I had finally decided to go 4790k until...

[Tremulant]

Newegg sent a weekender sale flyer. A 5820k and the MSI X99S SLI Plus for $519.98 PLUS a $20 rebate from MSI. $499.98 makes the extra price of DDR4 almost not as prohibitive compared to the 4790k/Maximus VII rig I was considering. The money part of the equation is always part of the decision process when discussing 4790k v. 5820k but at this moment, not so much. The 5820k is still a bit of a mystery to me though when it comes to ram, does it need to have 4 sticks, are two fine to start, its hard to find definitive answers.

After so much going back and forth the things I've deduced are:
-When the 4790k becomes obsolete, I'm sure the 5820 will be right there with it but with close to equal pricing the 5820 may be worth it.
-Buying DDR4 now with the sole idea of future proofing seems somewhat foolish as DDR4 in a few years from now will be a different beast(think early DDR3 vs the DDR3 2400 sticks today that really take advantage of Haswell's memory controller). It would get you up and running on a future build though.

Ahh, first world problems. I really do need an upgrade, I'm embarrassed to even say the rig I'm running on right now. I tend to make computers last for a long time though-think 7 years...

I use this for gaming, general use and if I get off my lazy butt, more music production. I will be buying a GTX970 as well(I'm on a BenQ GTG XL2720Z I just purchased)

 

[[21CW]killerofall]

Get a 4790k. Memory is not a bottleneck for games (or most other applications). Even DDR3 1333 is fast enough for gaming in most cases. Get the 5820k if you need the extra cores. Here is a good apples to apples (or as close as you can get) comparison between DDR3 and DDR4 https://www.youtube.com/watch?v=utWnjA4NzSA

Also if you are on a i7-920 or newer, you are unlikely to notice much of a performance upgrade by getting a new CPU. Your best bet is to just upgrade the GPU first. I am still rocking my i7-970 and will be until it fails or Skylake k is released (at the earliest).

 

[Dangman]

Newegg sent a weekender sale flyer. A 5820k and the MSI X99S SLI Plus for $519.98 PLUS a $20 rebate from MSI. $499.98 makes the extra price of DDR4 almost not as prohibitive compared to the 4790k/Maximus VII rig I was considering.

Not exactly a fair comparison there man: Assuming that you're talking about the Maximus VII HERO, that Maximus VII Hero motherboard is of definitely higher quality than that MSI X99S SLI Plus motherboard. So a more apt-comparison would be the 4790K and an Asus Z97-A motherboard.

The 5820k is still a bit of a mystery to me though when it comes to ram, does it need to have 4 sticks, are two fine to start, its hard to find definitive answers. )

I've built X99 platforms with just one stick of RAM to start. Personally I'd go with two but even one is fine.

After so much going back and forth the things I've deduced are:
-When the 4790k becomes obsolete, I'm sure the 5820 will be right there with it but with close to equal pricing the 5820 may be worth it.

Yes and no. If the games you're playing by then aren't taking advantage of multiple cores, then yes the 5820 would actually be obsolete BEFORE the 4790K. If the games you're playing do take advantage of multiple cores and/or you're doing a ton of music production work, the 5820K would last quite a bit longer (about an extra year or so)

-Buying DDR4 now with the sole idea of future proofing seems somewhat foolish as DDR4 in a few years from now will be a different beast(think early DDR3 vs the DDR3 2400 sticks today that really take advantage of Haswell's memory controller). It would get you up and running on a future build though.

Ehhh, not really. Current and even past Intel platforms don't really take full advantage of high-speed RAM. I've yet to see any sort of data showing noticeable improvements in real world apps and games with higher speed RAM. If the trend continues, I highly doubt higher speed DDR4 RAM would be beneficial to you.

