Question I-5 6600k vs I-7 7700(non k)

[Gmoney06ss]

So my work is recycling all the old pc's. All they really care about is getting rid of the HDDs so the cpus are free to me. I grabbed a 7700 and I've been doing a little reading about it vs my current 6600k. My current 6600k is oc to 4.2, though I've had it up to 4.5.

From what little reading I've done, it seems as though the i7 Will slightly outperform the i5. And like I said it was a free cpu, so is it worth it to swap it in? I know it won't overclock, and I'm planning a cpu refresh late this year or early next. But figured for zero cost, if it performs better, why not do it for now.

The rest of my specs:
Asus maximus viii
Gskill ripjawz v 3000mhz 16gb
evga rtx2080
corsair hx750i
full custom loop

Any insite is welcome. Thanks.

 

[Gmoney06ss]

It is very rare for an Intel processor to go bad.
One real possibility is that a pin in the cpu socket was damaged during the process.
They are delicate and specced for only 15 insertions.
When you go back to the 6600K examine the pins closely.

Well dang, glad you mentioned this. Just removed the 7700 and have several bent pins from the looks of it. Thought I was careful but, it happened.


Any good tips for straightening? Or is the mobo just toast?

 

[NightHawkRMX]

View: https://www.youtube.com/watch?v=O1H5_FVX9lU
Bent pins can be repaired, but its not easy.
Sometimes the damage is too far.

 

[Gmoney06ss]

Pin surgery done. Couple bent pins fixed. Putting it all back together now. Any luck and I'll have a working cpu again. If not I'm guessing it's build time. Which kinda sucks.

 

[mdd1963]

Pin surgery done. Couple bent pins fixed. Putting it all back together now. Any luck and I'll have a working cpu again. If not I'm guessing it's build time. Which kinda sucks.

Awesome news!

 

[Gmoney06ss]

She booted!! Running some stress tests to verify stability now. So far so good!! Cb20 done. Going to try the games that crashed it first and go from there.

@remixislandmusic thanks for the vid, it certainly helped.

@geofelt thanks for the heads up on possible pin damage. Honestly the though never wouldve crossed my mind.

Luckily it was only three pins and it was an easy enough process. Temps are stable and most important no more leds or q codes! I'll post my finished thoughts after some more testing.

Thanks all!!

 

[Gmoney06ss]

So far so good!! Seems to be stable, have ran several benchmarks as well as some gaming.

Attempting to enable xmp, but have a question. When enabling xmp it raises the bclk to 102.3. I haven't booted with this setting, as I've always been told not to touch bclk.

Should I run it like that, or just manually oc the Ram? I've never done that before, but planned to copy xmp specs and just manually input. I made a post in the memory section, but figured I'd ask here, since there is people watching it.

Thanks

 

[Gmoney06ss]

After a few hours of runtime and some heavy benchmarking/stress testing I feel confident in posting my final results.

The cpu seems to average 4.1ghz, leaving mce on. Slightly higher with xmp enabled, but very similar, 4.15ghz.

Cinebench R15 single - 168, multi - 839. With xmp enabled multi jumps up to 874.

Cinebench R20 multi - 2029. Xmp 2116. I didn't run single, as the cb15 score was so close to the 6600. Didn't really see a point.

Timespy - gfx - 11,194 cpu - 4773 score - 9314.
With xmp and a very slight gpu oc, +100 core and +500 mem. Cpu - 5108 gfx - 12009 score - 9985. I'm sure I could finally break 10k with a few more minor gpu tweaks.

Firestrike - gfx - 27601 com - 7361 phy - 12232 score - 18860.
With xmp and the same gpu oc. Gfx - 29164 com - 7889 phy - 12558 score - 19866. Again 20k probably easily with a few tweaks. (Quick edit. Timespy final score with +125/+1000 is 10195, firestrike score 20245!)

My realbench scores jumped an insane amount, but I didn't record the 6600k in here, so I'll skip it.

