very interesting.

I have tried so many things in bios and different OCs. Put the nvme drive in the second slot so it won't get direct pcie lines to the gpu, put the pcie in gen4 and gen4,...etc. There is something wrong in my set-up or configuration regarding graphics. I'm coming from 8700k @5GHZ.

It's not a intel-amd war. I'm only looking for ideas about how to find a solution.

Any ideas?

 

Return the AMD board and CPU. Keep the 10900k until Rocket Lake comes out. Then you can sell the 10900k and upgrade to the next gen CPU which will work with your Z490 board. Intel CPUs have great resale value. You should get at least $400 for the 10900k after Rocket Lake comes out. I sold my 7700k in July for $270.

 

It's really hard to say. You aren't posting numbers so we don't have anything to compare too.

I see a few possibilities

1) Your benchmarking is not equal. Are you sure that all the setting you used are identical between the Intel and AMD systems? Game settings, nvidia control panel settings, ect?

2) Your benchmarking is not accurate. Are you comparing the 8700k from memory? How sure are you that the intel system is actually faster?

3) Your 8700k is faster and there is something about the games you play and the settings you use that says something interesting (and unflattering) about Zen 3.

Let's start with an easy one. Would you be able to run the Shadow of the Tomb Raider benchmark and Farcry 5 benchmarks at Ultra 1080p? Those games have been benchmarked heaps on both the 5900x and 3090, and so your results should tell us a lot about whether or not you have a slow AMD CPU or a particularly fast intel one.

 

Thank you for replying. Let me try to explain a bit better with your advice.

1) Settings are similar as they can be being two different platforms. Nvidia control panel settings and driver are the same. Overclock in the rtx 3090 is the same as well. Ram is the same running 3900mhz in both systems with the same main latencies 15-15-15-35 (the rest is on auto in motherboard. 1T for amd, 2T for intel). The overclock in both systems are different because of the arquitecture. The 8700k is on 5ghz and with the 5900x I tried with PBO and all cores @ 4.7ghz. Temps ok with both because I have a bigc coooler The rest of the setup is the same. Psu Platinum 1200w, cooler, ssd, case.
Two different platforms but as close it can be in every way.

2) The system overall is way faster in 5900x. Even more than double in some tests. It's just games and graphic benchmarks are lower. I changed the my old set up for better gaming even knowing that at 4k 120fps there won't be much difference. Some 5-7fps up but specially higher lows which is the reason for my upgrade. I'm getting 5-8fps lower with the new system is something is not right I guess.

3) The 5900x should be faster for sure than 8700k. I just have a problem than its slower in my case. A setting or I don't know but I have been trying for a week everything I can imagine with no luck. Port royale 8700k gets around 14500. With 5900x barely 14000. Quite a lot.

4) I play only at 4k so I only bench for that resolution. At the end the 5900x should outperform or at least equal the 8700k, but it's not in my case when I get worst performance. I give you some numbers:
Far Cry 5.
110fps with 8700k@5ghz.
103fps with 5900x PBO. 101fps@4.7ghz.

Shadow of tomb Raider.
99fps 8700k 5ghz
95fps 5900x PBO

Shekiro die twice.
155fps 8700k 5ghz
150fps 5900x PBO

Assassin's creed odyssey.
61fps 87000mhz
59fps 5900x PBO

The lows are lower in all games with the 5900x as well.

Something is not right in my set-up according to reviews and forums. I'm not saying the 8700k and intel better. Not at all. I'm just trying to find a solution for my problem because the numbers don't match.

If the cpu scores benchmarks were bad as well maybe I have a bad chip, but it's not the case. The 5900x crashes the 8700k in those and temps are good.

If I can put the link of the study where shows the 5900x beats 10900k in games let me know. It will be helpful but I don't want to get baned in the forum.

I hope with this details I can get a little more help.

Thank you.

 

Did you reinstall windows when you changed from intel to amd ?

 

Windows clean install from scratch. Even my old set-up had 2 years Windows being used.

 

What is your memory interleaving setting in the BIOS?

 

That's something I have no idea what it is.

If it's single or dual rank I know. Mine is single rank 2x8gb samsung b-die.

 

GN had something similar occur in their cyberpunk CPU benchmark, when the game got GPU limited that ended up somehow hurting the performance on ryzen. Maybe that's what you're seeing aswell.

 

Cyberpunk was also found to be using some sort of Intel compiler too that wasn't favoring AMD. There was a huge Reddit thread that blew up about it.
On intel it's using the entire CPU. On AMD it's only uses the threaded cores, not physical. So there was a hex edit guide on how to fix that in the EXE that raised cpu usage another 25% and people were getting 10-15FPS extra.

Showed my buddy with a 5950x and his cpu went from 30% roughly up to almost 60% and frames shot up.

 

Wow. I'm going to make deep research now of dual rank...maybe that's the problem

 

But then you got to retest your 8700K with DR as well😜

 

Dumb question, have you tried your GPU stock? Sometimes an OC ends up hurting performance because the OC isn't actually stable. Steve from GN keeps stressing that and I noticed as much on my 1080 Ti as well.

My GPU runs much better undervolted and with slighlty lower clocks than with max voltage and higher clocks. For example, I got around 6200 score at 2000MHz at 1.025V, then I upped the voltage to the max (without custom VBIOS), which is 1.093V, and the clocks to 2050MHz (higher was unstable), the performance was around 6000 or under! I tested 5 runs with each setup and the run-to-run variance was around 60 points.