 

[dasa]

i would go the 5820k if your overclocking the extra cores may come in handy in something like dying light
http://gamegpu.ru/action-/-fps-/-tps/dying-light-test-gpu.html
asrock mb like x99 extreme4 are a fair bit cheaper and are usually decent quality
if your not overclcoking though then the higher clock speed on the 4790k compensates for the loss of cores in games and in games that dont have good threading like arma\dayz the 4790k will be a fair bit faster

as for ram not helping thats not true it helps cpu performance a lot
its just that a lot of reviews test ram performance in games that are so badly gpu bottlnecked that there is no difference between a pentium and a 5ghz 5820k so how can you expect to see ram make any difference?

some ram tests in games that are more cpu limited can see up to 20% gain
http://forums.bistudio.com/showthre...rmance-comparison-1600-2133-up-to-15-FPS-gain
http://forums.atomicmpc.com.au/index.php/topic/55771-cpuram-performance-in-thief/
http://www.team-greatbritain.com/call-of-duty-ghosts-vs-battlefield-4-multiplayer-benchmark/
http://www.xbitlabs.com/articles/memory/display/haswell-ddr3_7.html#sect0

 

[lodoss900]

"The money part of the equation is always part of the decision process when discussing 4790k v. 5820k "

You answered your own question. Get a 4790K and close this chapter on your live.
The 4790K is not going to be obsolete any time soon. You can use DDR3 memory, and have a variety of motherboards. You can get these new or used.

The 5280K is not going to drop in price any time soon.
DD4 memory, which will not drop in price for the foreseeable future.
By your post, there is no advantage to DD4 memory
Extra cores have minimal impact on gaming. Gaming is all about the GPU.

 

[Dangman]

as for ram not helping thats not true it helps cpu performance a lot
its just that a lot of reviews test ram performance in games that are so badly gpu bottlnecked that there is no difference between a pentium and a 5ghz 5820k so how can you expect to see ram make any difference?

some ram tests in games that are more cpu limited can see up to 20% gain
http://forums.bistudio.com/showthre...rmance-comparison-1600-2133-up-to-15-FPS-gain
http://forums.atomicmpc.com.au/index.php/topic/55771-cpuram-performance-in-thief/
http://www.team-greatbritain.com/call-of-duty-ghosts-vs-battlefield-4-multiplayer-benchmark/
http://www.xbitlabs.com/articles/memory/display/haswell-ddr3_7.html#sect0

Well the first two links don't count as much considering that they're from YOU. So there's a bit of bias right there.

The third link is close but the problem is that it seems rather improbable that the reviewers ran that many test to test each RAM speed: "To ensure that we provide an accurate representation of real world multiplayer usage, we decided to play 10 rounds on different maps". So that would be 40 runs for each game for a total of 80 runs total. Nor does it mention how long said runs were made.

The last link only really matters if you're actually playing at 1280x800 or Theif the entire time. For the other 75% of the games at 1080P, the performance difference was miniscule.

 

[Dayaks]

I plan to wait until Skylake to go anywhere near DDR4. The premium is too much for me and I own a 980... I planned my next build for 2016 Q4 in hopes DDR4 will be lower/16nm GPUs/Windows 10/Skylake. (I realize these might not all be released.)

If I was building today I would get something DDR3 no question.

 

[jolli]

I would go for the 5820K. I wish I had waited for X99 to launch instead of building another socket 1150 system.

You can get 16GB of DDR4 for ~175 dollars. Yes there is a price premium, but it's not that great.

 

[dasa]

The last link only really matters if you're actually playing at 1280x800 or Theif the entire time. For the other 75% of the games at 1080P, the performance difference was miniscule.

same could be said for i3\fx6300 vs i5 but most people still pay twice as much for a i5 for the games that do perform better with it when faster ram only costs ~$10-20 more? and then there is the i7 which doesnt help at all over a i5 in most games while costing $100 more
those xbit results at a low res would be similar if you had say gtx980 sli at 1080p and were trying to run over 120fps for a high hz monitor

yes the first two are from me but if you look through the thread you will see a lot of other people getting similar results with arma as its one of those games that is not a gpu bottlnecked console port that runs fine on a pentium as long as you have enough gpu grunt
a lot of people still spend a lot so they can overclock there cpu despite the fact that most games dont need it when faster ram can provide similar gains for much less outlay