Onto some gaming. All setting maxed at 2560x1440, no gpu oc or xmp.

Doom 2016. Cpu usage came down quite a bit averaging in the mid 60s with a peak to @88. Gpu and fps stayed the same. No difference in gameplay as it was already as smooth as possible.

Resident evil 2. Again a late drop in gpu usage. Averaging low 50, with peaks in the mid 80s. Gpu usage stayed the same, but it did pick up a little fps, averaging around 125.

Ghost recon. Cpu usage dropped again, with averages in game of 65. Benchmark showed min 45.7, max 81.8, and an average of 57.5. Gpu and fps stayed the same, but gameplay was a whole other story. So much smoother with no stuttering that I saw.

Both far cry titles saw very similar results, with minor bumps in fps and average cpu usage in the mid 60s.

GTA V saw large swings in cpu usage, with only peaks in the 80s, average were much lower, and a slight fps average bump of 10-15. But gameplay benefited most. Smooth as butter with no stuttering at all.

Battlefield. Similar story here. But usage was staying much higher in the 90s, just the gameplay was sooooo much better. I can actually play now, no more twitch or stutter. Stable fps, with averages in the 80s.

Overall very pleased, especially for free. As stated I didn't see any massive fps increase, but gameplay, oh my the gameplay. Better in every way. Games are now smooth, no more lag or stutter. I never felt my system was holding me back. But after these tests, I can see it was slightly cpu bottlenecked, as much as I dislike the term. The upgrade really opened up the system and let it run so much better.

 

[boju]

Good stuff, glad it worked out.

Thanks for giving your time to analyze, similar to how my pc performs with HT on.

When running games in 1440p, do you turn AA off? I find the picture is fine without it on my 27" and usually gives more fps.

 

[Gmoney06ss]

Good stuff, glad it worked out.

Thanks for giving your time to analyze, similar to how my pc performs with HT on.

When running games in 1440p, do you turn AA off? I find the picture is fine without it on my 27" and usually gives more fps.

Of course. It was honestly pretty fun and a good learning experience. I'm quite surprised at the results. I expected a little change, but not as much as I actually saw. Especially with gameplay, the overall experience is much smoother and makes playing even more fun!

I also have a 27", but no, I usually have AA turned up pretty high. I'll have to attempt to turn it off and see. But as long as it's smooth, and not stuttery, I tend to go for eye candy vs fps. But if there's not much in the way of visual impact, I definitely won't say no to more fps.

 

[boju]

I've been involved in a few discussions here debating hyperthreading in modernish & modern games and there are few believing there wouldn't be much difference or even at times make things worse. They can think what they want, everyone is entitled to their opinions ;P

 

[NightHawkRMX]

If you look at older cpus, you will see how much hyperthreading helps.
An I5 3470 with 4 cores and no hyperthreading, it performs well overall, but will struggle for playable framerates in some demanding modern games.
An I7 3770 with 4 cores and hyperthreading will most certainly perform better in modern games.

 

[boju]

True, older i7s are still showing relevance from my own experience, probably not for too much longer though and i agree, the i5s in the same era have pretty much left the scene awhile ago.

Id be bold and say that Sandy/Ivy i7s are giving later gen i5 4cores a run for their doe. Not in terms of frame rates but overall cpu usage stability at the frame rates the cpu is capable of.

Any cpu capable of setting a decent fps cap needs the resources to provide otherwise they run themselves to slavery by the gpu and become overwhelmed. Same deal i believe extends to the latest i5's going by others experiences.

 

[geofelt]

The real test is if your gaming/fps gets significantly better with 8 threads vs.4
Do not be impressed if your cpu% drops with 8 threads.
That is simply windows distributing the cpu activity across all available threads.
In fact, as strange as it might seem, it might indicate that you are only getting marginal benefit from the extra threads.

