R-T-BIt's not nostalgia. It's that the money to performance ratio just isn't there. Another video card would've served me far better.

Mind you, I was going from a 990x i7, not quite the same as the 920... Same damn core though minus AES instructions.

well then if it was a 990X i understand

tho i still find the fact of being deceived delusional since even if you went from the higher 1366 to the lowest 2011-v3, even if you took the lowest it should still yeld substantial advantage over ... (one beingt the PRICE... i mean com'on people still try to sell the 980 and 990 above a 5820K price ... ) well the 990X was a bit more efficient (5%) on the tdp but at 999$ versus 389$ i don't think a 5820K has a worse price to perf ratio ..., roger?

apple to apple a 990X successor is a 5960X for me

Caring1They do, they're called XEON!

There are Xeon's with iGPUs.

R-T-BMore like HEDT

This. More like a real server platform that has use no use for an iGPU. Most consumers will probably use an iGPU so it makes sense that most mainstream CPUs have an iGPU on them.

GreiverBladewell then if it was a 990X i understand

tho i still find the fact of being deceived delusional since even if you went from the higher 1366 to the lowest 2011-v3, even if you took the lowest it should still yeld substantial advantage over ... (one beingt the PRICE... i mean com'on people still try to sell the 980 and 990 above a 5820K price ... ) well the 990X was a bit more efficient (5%) on the tdp but at 999$ versus 389$ i don't think a 5820K has a worse price to perf ratio ..., roger?

apple to apple a 990X successor is a 5960X for me

990x was purchased years later at well below msrp (I got it for about $350)... Upgraded from a 920 then.

Don't get me wrong, I don't feel decived, I just feel Intel could be doing a lot more with this die shrink given proper competition...

R-T-B990x was purchased years later at well below msrp (I got it for about $350)... Upgraded from a 920 then.

Don't get me wrong, I don't feel decived, I just feel Intel could be doing a lot more with this die shrink given proper competition...

tho Skylake is not a die shrink .... broadwell was,

ok i get your meaning and i agree.

I hope it doesn't cost the same as broadwell...

I feel a small need to upgrade from a 990X.

RejZoRThe 6700K CPU should be a hexacore. It's 2015 and they still consider a quadcore to be "enthusiast" level. C'mon, really!? I see ZERO point in switching and I have a Core i7 920. Only thing that I'd realistically gain is power consumption and some new instructions. Do your math how long can I use my existing i7 920 to justify the price difference in electricity bills...

From what I've checked, everything is identical. Cache sizes, core count, thread count etc. Hell, I even have triple channel on my ancient grunt and Skylake is only dual channel. Like ugh!? Totally pointless product. It only makes sense if you don't have a computer and you're buiying from scratch. Or you have some shitty dual core from 10 years ago...

I'll have to disagree with you, I upgraded from i7 920 like you which I had overclocked to 3.8 Ghz on all cores. When I got my 4770k, at stock it was faster in almost every game. Same video card 2 x Radeon 7950 in Crossfire.
Once I overclocked the 4770k to 4.4 Ghz, it became about twice as fast so it's not a linear comparison.
And then I got a 780ti and haven't had to think about upgrading yet.

I can't imagine going from 920 to 6770, that would be a huge increase.

Clock for clock, I had no decrease in performance going from triple channel to dual channel either, EXCEPT when I mad a RAM Drive, but then again, what's the difference between 8GB/s to 3GB/s? That sounds like a lot, but it made 0 difference in load times on the games I played on that ram Drive.

* cough * bullshit * cough *

Unless you take only most CPU bottlenecked games, you'll see no real difference.

Change entire platform for 800-1000 € and gain 2fps or buy a new graphic card for half that and gain 30+ fps... hm...

How does going from a 920 to a 4790k effect minimum fps. Average fps there would be no difference, but minimum there should be a noticeable difference. I mean noticeable in game. If you average 80 fps but drop to 40 fps that is bad, as compared to 85 fps average and 60fps minimum.

RejZoR* cough * bullshit * cough *

Unless you take only most CPU bottlenecked games, you'll see no real difference.

Change entire platform for 800-1000 € and gain 2fps or buy a new graphic card for half that and gain 30+ fps... hm...

Well it's not bullshit, but I did have to get a new Motherboard, Processor, RAM etc. So other efficiencies were also added in which probably contributed to the increase in performance.
Same video cards though, those beasts lasted me a long time, loved those Windforce 7950s.

james888How does going from a 920 to a 4790k effect minimum fps. Average fps there would be no difference, but minimum there should be a noticeable difference. I mean noticeable in game. If you average 80 fps but drop to 40 fps that is bad, as compared to 85 fps average and 60fps minimum.