20% is the difference between sandy and haswell so a 2500k with 2133c9 ram could perform much the same as a 4690k with 1333c9\1600c11 ram

 

[mda]

same could be said for i3\fx6300 vs i5 but most people still pay twice as much for a i5 for the games that do perform better with it when faster ram only costs ~$10-20 more? and then there is the i7 which doesnt help at all over a i5 in most games while costing $100 more

FX6300 vs i5 --There IS a noticeable difference in terms of CPU alone with lightly threaded games at resolutions that matter unless you want to do some 4.5ghz overclock on the FX6300 and use a lot of power while you're at it.

Also, if you look around at the CPU recommendations being made around here, the i5s are recommended waaaaaaay more often than the i7s when only gaming is concerned.

 

[Dangman]

same could be said for i3\fx6300 vs i5 but most people still pay twice as much for a i5 for the games that do perform better with it when faster ram only costs ~$10-20 more?

Most people don't know about the cheaper non-overclock capable Core i5 CPUs, don't know how to do a proper cost to performance analysis, focus on just one part rather than the system as a whole, vastly over-estimating their willingness to overclock, and are willing to make long-term sacrifices for short-term benefits. So just because other people don't know how to research and buy a CPU properly does not provide additional justification for your reasons for getting higher speed RAM.

and then there is the i7 which doesnt help at all over a i5 in most games while costing $100 more

I agree with that.

those xbit results at a low res would be similar if you had say gtx980 sli at 1080p and were trying to run over 120fps for a high hz monitor

Prove it properly then.

yes the first two are from me but if you look through the thread you will see a lot of other people getting similar results with arma as its one of those games that is not a gpu bottlnecked console port that runs fine on a pentium as long as you have enough gpu grunt

They're still user reviews that aren't properly done and prone to confirmation bias or placebo effects.

20% is the difference between sandy and haswell so a 2500k with 2133c9 ram could perform much the same as a 4690k with 1333c9\1600c11 ram

20%? The difference between Haswell and Sandy Bridge IPC is roughly 8% to 10% in games. Nor have you actually proven that SB with higher speed RAM will perofm as well as Haswell with lower-speed RAM. That's just conjecture.

 

[dasa]

can you disprove it?
unfortunately nobody has tested it and i dont have the money to buy a heap of hardware but even if i did it seems you wouldnt believe my results anyway would you
but lower the res and you lower the gpu power required if you double the gpu power then the result is similar to lowering the res in that the bottleneck starts moving away from the gpu and over to the cpu as the fps clime

hmm maybe my memory is a little rough but i thought sandy to ivy was up to 10% and ivy to haswell was up to 10% more maybe it was only 5% each step though
but then if your looking at gpu limited games then the difference could be anywhere from (20% assuming this is the max) to 0% i guess thats something i need to go back over

if somebody doesnt want to oc by all means suggest 1600 c8 ram to go with a 4590\4690 but its bad value to put 1333c9\1600c11 ram with them when 4460 with the faster ram would perform similar for less outlay
if somebody does want to oc then i suggest 2133c9-2400c10 if the cost is similar to other ram
but if the budget is tight i agree gpu first cpu\ram second (unless arma\dayz are the main use for the pc) all i was saying is that ram does make a difference vs the first reply that said it doesnt

 

[dasa]

FX6300 vs i5 --There IS a noticeable difference in terms of CPU alone with lightly threaded games at resolutions that matter unless you want to do some 4.5ghz overclock on the FX6300 and use a lot of power while you're at it.

yes most people know this trouble is reviews like this which show them to perform the same now most tech savy people know thats just due to them testing gpu bottlnecked games or in the case of bf4 a gpu bottlnecked section of the game
http://www.anandtech.com/bench/product/1260?vs=1197
but then they go and test the same games with a 4770k@4ghz\hd6950 and everyone says ram doesnt make a difference?
http://www.anandtech.com/show/7364/memory-scaling-on-haswell/7

 

[Dangman]

can you disprove it?

