174790K upgrade or not?

Hi

I have an i7 4790K with 32GB RAM and a RTX 2080; I fly only Rift S. My CPU is fairly old but problem free but I do like to have an approx 4/5 year upgrade cycle. I only use this PC for DCS.

The question: is an upgrade worth the £1200? That is will I see a noticeable performance say if I went to a latest gen Intel i5 or i7. (I am not considering AMD, because I am old and change my views like an oil tanker turning!).

Thanks

Neal

The question: is an upgrade worth the £1200?

No .. the newer processors give higher performance ratings mostly because they have 6 or 8 cores, while yours has 4 .. but the single core performance is not really much better .. I would postpone the upgrade until the single core performance of the processors has increased at least 50% .. or until DCS starts to really use more than 2 cores.

If you upgrade to a 9900k it is a noticeable performance upgrade. If it is worth the money you must decide on your own.

If you upgrade to something below a 9900k it is a waste of your money. Amd offers far superior bang for the buck in every aspect.

So upgrading today -> Amd

Or wait till Intel is forced to lower their prizes or offer something with value.

If you upgrade to a 9900k it is a noticeable performance upgrade..

The single core passmark of a 4790k is 2530, the 9900k is 2890 .. a 14% difference .. in my opinion not worth it. The amd 3900x SC passmark is 2920, almost the same as the 9900k.

On the basis of your comments and what I have read, I think if it was a case of just a new CPU it would be ok, but when I then need a new MOBO and RAM and SSD the upgrade is much less attractive.

Thanks for your time and thoughts.

Neal

If you want to know how hardware is performing with DCS you need a DCS comparison.

In 2016 it was done by tiborrr ( https://forums.eagle.ru/showthread.php?t=157374 ) and he produced very clear and usefull boards compairing cpu's, gpu's, frequencies etc …

Since that nothing usefull is available to compare HW performances in DCS. And DCS is not included in test panel of any testers teams. (only AAA titles, and at best on a website I've found XP11 being part of the test panel for HW testing)

1 thing that is now about sure for me is that Single Core Performance is just another synthetic benchmark that is not fully relevant in game.

For exemple you can read here very often "AMD current stock Ryzen are equally or more performant in DCS than a 5.0Ghz 9900K because of Single Thread Perf". But it's just wrong.

In the website I've readen testing HW with XP11, all new stock ryzen (3600x, 3700x, 3800x, 3900x) are about 12% slower than a stock 9900K, and all slower than stock 8700k, even 7740K ! So considering all intel can OC easily near or over 5.0Ghz at conservative Vcore and temps... while Amd can't … I personnally don't consider the Single Thread Performance valid anymore.

But a proper HW comparison test on DCS, like tiborrr did, is needed.

going from a 22nm 4790 to a 14nm 9900k will give you about 20% increase in single thread performance. just from the shrink in size. that will be noticeable.

its not all about multiple cores. but the death of moores law does mean multiple cores are the only way to see speed improvements in the future.

everybody is struggling to go smaller than 10nm. the limits of silicon are just over the horizon.

and quantum computers don't exist yet.

the DDR 4 RAM, nvme ssd and pci-3.0 will all reduce latency. which is especially important in VR.

your RIFT s will like it. as its not a complete system hog like the reverb or index.

going from a 7700k to a 9900k would be a waste of money.

the new ryzan are good, but not as good at single core speed. but not by much.

if you have over 64 threads they are 50% faster than the 9900k...

which does not help at all with DCS.

so you will see a definite improvement. if you make the change.

ive been thinking about it myself. I'm just getting lazy in my old age.

.... So considering all intel can OC easily near or over 5.0Ghz at conservative Vcore and temps...

That generalization is so not true ... I've seen a lot of Intel processors that cant reach 5.0, its owners later asking on OC Forums why theirs couldnt reach the magic numbers, feeling like they got "a dud"

That's why I prefer to do my comparisons without regarding overclocking.

That generalization is so not true ... I've seen a lot of Intel processors that cant reach 5.0, its owners later asking on OC Forums why theirs couldnt reach the magic numbers, feeling like they got "a dud"

 

 

That's why I prefer to do my comparisons without regarding overclocking.

Since 7700K it is true. My bios has a default "Game Mode" which is a 4.8Ghz all cores OC, and 4.8 is near 5.0.

BTW the quotes "all intel can OC easily near or over 5.0Ghz" and "I've seen a lot of Intel processors that cant reach 5.0, its owners later asking on OC Forums why" are not necessarly contradicting.

But I feel it is starting arguing on futil thing.

I have not seen one post claiming the 3800X/3900X are faster than the 9900K at 5 GHz, but only equal. Which is true since the new BIOS updates for the Ryzen platform came out.

Never were the AMD "Hertz" 1:1 comparable to the Intel "Hertz" when it came to performance output and they still aren't, although they should give an exact indicator of the processors speed in comparison, since both use the Hertz unit for measurement. The new Ryzen CPUs at 4.6 GHz have the same single core performance as the 9900K does at 5.0 GHz. Which one is faster solely depends on the titles you're testing. And testing one application, in this case XP11, doesn't qualify someone to claim: "The 9900K is faster than the new high end Ryzen chips all the time and in every case". This would just be a lie.