I've tested it easily. Stock i7 920 and overlocked i7 920. And I could hardly notice any difference until game started hammering CPU very hard. Just a mild bump to 3,2GHz and i frankly couldn't notice much difference compared to 4,2GHz. I basically run it at this clock just because I can. No other reason.

RejZoRI've tested it easily. Stock i7 920 and overlocked i7 920. And I could hardly notice any difference until game started hammering CPU very hard. Just a mild bump to 3,2GHz and i frankly couldn't notice much difference compared to 4,2GHz. I basically run it at this clock just because I can. No other reason.

We're not talking about the same thing though, I'm not talking about just overclocking the 920, I'm talking about changing to a new CPU architecture. The difference between the 920 and 4770k at the same clock speeds, was huge. (Exaggerated a little)

If i7 920 at clocks beyond 3,2GHz shows very little difference, what makes you think 4770k would make any difference? It's still just a CPU doing the exact same workloads as 920... It's not like these new CPU's are so incredibly radical for them to make a difference by that alone. Realistically that happened with Nehalem when they re-introduced HT which was pretty much useless on P4's because there was no software to really use it. HT is a very radical approach at doing computation. We haven't seen anything similar since...

There are various specualtions about eDRAM on Skylakes (probably Skylake-E variants), but no one really knows how they'll use it. Could be just for iGPU and wouldn't make any difference for anything else or they might use it for general CPU tasks and that might be interesting. But other than that, nothing really new worth mentioning. Die shrinks are like "meh" these days really. It's nice to have it, but essentially yawn inducing...

There's a lot more to CPUs than just clock speeds man......you know what I do have the 920 at home, I'll try it tonight with some CPU benchmarks.

I don't care about synthetic benchmarks. I care about real world performance and differences you can actually see and feel.

EDIT: Just realized I have an i7 930, not 920. Sorry about that.

RejZoRI don't care about synthetic benchmarks. I care about real world performance and differences you can actually see and feel.

Hey man I'm right with you there, I want to know if it was all in my head or was real. So I did start with a synthetic benchmark, I just used the benchmark tool built into ThrottleStop. Both processors are running all cores/ht at 3.8Ghz.

i7 930:


i7 4770k:



So for that benchmark, it went from almost 12 seconds to 7 seconds. Not bad.

Here I'm trying waframe with everything graphical turned down (including resolution) to keep GPUs out of the equation as much as possible.

i7 930:



i7 4770k:



If you check the FPS counter at the bottom, 930 is around 548 fps, while the 4770k is around 847 fps.
Almost double the performance. Okay so it wasn't just me, there is a massive difference.

EDIT 2: Even though I had everything graphical turned down, the GPUs were still being used a lot. The Radeon 7950 with the i7 930 was at 45% usage, and the 780ti with the i7 4770k was at 32% usage, so take those results with HUGE grain of salt man.

I really applaud the effort to actually do some benchmarks. That is a 55% boost in performance clock for clock. That is very nice.

I don't know if warframe is the best game for testing if it is noticeable. Yes it shows a real world boost, but you were already at 550 fps. Anything more is not even noticeable. I don't know of any game, but it would be more ideal to have a game that was getting low fps with the 930 and 55% more fps with the 4770k.

AluminumHasteIf you check the FPS counter at the bottom, 930 is around 548 fps, while the 4770k is around 847 fps.
Almost double the performance. Okay so it wasn't just me, there is a massive difference.

EDIT 2: Even though I had everything graphical turned down, the GPUs were still being used a lot. The Radeon 7950 with the i7 930 was at 45% usage, and the 780ti with the i7 4770k was at 32% usage, so take those results with HUGE grain of salt man.

While it is important to note that the difference is there, the gap between those 2 CPUs close in when the graphical settings are set to realistic points. Now, I'm not even sure to guess how many people who get CPUs like the 4770k/4790k will have a good GPU to accompany it, but using that as a base, if graphical settings were scaled properly when comparing games of noticeable grunt, the CPU does little to increase a playing experience for games that are GPU-bound (Ex. Crysis 3, Far Cry 4, Dragon Age). Hard to justify a CPU upgrade when it will only net you ~10% gain while spending $300+.

I feel like CPUs are no longer a factor in this day & age when considering a PC gaming build. Seems like even if you get a low-end/mainstream CPU and just give it a small OC, it will keep you going for a long time. Case and point, my 2500k has not slowed me down one bit. I might wait for Cannonlake or Zen depending on how AMD plays it's cards this time around.

RejZoR* cough * bullshit * cough *

Unless you take only most CPU bottlenecked games, you'll see no real difference.

Change entire platform for 800-1000 € and gain 2fps or buy a new graphic card for half that and gain 30+ fps... hm...