You're the one who made that GPU performance claim so it's on you to prove it, not me. I have not seen anything that would prove your claim.

unfortunately nobody has tested it and i dont have the money to buy a heap of hardware but even if i did it seems you wouldnt believe my results anyway would you

Assuming that English is your first language, yes I would have a hard time believing you due to your grammar, punctuation, and post style.

but then they go and test the same games with a 4770k@4ghz\hd6950 and everyone says ram doesnt make a difference?
http://www.anandtech.com/show/7364/memory-scaling-on-haswell/7

Want to know how I know you didn't actually read the page? If you had, you would have noticed that the performance difference would not be noticeable to your common computer hardware enthusiast let alone everyone else. The majority of the games tested at that low resolution of 1360x768 saw less than a .5 FPS to 1FPS increase with a few 3FPS here and there. THose are not going to be noticeable at all to anyone during gameplay.

 

[dasa]

so by that logic a autistic person who has trouble communicating couldnt possibly be any good at math?

you got it all wrong that anandtech review is the kinda crap that everyone believes as proof that ram makes no difference its not a example of ram performance its just the performance of the gpu yet they call it a ram review
same goes for there cpu reviews showing that there is no noticeable difference between a i3 and a 4790k when there is
both are false yet you believe one and not the other

 

[melk]

How comfortable are you with overclocking the 5820k? I haven't read up much on what those owners are getting (edit, quick read of this thread seems to show that most are in the 4.2 or 4.3 range max.)

If you like to fiddle around and try to eek out every last mhz from your chip then either way you are getting a K chip to play with, the only difference being that the 4790k is a guaranteed 4.4 and has the potential (not much) to go even higher.

Unless you know for a fact that whatever work you are doing (not games) pegs all CPU cores at 100% then an extra 2 cores aren't going to do much besides sit there idling away in your Task Manager window

 

[faugusztin]

The 5820k is still a bit of a mystery to me though when it comes to ram, does it need to have 4 sticks, are two fine to start, its hard to find definitive answers.

Any number of RAM modules is fine. I was running it for around a month with a single 8GB RAM module, later when i got my 4x8GB kit i ran it with 4 modules.

 

[omega-x]

so by that logic a autistic person who has trouble communicating couldnt possibly be any good at math?

you got it all wrong that anandtech review is the kinda crap that everyone believes as proof that ram makes no difference its not a example of ram performance its just the performance of the gpu yet they call it a ram review
same goes for there cpu reviews showing that there is no noticeable difference between a i3 and a 4790k when there is
both are false yet you believe one and not the other

Fine, lay out some formulas that prove your point. It won't help you without benchmarks.

 

[dasa]

if the benchmarks i provided are not enough to even cast doubt on there being more than ~0-3% difference from ram in a cpu limited game for you to look into it more yourself then thats your loss
the thing that bugs me is that it means you could be passing on bad advice to others

 

[Dangman]

so by that logic a autistic person who has trouble communicating couldnt possibly be any good at math?

No, because at there's an excuse for that autistic person to have trouble communicating. Simply put, if you can't take the little bit of time to properly format, write, and ensure proper grammar for something , that casts doubt on your actual results.

you got it all wrong that anandtech review is the kinda crap that everyone believes as proof that ram makes no difference its not a example of ram performance its just the performance of the gpu yet they call it a ram review
same goes for there cpu reviews showing that there is no noticeable difference between a i3 and a 4790k when there is
both are false yet you believe one and not the other

Again, show proof. Yes you can try to muddle the waters all day long but without any sort of proof, you're pissing into the wind here.

the thing that bugs me is that it means you could be passing on bad advice to others

Ironically, that's what bugging me as well about your RAM advice: Bad advice to others.

