I7 7700k

[kaydub 3]

I went from i7 3770K OCed to 4.5Ghz to i7 6700K OCed to 4.5Hhz in 3 years and saw some improvement with P3D

 

But its not as much as I wanted or expecting.

 

I think we have to wait 6-8 years between upgrades before seeing real big improvements. The positive to this is, you don't have to spend money every 2 years. LOL

Yep.  I don't worry about performance anymore.  I worry about my overclocked CPU degrading because if it fails I can't replace it.  The technology that I am on isn't available.  If my motherboard fails I will have to buy memory, a motherboard and a cpu, and possibly a new OS if I can't get W7 to install.  All for very little performance gain.

 

Kdub

[Ted Striker 338]

I went from i7 3770K OCed to 4.5Ghz to i7 6700K OCed to 4.5Hhz in 3 years and saw some improvement with P3D

 

But its not as much as I wanted or expecting.

 

I think we have to wait 6-8 years between upgrades before seeing real big improvements. The positive to this is, you don't have to spend money every 2 years. LOL

Thanks for posting that Manny. I am currently on an I7 3770K at 4.5Ghz and was thinking about upgrading. Think I'll just stay with what I have.

Ted

[jabloomf1230 2,445]

Yes, but it's getting to the point where upgrading to a new PC is no longer cost effective.

 

 

Recently, Intel hasn't been able to improve single core performance very much from generation to generation. These days, the only way to improve performance is to upgrade one's GPU. And even that approach is getting strangled by the CPU bottleneck. Look at these single thread benchmarks:

 

https://www.cpubenchmark.net/singleThread.html

 

The i7-4790k even beats the i7-6700k by a small amount.

[martin-w 3,234]

The i7-4790k even beats the i7-6700k by a small amount.

 

 

It depends who's doing the testing.

 

Single core 6700K 6% up...

 

http://cpu.userbenchmark.com/Compare/Intel-Core-i7-6700K-vs-Intel-Core-i7-4790K/3502vs2384

[SteveW 1,792]

In Jays example the results are normalised to equivalent fixed frequencies, in Martins example the 6700 results are skewed as benefits from a better motherboard chipset and memory, differing frequencies and boost. The 6700 is slightly slower than the 4790 as the increased heat density of the 6700 delivers less work by around 8% but well compensated for with the later z170 chipset results in as good as or slightly better performance.

[Christopher Low 5,352]

In other words, nobody with an i7 4790k should even be thinking about "upgrading" :wink:

[martin-w 3,234]

In other words, nobody with an i7 4790k should even be thinking about "upgrading" :wink:

 

Of course not, no one has said anyone should... unless they are well off enthusiasts that just love new kit.

 

I came from a 3770K, upgraded because my son was having my old rig. Glad I did though, I'm thoroughly thrilled with Skylake and the GTX 980Ti.

[Christopher Low 5,352]

That was just a little side comment, Martin. I wasn't suggesting that anyone on this thread had mentioned it.......but I have seen posts elsewhere to suggest that others have considered it.

[ryanbatc 12,832]

I'm still on a 2500k lol

 

I expect a large perf increase when I upgrade

[martin-w 3,234]

Don't know about the sim, not installed yet, but Iv'e just been playing the BF1 beta at between 140 and 180 frames per second on Ultra settings.

[garrett_frank 150]

So it doesn't seem like there's a consensus as to whether the 6700k is all that much better for P3D than Sandy or Ivy Bridge. I have a pseudo-silicon-lottery 3770k that I can run at 4.9Ghz at well under 1.5v and well under TJMax (stress tests under 90*C and real-world P3D use (if that makes any sense) is in the mid-70s). All thanks to an AIO cooler. With 2200Mhz RAM at 9-11-11-31-2T I get an XTU score of 990 with HT off (which I keep disabled when simming anyway).

 

All of that said for some context. The more apples-to-apples comparisons I've used simply involves looking at XTU scores on HWBOT.org for 6700ks around the same frequency. Granted the RAM is different (I found some G.Skill that runs at 3200Mhz at CAS14 so I've been using that in score searches), I've seen XTU scores averaging around 1600, a ~60% increase over my score. I know that XTU likely wouldn't translate to a proportional performance gain in P3D but it seems like it could be far more than anecdotes in this thread indicate... Thoughts?

[martin-w 3,234]

So it doesn't seem like there's a consensus as to whether the 6700k is all that much better for P3D than Sandy or Ivy Bridge.

 

 

 

Well yes, it's better.

 

6700K is 37% better than 3770K in single core performance, taking into consideration all else that makes up the new platform, motherboard chip-set, ram etc. The point being made in this thread is that gains are minimal compared with the previous architecture. In my opinion if you are on Ivy Bridge or older, an upgrade may be worthwhile.

 

I can confirm how good Skylake is compared with Ivy Bridge tomorrow when I install FSX.

 

Intel seem to be working to a triphasic pattern rather than the old "tick tock". "Process" where an existing design is modified slightly in order to shrink it, "architecture" where a new architecture is introduced, and finally "optimisation" where the new architecture is revised. 

[ryanbatc 12,832]

By the time I'll be able to part with the cash for a new system (darn house projects!) it'll be too late and all gaming will be done on mobile BAH!

 

I hate how that's where the trend is right now.  I've grown up in some of the best PC game releases of all time - and there just aren't many out there anymore.

[garrett_frank 150]

Thanks Martin. The biggest problem we, as flightsimmers, have insofar as CPU news and benchmarks go is that they're almost always in the context of the rest of the gaming world. The mainstream gaming universe has been scoffing at Intel with the last several generations because they don't see 5-10% IPC improvement as substantial since, lo and behold, few mainstream gamers are using titles that are CPU bound.

 

So all we have are synthetic benchmarks. That being said, the reports based on the leaked SiSoft benchmarks don't show Kaby Lake to be much of an improvement over SkyLake. But, as Martin points out, Kaby Lake is the first release in Intel's 3-phase product generation (Kaby Lake being the 3rd and final). So perhaps this isn't surprising? That the new 3rd step, "optimization," is power and efficiency, not so much IPC, at least on the unlocked desktop models?

 

Kaby Lake will also, apparently, only work on Windows 10. As much as I've resisted, I may move there regardless since Windows 10 at least supports DirectX12. If only we had word that LM was either about to make performance improvements with P3D or if an upcoming version would utilize some of the CPU threading capabilities of DirectX12.

 

So many questions, so few answers. Upgrading now, anyway, may be the only option...

[westman 257]

Think that win7 gone work on the KabyLake 1151 series as on the Skylake.

and if not they gone have a unoficial bios for it , like the Skylake win XP up to Win10 bios.