You're absolutely right.
I've upgraded recently from a Phenom II X4 to a Haswell i5, and although I get ~40% more performance in pure CPU synthetics, 3dmark score has hardly moved and I don't feel any significant improvement in games. There is some improvement during scenes where frame drops, if you usually game with FRAPS on, but nothing which you could notice if FRAPS would be off, and drops and stutter if they were there before they are still there even after the upgrade.

Even badly coded games, which rely heavily on single thread performance, like Starcraft 2, which was the main reason for which I went i5 don't feel that much better. (Badly coded because this game uses ~70% from one core and 40% from another ... rest of them relax doing nothing)

Anyway I upgraded mostly because my platform was very old, because I could find good looking settings which got me over 60 fps in all the games even on the phenom.

BTW I also have a 7850, which I want to upgrade, but again it is because of its age and because at 1GB it has quite low memory, not because it doesn't give me good performance, in fact it is still a monster running flawlessly everything I throw at it.

One more thing about the memory, I don't see how an iGPU even with 128 EDRAM or whatever can compete with a card with dedicated memory. This iGPU stuff is very good for laptops, otherwise for desktop it is completely useless and any reasonable gamer will eventually get a discrete card.

This makes this intel iGPU push very stupid from my perspective, from the same die size they could make 2, maybe 3 CPUs, (even with slower 2d/3d integrated graphics) and sell them a bit cheaper and with much better margins.

RejZoRShow me the difference in games between i7 920 at 4,2GHz and that Skylake. It'll probably be identical. Paying premium for 3 seconds less in 7zip compression, I couldn't care less...

Funnily enough those benchmarks aren't out yet, however due to the performance increase they will certainly not be identical. I myself upgraded from a 920 at 4Ghz to a 4930k at 4.5Ghz, and the difference was noticeable before I even overclocked the 4930K. CPU limited games like WoW benefitted and allowed me to raise settings, GPU limited games only saw a minor increase in max FPS however the average FPS took a noticeable bump, again allowing settings to be raised.

If you're getting 500+ fps in a game, of course you'll see huge differences. It's EVERYTHING down to CPU to make ANY more difference. You're already at such ridiculously high framerate it doesn't even matter anymore if it's 500 or 1500 fps.

john_Cheap prices where a necessity for AMD most of the time. Even when they had faster processors, Intel was controlling the OEMs. Every OEM, or retail store was selling Pentium 4s. After that the Phenom processor wasn't that fast and the Bulldozer architecture a pure disaster. So how can you go out and charge equally or more? The competition is controlling the OEMs, the retail stores, the press. People are used in blaming AMD for the same things they will find plenty of excuses for Nvidia or Intel. When everything is against you, can you really expect to empty your warehouses with prices that are not ultra competitive? Fury X come out at the same price as 980Ti and guess what. Everyone was looking the second decimal on the fps counter to say that the card was a failure. Suddenly a pump noise was compared to a jet engine and tech sites rush to write articles about how AMD failed there. And they rushed because it was already known that the problem was fixed. When everyone is pointing a gun at you, can you really charge extra?

Boom. There it is.

Moving on to CPU, and the performance-increase is a predictable 10-20% single/multi-thread CPU performance, over "Broadwell." This is roughly similar to how "Haswell" bettered "Ivy Bridge," and how "Sandy Bridge" bettered "Lynnfield."

I don't recall Haswell beating Ivy by 10-20%, and neither Ivy beating Sandy the same way. Indeed Sandy was a big leap from Lynnfield, but after Sandy the progress was much smaller, it was more evident on the iGPU side.
:lovetpu:

I think I'll go with the upgrade anyway. I need to refresh my system since it's acting a bit funny lately and buying new LGA1366 board just doesn't seem viable at this moment (and hard to actually do since they are rather hard to come by these days)

A Core i7 6700K paired with 32GB of RAM, new AX 760i PSU and new case, probably Corsair Carbide Silent. I'll be keeping the graphic card, soundcard and HDD/SSD. Should be fine for quite some time even without any overclocking. If it'll last for 6 years like this one I'll be happy. Lets wait for august...

RejZoRShow me the difference in games between i7 920 at 4,2GHz and that Skylake. It'll probably be identical. Paying premium for 3 seconds less in 7zip compression, I couldn't care less...

Maybe you should consider yourself to watch this Intel Skylake Core i7 6700K vs 4790K/3770K/2600K Gaming Benchmarks comparison. Well, I know that video is not directly compared i7 SkyLake to your i7 920, but seeing it faster than i7 3770k which I know is many times faster clock for clock than your 1st gen i7, you can notice the difference.

And click this for Synth. benchmark from anandtech.
Well, don`t get me wrong, but I just want to make you know about how every generations of i7 improved it`s performance.