 

[dasa]

well i know how it performs on my old sandy system but not on newer systems other than what i have seen from others on forums and hints in a few reviews that you are happy to dismiss

until somebody other than me does a proper ram review there is no way to prove it to you
do you at least understand where im coming from that its useless testing ram or cpu in a gpu bottlnecked game?

if im wrong then my advice costs somebody an extra $5-30 on a 1.5k+ build for a unnoticeable performance gain
if your wrong you may be gimping a 4690k system back to 3570k performance possibly even 2500k level wiping out ~4 years of cpu progress
not a noticeable difference in most games but in some it is

 

[Dangman]

well i know how it performs on my old sandy system but not on newer systems other than what i have seen from others on forums and hints in a few reviews that you are happy to dismiss

until somebody other than me does a proper ram review there is no way to prove it to you

If someone other than you who has the proper methodology, presentation skills, credibility, and can be held accountable to his or her reports/article can actually show NOTICEABLE gains in performance in games at 1080P or above, yes that would be sufficient proof for me. There's more than sufficient evidence on my side to discount the so-called advantages of higher-speed RAM for gaming. For Intel platforms any way.

do you at least understand where im coming from that its useless testing ram or cpu in a gpu bottlnecked game?

Until there's proof showing that the GPU was a major bottleneck, the testing wasn't totally useless. It may be an older GPU but it does not mean it was a major bottleneck for those games at that resolution.

if im wrong then my advice costs somebody an extra $5-30 on a 1.5k+ build for a unnoticeable performance gain
if your wrong you may be gimping a 4690k system back to 3570k performance possibly even 2500k level wiping out ~4 years of cpu progress
not a noticeable difference in most games but in some it is

If you're wrong, you costed someone an extra $30 in a $1000 build that could have gone towards a better HSF, PSU, case, motherboard, speakers, and/or larger SSD.

Even if I'm wrong according to your faulty data, he loses out on 8% performance going back to 2500K levels. Whoopty fuckin' doo. There's a reason why we've been telling people who have Sandy Bridge CPUs not to upgrade until Skylake comes out in 2016.

 

[dasa]

you have no proof that my data is faulty thats just your opinion

well dirt 3 shows some of the largest gains from ram in the anandtech review but nothing note worthy
here a 2500k@ 4ghz is the same speed as a i3 2100 in it with a 7970
http://www.tomshardware.com/reviews/gaming-fx-pentium-apu-benchmark,3120-7.html
is that proof that there tests are gpu bottlnecked?
i guess not since they are running a higher res and aa and yet there only 20fps lower

kinda hard when nobody tests at the same res it can only really be proved by the people that did the test so your asking the impossible

here they have AMD A8-5600K getting 96fps and a i7 3770k getting 109fps only a 12% difference between them? i think not while in there ram review at 1360 the 4770@4ghz only managed ~105fps odd but still nothing solid
http://www.anandtech.com/bench/CPU/340

also this suggests i may have been right about sandy-haswell with ~18% difference at stock clocks although it does shrink to ~10% with them both at 4.5ghz
http://translate.googleusercontent....7.html&usg=ALkJrhijELTmgETZsWGePHaRF9P9nLxzlw
17% difference both at 4.5ghz
http://pclab.pl/zdjecia/artykuly/radek/2014/centurion/charts/oc/fsx_1920.png
16%
http://pclab.pl/zdjecia/artykuly/radek/2014/centurion/charts/oc/sc2_1920.png
granted most games are much less

installed dirt 3 gave it a run with a r9-290 on my [email protected] 1360x768 ultra preset and managed 188fps avg
141fps 2600k@stock 1600c11
155fps [email protected] 1600c11
166fps 2600k@stock 2133 9-10-10

well thats proof enough for me that they were gpu bottlnecked in this crappy old console port and to think you had me doubting myself for a while there but obviously it means nothing to you unless you take the time to do it yourself instead of preaching how im wrong and your right
i will admit though that the dirt 3 benchmark is not all that consistent and i didnt do repeat runs on these tests sometimes you come first if there is a big crash other times you come 8th
but the % stacks up with a 12% increase in cpu clock speed comes a 10-12% increase in performance in those tests
also as the cpu clock speed increases so does the % gain from faster ram (unless you start running into a gpu limit) due to the cpu becoming more starved for bandwidth

 

[nycdarkness]

I upgraded from a 2600k @ 4.7ghz to 5820k @ 4.6 ghz. If you do any type of cpu intensive work there is a definite difference. Gaming not so much, but an improvement can be felt in general. Placebo or not.