Here is a series of benchmarks for some games that explore this.
If your games were on the list of those tested that will be very interesting:

Gaming benchmarks: Core i7 6700k hyperthreading test (2018 update with GTX 1080 Ti)

Back in January 2016 i made a test about Intel Core i7 6700K virtual cores performance numbers in gaming. Short story back then was that hyperthreading offered no FPS improvement in any of my tested 21 games. Not only that, but about half of the tested games showed slight performance decrements...

www.techpowerup.com

 

[boju]

Geofelt, fps isnt drastically increased i agree, that's not what i nor op is saying. What we've both experienced is performance stability and smoother game play not having the cpu running too hard.

Comparison or benchmark sites with any cpu, fps cap is one thing but cpu usage during?, is another story and is rarely if not ever covered and is misleading. Plenty of people are disappointed in 8th/9th gen i5s in cpu intensive games at high refresh rates because they stutter with high activity. But guess what, benchmark sites say they should be getting decent fps. They can but at a cost of stability.

 

[Gmoney06ss]

There were fps gains in all but two titles, some were small, but some upwards of 20fps. Doom and wildlands being the no gain. Doom was already maxed at 200. The main impact was playability, it made everything much smoother and responsive. Where I saw lag or stutter before was gone, most notably in bfv and GTA v. So average fps may not have changed much, but min and lows did.

In my case, I found it to be a worthy up grade. Had I paid for it, it still would have been worth it, but maybe not as much. Would I recommended you rush out and buy a cpu with ht vs your current non ht Cpu? Only if you feel your being held back.

Still happy with my results.

 

[geofelt]

Happy is what counts

If/when you need a cpu upgrade, look at this review of the 9700K.
In particular look at where even a small difference in clock rates is more important that the increase in thread count .
I might note that the games tested do not show the current ryzen processors to be able to run as well as Intel.
No doubt that will be different with ryzen 3000 or, for that matter with whatever the intel response will be.
https://www.tomshardware.com/reviews/intel-core-i7-9700k-9th-gen-cpu,5876.html

 

[Gmoney06ss]

The real test is if your gaming/fps gets significantly better with 8 threads vs.4
Do not be impressed if your cpu% drops with 8 threads.
That is simply windows distributing the cpu activity across all available threads.
In fact, as strange as it might seem, it might indicate that you are only getting marginal benefit from the extra threads.

Here is a series of benchmarks for some games that explore this.
If your games were on the list of those tested that will be very interesting:

Gaming benchmarks: Core i7 6700k hyperthreading test (2018 update with GTX 1080 Ti)

Back in January 2016 i made a test about Intel Core i7 6700K virtual cores performance numbers in gaming. Short story back then was that hyperthreading offered no FPS improvement in any of my tested 21 games. Not only that, but about half of the tested games showed slight performance decrements...

www.techpowerup.com

I have a few on that list, two that I've tested so far and two I haven't. Doom and wildlands being the ones I did test. As far as doom goes, I didn't see any I'll effect on either fps or gameplay with the hyperthreading. Honestly game played exactly the same as before. 200fps and super smooth. Imo this is one of the best optimized games in a while. Wildlands saw no fps average gain, with minor min/low gains, gameplay was smoother. Unlike the test he did on his system, I don't feel like ht hurt at all.

Rise of tomb raider was on his list, I've not tested that but did test shadow. And found the results to be good overall. Gains all across the board in fps, and the benchmark now showing the game being more gpu bound.

I was planning an upgrade for the end of this year, but now feel I may wait. I mean I probably won't, cause it's like I just hate money, lol.

I'm patiently waiting benchmarks of the new ryzen vs 9th gen stuff to see which direction I take.

 

[boju]

For me doom was pretty smooth on 4cores too, Vulkan is an excellent api and is very tuned. It would be my ancient cpu but did have mini pauses every now and then running high %'s, though doom does achieve much higher fps than other games I've play so that's probably why i got pauses in doom and not in other intensive games like Ghost Recon / FC 5 with fps not as high. Id say 2600ks ipc obviously wasn't keeping up compared to later 4core gens albeit with HT those pauses stopped.