 

Use monitoring software and watch thread usage. Cyberpunk used generic FX era scheduling and doesn't use SMT on Ryzen. I suspect we will find even more of these "accidents" in the future.

If your hardware is running properly than it's a software issue.

 

something off with your results dude...

in my testing with a friend rig.. 5900x beats the 10900k and thats a optimized rig on both side at 1080p handsdown..
at 1440p.. they trade blows.. but 4k they are the same..

so.. SOTR.. the scenes where the gpu utlization drop.. the cpu max fps for 5900x is way way way higher than 10900k.. its a astronomical amount .

 

something off with your results dude...

in my testing with a friend rig.. 5900x beats the 10900k and thats a optimized rig on both side at 1080p handsdown..
at 1440p.. they trade blows.. but 4k they are the same..

so.. SOTR.. the scenes where the gpu utlization drop.. the cpu max fps for 5900x is way way way higher than 10900k.. its a astronomical amount .

I believe you when you say its faster, but by a"astronomical amount"??
Could you back that up with a benchmark screen? Here is my 10900k @1080p medium preset and 720p highest preset

 

Try running your gpu at stock. Like others have said these cards these days run better the cooler they run. Also max the fan out

 

The numbers you are getting with an overclocked 3090 are right in line with what Toms Hardware got in their 3090 review. They used a 9900k in their testing:

Nvidia GeForce RTX 3090 Founders Edition Review: Heir to the Titan Throne

The GeForce RTX 3090 is a Titan-class card, with a bigger performance benefit than older GPUs, but minus a few professional features.

www.tomshardware.com

Guru 3d did benchmark the 5900x at 4k with a 3090 and their numbers in farcry 5 are identical to your 8700k:

AMD Ryzen 9 5900X and 5950X review (Page 29)

We review ZEN3, the new Ryzen 9 5900X, and 5950X. Released by AMD as a new architecture that will once again attack Intel, this round with a heavy focus on your gaming performance. Overall, this pro...

www.guru3d.com

A less than 7% difference in performance in what should be GPU bound scenarios could be explained by something going on in the chipset or motherboard. Any chance you may be running at an 8x PCIE link somehow? Sometimes onboard devices like extra sata controllers or NVME drives can take up PCIE lanes. Something else that could be hurting it, maybe the boost profile on your new system for the card could be slightly different for whatever reason. I would check those next.

 

Now I can compare properly with a 10900k. Something must be very wrong.

 

This is my tomb raider bench comparing my 5900x with the 10900k it was posted here. I put the same settings. I wish I did this screenshot with the 8700k but with the 10900k is quite big. Something's wrong in my set up.

 

This is my tomb raider bench comparing my 5900x with the 10900k it was posted here. I put the same settings. I wish I did this screenshot with the 8700k but with the 10900k is quite big. Something's wrong in my set up.

The 10900k clearly says Medium settings while you're running High on the 5900x. Run the same settings.

 

Hi,
How do you know what frequency that 10900k is running at ?

 

5.4 with ht or 5.5 ht off(my two daily settings). Can't remember exactly coz I had to grab the screens from another forum since I'm offshore atm.

That said, I've seen the 5900x do a lot better than sub 200ish fps range

But I'm curious about that 5900x that is better by "astronomical amount"

 

It's probably pretty high but that shouldn't matter. The 5900x shouldn't be 30%+ behind a 10900k in any benchmark.

 

What is the memorylatency differene?

 

5900x Cpu stock
ram 2x8 3200 CL14 (still w8 for new kit 2x16)
Dark Hero 3003 bios
3090 @2040
TR Demo benchmark

 

Your results make perfect sense. 8700k is actually a 10900k with 4 less cores, and slightly worse silicon. Intel has been recycling the thing since 2015 but with moar cores.
Zen3 while caught up for the most part it is still slower than Skylake in gaming generally speaking, so if that is only your path, your only reaction is to be unimpressed by the performance sidegrade in games that don't benefit from the extra cores provided by the 5900x.
Nothing to see here and everything is operating normally on both rigs.
Everybody was impressed with Zen3 because AMD is now competitive with Intel more than ever when it comes to gaming performance, in which they sucked at since 2006, not because they are faster. It is very easy to get confused because of all the hype created after the Zen3 praising reviews.

 

RTX 3080 + 5800X
Me I do not see a problem here with 200+FPS with 3090 /3080 at 1080p. Use real setting at least.Unless your one of those people that buy the best gear to play medium 720p/1080p. Then all is good,Intel is faster ,return your AMD gear or sell it.

Here is a possible reasonfor AMD Ryzen with RTX 3090. Clock stretching. Although I like what @HeadlessKnight said in his post

 

RTX 3080 + 5800X
Me I do not see a problem here with 200+FPS with 3090 /3080 at 1080p. Use real setting at least.Unless your one of those people that buy the best gear to play medium 720p/1080p. Then all is good,Intel is faster ,return your AMD gear or sell it.

I have tried real settings. You just only need to read the post to see before comment.

I just did this 1920x1080 bench to compare with the 10900k because all my other bechmarks and games are lower than they should. Even lower than my 8700k at 4k or 1080p in benchs and games. And it shouldn't be like this because the 5900x is more powerful.

That's the reason of the post.