 

[Tsumi]

So what if it's GPU bottlenecked with high end cards at 1080p or higher? It just tells us that the logical thing to do if the person was playing those games would be to put more money into the GPU rather than spending more on the CPU or RAM.

 

[dasa]

So what if it's GPU bottlenecked with high end cards at 1080p or higher? It just tells us that the logical thing to do if the person was playing those games would be to put more money into the GPU rather than spending more on the CPU or RAM.

yes
spending money on cpu\ram will do nothing at all for games that are gpu bottlenecked which is most games provided you have a half decent cpu

when trying to make use of a 144hz monitor there is a fair few games that can be cpu limited with enough gpu power

with a 60hz 2560 monitor and a single gpu there is only a few games that are still cpu limited so it really depends on what games you play

 

[Dangman]

you have no proof that my data is faulty thats just your opinion

ACtually I do have proof: You never actually properly set up the conditions for a proper test. As such, your data is inherently faulty. For your first link, you only ran two tests per clock speed. That's hardly sufficient runs for an accurate measurement of the average runs. For your second link, both your results are incompatible with one another since the system was not exactly the same since one of your 7970 cards died. In addition, in your own words:

there is obviously some inconsistency with the tests since i only bothered to run them all once with 1600 8-8-8 performing the same as 9-9-9 in one test and 9-9-9 performing the same as 10-10-10 in another this is just a hickup

So yes I do have proof that your data is faulty. NOt to mention that you yourself are going to point out the big flaws with your recent Dirt 3 tests.

installed dirt 3 gave it a run with a r9-290 on my [email protected] 1360x768 ultra preset and managed 188fps avg
141fps 2600k@stock 1600c11
155fps [email protected] 1600c11
166fps 2600k@stock 2133 9-10-10

well thats proof enough for me that they were gpu bottlnecked in this crappy old console port and to think you had me doubting myself for a while there but obviously it means nothing to you unless you take the time to do it yourself instead of preaching how im wrong and your right
i will admit though that the dirt 3 benchmark is not all that consistent and i didnt do repeat runs on these tests sometimes you come first if there is a big crash other times you come 8th /QUOTE]
So because your standards for properly done tests and proof are almost non-existent, that means you're right?

Again, if you want intelligent hardware enthusiasts to actually believe you, first and foremost, take the time to actually prsent your information in more clear and concise manner. I have had to re-read your posts multiple times due to their incoherency. Second, actually develop habits and habits for getting accurate information. Yes that also maintaining consistency as well.

also this suggests i may have been right about sandy-haswell with ~18% difference at stock clocks although it does shrink to ~10% with them both at 4.5ghz
http://translate.googleusercontent....7.html&usg=ALkJrhijELTmgETZsWGePHaRF9P9nLxzlw
17% difference both at 4.5ghz
http://pclab.pl/zdjecia/artykuly/radek/2014/centurion/charts/oc/fsx_1920.png
16%
http://pclab.pl/zdjecia/artykuly/radek/2014/centurion/charts/oc/sc2_1920.png
granted most games are much lessh

Whenever people talk about performance difference between different generations of CPUs, they're talking about IPC, not clock speeds of the various different models. In fact, your linked pictures proves my point in regards to the performance difference in IPC between Sandy and Haswell being roughly 8 to 10%. The Rome II figures shows only a 10% difference between Sandy and Haswell at the same clock speeds.

In addition your numbers for the last pictures are off. For FSX, the Core i7 4770k had 42.2 FPS and the Core i7-2600K had 35.9 FPS. That's a difference of 15%. There's only a 13% difference with SC2. Yes those are higher than my earlier quoted 10% but those increases could easily be chalked up to testing variances.

 

[dasa]

For your first link, you only ran two tests per clock speed. That's hardly sufficient runs for an accurate measurement of the average runs.

Actuly they are two different levels benchmarked and each individual test was run three times to make sure the numbers were consistent on the first two arma graphs i even charted the fps

For your second link, both your results are incompatible with one another since the system was not exactly the same since one of your 7970 cards died. In addition, in your own words:

So? its not like one card died half way through the tests its a new lot of tests carried out months later with the game updated and in a new graph and yet they still show large gains
Yes i admitted there was some variance but it was under 1fps!
For example if i averaged them out over several runs they may have been 1600c10=54.5fps 1600c9=55.5fps instead of both 55fps talk about nit picking

NOt to mention that you yourself are going to point out the big flaws with your recent Dirt 3 tests.

I ran the same test anandtech used they would see the same variance its a shit test but thats what they used so thats what i had to worked with i could run it more times to get a consistent result but you wouldnt believe it anyway so why bother
the fact that my fps were almost double theres simply due to a more powerfull gpu was enough for me to be satisfied that they were gpu limited
but if you want me to run it again at a higher res so that my fps drop to there level then show you how changing the cpu doesnt effect fps i can

my apologies that is the % increase i listed maybe there just test variance but maybe there not you dont know

edit
just realized you were talking about my second link first
with thief the first run was different to the second but the third and fourth run were always the same as the second (within 1fps) so after a few runs i stopped testing them and didnt see any point including the data

 

[R3MF]

Newegg sent a weekender sale flyer. A 5820k and the MSI X99S SLI Plus for $519.98 PLUS a $20 rebate from MSI. $499.98 makes the extra price of DDR4 almost not as prohibitive compared to the 4790k/Maximus VII rig I was considering. The money part of the equation is always part of the decision process when discussing 4790k v. 5820k but at this moment, not so much. The 5820k is still a bit of a mystery to me though when it comes to ram, does it need to have 4 sticks, are two fine to start, its hard to find definitive answers.

After so much going back and forth the things I've deduced are:
-When the 4790k becomes obsolete, I'm sure the 5820 will be right there with it but with close to equal pricing the 5820 may be worth it.
-Buying DDR4 now with the sole idea of future proofing seems somewhat foolish as DDR4 in a few years from now will be a different beast(think early DDR3 vs the DDR3 2400 sticks today that really take advantage of Haswell's memory controller). It would get you up and running on a future build though.

Ahh, first world problems. I really do need an upgrade, I'm embarrassed to even say the rig I'm running on right now. I tend to make computers last for a long time though-think 7 years...

I use this for gaming, general use and if I get off my lazy butt, more music production. I will be buying a GTX970 as well(I'm on a BenQ GTG XL2720Z I just purchased)

5820k gives you enough PCIe 3.0 to have a 4x 3.0 NVME SSD and an expansion card for USB 3.1 type C.
That will never be possible on s1150.

that is what decided me for the 5820k/X99 SLI-Plus

 

[kierwest]

5820k is overkill on gaming. Not enough performance difference to warrant the price.

 

[nycdarkness]

5820k is overkill on gaming. Not enough performance difference to warrant the price.

The price argument really depends on 2 things

#1) Do you live near a microcenter

#2) Do you care for more than 1 gpu

If both things are true the price difference is something like 40 dollars.

 

[Arct1c0n]

I upgraded from a 2600k @ 4.7ghz to 5820k @ 4.6 ghz. If you do any type of cpu intensive work there is a definite difference. Gaming not so much, but an improvement can be felt in general. Placebo or not.

I went from a i5 2500k at 4.6 Ghz and I can tell a difference on my 4790k, I have more cpu headroom then before, expecially dragonage where my 2500k was really getting a workout. Future games are gonna really benefit from a i7s extra grunt. Witcher 3 and Star Citizen are gonna be proof of that

 

[UnrealCpu]

I went from a i5 2500k at 4.6 Ghz and I can tell a difference on my 4790k, I have more cpu headroom then before, expecially dragonage where my 2500k was really getting a workout. Future games are gonna really benefit from a i7s extra grunt. Witcher 3 and Star Citizen are gonna be proof of that

haswell processors have a huge improvement when it comes to emulation especially dolphin

own a sandy bridge i5 3770k now 4790k

unless you like doing movie editing , photoshop, other cpu related intensive applications then get a hex or octo core processor.
Gaming hands down the i7 4790k is the best on the market and soon to be broadwell

Once good mobos come out for skylake i will upgrade to a hex or octo core in 2016.

Right now only a few games use 4 cores , ARMA for example. Put your money into a video card like maybe a upcoming 390x 6gb or new Nvdia 6gb card in the future

ops just noticed you bought a 970 and that benq

really to be honest you really cant keep your computer 7 years and expect it to perform well. Computer parts depreciate so fast you are better off selling off components every 6 months to a year and just upgrading by investing a few hundred a year.




Also put more money into your monitor and get one of the best IPS monitors you can buy . I am waiting for the new Asus 120hz IPS coming out in the next 60 days

 

[Zohar78]

not to thread crap or steal, but it sounds like the i5-4690k is a pretty good deal. games is about the only thing i do that pushes the computer hard, most of the performance of a 4790 chip. Im looking to upgrade in a month or two, and that looks to be the one. Now to hopefully get an idea of a motherboard to go with. coming from an q8200.



I think i posted in the wrong thread... bu

 

[durquavian]

haswell processors have a huge improvement when it comes to emulation especially dolphin

own a sandy bridge i5 3770k now 4790k

unless you like doing movie editing , photoshop, other cpu related intensive applications then get a hex or octo core processor.
Gaming hands down the i7 4790k is the best on the market and soon to be broadwell

Once good mobos come out for skylake i will upgrade to a hex or octo core in 2016.

Right now only a few games use 4 cores , ARMA for example. Put your money into a video card like maybe a upcoming 390x 6gb or new Nvdia 6gb card in the future

ops just noticed you bought a 970 and that benq

really to be honest you really cant keep your computer 7 years and expect it to perform well. Computer parts depreciate so fast you are better off selling off components every 6 months to a year and just upgrading by investing a few hundred a year.




Also put more money into your monitor and get one of the best IPS monitors you can buy . I am waiting for the new Asus 120hz IPS coming out in the next 60 days

With dx12 the possibility of even longer life CPUs becomes more plausible.

 

[Voxata]

If you like to overclock easy, go with X99.. If you're willing to delid, go 4970K because sky is the limit on this chip, I do 4.8 @ 1.32, 5.0 is unstable for me at 1.4, not pushed for 4.9 yet I don't think I will either given the voltage curve after 4.8. The moment you push 4.3-4Ghz+ on all cores before delidding You'll be 100C and throttling in Prime28/IBT.

 

[Elios]

just upgraded to an i7 4790k @ 4.5Ghz from an i7 860 @ 4.0Ghz
and it is a huge speed up
very happy with it

 

[sk3tch]

Lots of great information and opinions in this thread. However, OP - you just need to decide what you want. Sometimes it does not need to be majorly logic/fact based as an enthusiast. Do you want Z97 or do you want X99? X99 will have much longer legs than Z97 as a platform. Does that mean much for gaming? Not necessarily. But if you think you're the type to swap out hardware/etc. etc. then the investment in X99 can pay off. Otherwise, yeah, go Z97 it cheaper and faster for games.

 

[UnrealCpu]

when skylake is released you will wish you just bought the 4790k so you can sell it and upgrade with your z97

 

[UnrealCpu]

not to thread crap or steal, but it sounds like the i5-4690k is a pretty good deal. games is about the only thing i do that pushes the computer hard, most of the performance of a 4790 chip. Im looking to upgrade in a month or two, and that looks to be the one. Now to hopefully get an idea of a motherboard to go with. coming from an q8200.



I think i posted in the wrong thread... bu

i paid 299.00 after rebates at microcenter for a 4690k and a gigabyte g7 gaming

does 4.7 with the H110