i5-760 to i5-4690K, worth the upgrade?

[jihe]

I sidegraded from a 4GHz i5-750 to a pentium G3258, the most noticable difference is in the fast boot stuff.

That's a downgrade

 

[jihe]

Alright so an update on my i5-4690K. I got it stable at 4.7Ghz @ 1.265V. On 4.8Ghz I can make it POST but fails on stress testing within 15s. It's being cooled by a Noctua NH-U14s.

I mounted it vertically (wind going up) and average core temps were 72.75C on Intel Extreme Tuning Utility and 95.25C on Intel Burn Test.

I then changed the mounting horizontally (wind going to the back) and average core temps were 69C on Intel Extreme Tuning Utility and 87.25C on Intel Burn Test.

This surprised me a lot as I have 2x 120mm fans on the top exhaust, while only having 1x 120mm fan on the back exhaust.

My theory is that the convex shape of the heatsink base really helps align with the die under the IHS of the i5-4690K.

But yes, I'm very happy with my i5-4690K @ 4.7Ghz. One thing I immediately noted was just how fast this PC boots up, it's almost 2x as fast as my previous setup, literally boots to Windows in 6-7s. Whether its the SATA2 -> SATA3 change or the processor or combination of both, I'm not sure. But I'm sure enjoying this!

Alright.. It's been a long day. I haven't even gamed on this beast yet, time to do that! See you guys later! I'll report back any improvements in games.

Wouldn't wind blowing up carry the hot air around graphics card to your cpu?

 

[Yuriman]

Wouldn't wind blowing up carry the hot air around graphics card to your cpu?

If there's enough airflow (and it doesn't take much) the air coming from near the graphics card shouldn't be much above any other part of the case.

 

[Ryanrenesis]

Wouldn't wind blowing up carry the hot air around graphics card to your cpu?

Yes, theoretically it will. But during the stress tests, my GPU temps were a cool 33-35C.

 

[Deders]

My theory is that the convex shape of the heatsink base really helps align with the die under the IHS of the i5-4690K.

But yes, I'm very happy with my i5-4690K @ 4.7Ghz. One thing I immediately noted was just how fast this PC boots up, it's almost 2x as fast as my previous setup, literally boots to Windows in 6-7s. Whether its the SATA2 -> SATA3 change or the processor or combination of both, I'm not sure. But I'm sure enjoying this!

Ideally both surfaces should be flat

Anything before the windows logo is down to your motherboard. I wouldn't expect Sata 3 to make too much of a difference but If you re-installed windows fresh then that would be noticeably quicker.

 

[Termie]

Alright so an update on my i5-4690K. I got it stable at 4.7Ghz @ 1.265V. On 4.8Ghz I can make it POST but fails on stress testing within 15s. It's being cooled by a Noctua NH-U14s.

I mounted it vertically (wind going up) and average core temps were 72.75C on Intel Extreme Tuning Utility and 95.25C on Intel Burn Test.

I then changed the mounting horizontally (wind going to the back) and average core temps were 69C on Intel Extreme Tuning Utility and 87.25C on Intel Burn Test.

This surprised me a lot as I have 2x 120mm fans on the top exhaust, while only having 1x 120mm fan on the back exhaust.

My theory is that the convex shape of the heatsink base really helps align with the die under the IHS of the i5-4690K.

...

Alright.. It's been a long day. I haven't even gamed on this beast yet, time to do that! See you guys later! I'll report back any improvements in games.

The cooler works better horizontally because it's getting cool air from the front intakes, rather than hot air from the video card. You did have an interesting theory, though.

Hope you have some benchmarks from your old system to test against this system - a lot of people are interested in the upgrade you just did and would love to see the difference.

 

[Ryanrenesis]

The cooler works better horizontally because it's getting cool air from the front intakes, rather than hot air from the video card. You did have an interesting theory, though.

Hope you have some benchmarks from your old system to test against this system - a lot of people are interested in the upgrade you just did and would love to see the difference.

Okay I tested one CPU-intensive game so far, SWTOR. And although unfortunately I don't have any proper benchmarking baseline to go from the other PC, I can say one thing for sure: minimum FPS has increased drastically.

I have NOT encountered any crazy frame dips to 15 FPS like I did with my i5-760 @ 4Ghz. The lowest FPS recorded is 37 FPS so far.

Another thing I noticed: GPU usage actually went UP! With my i5-760, my GPU wouldn't even overclock itself to max boost, just staying at 1019Mhz. Now my GPU actually boosts to 1202Mhz much more often.

Very impressed with this upgrade overall, going to try Shadow of Mordor next.

EDIT: I have 3DMark benchmark scores from previous CPU if anybody is interested I could benchmark 3Dmark on this CPU and see any difference. Would anybody be interested in seeing these? It is only synthetic tho.

 

[Termie]

3DMark won't really tell you anything. The Physics Score will be much higher, the Graphics score will be practically the same, and the overall score will be much higher due to the Physics Score.

Not a good test for this purpose. It will be far out of line with what you'd see in games.

 

[Flapdrol1337]

That's a downgrade

Nah, it's faster the games I play most.

 

[Ryanrenesis]

Okay so I played Shadow of Mordor intensively after this upgrade and WOW!

No more stutters... No more frame dips... It's 100% fully smooth.

So I guess my i5-760 @ 4Ghz was hindering my GPU after all.

Man, I'm super satisfied with my new i5-4690K, I recommend anyone on this upgrade path to pursue it.

 

[Deders]

What motherboard did you have? I didn't get frame dips at all in mordor, it was very smooth.

 

[Ryanrenesis]

What motherboard did you have? I didn't get frame dips at all in mordor, it was very smooth.

As stated in my OP, my previous mobo was the MSI P55-CD53.

I would stutter in certain areas, stutter when entering Wraith state, stutter when executing people, though most of the time it was fairly smooth.

Since this upgrade, all of that is gone, and average FPS has gone up, I'm now hovering 80-120FPS instead of 50-100FPS with occasional dips to 15 FPS.

I can't imagine how a 4th gen i5 with 700Mhz lead in frequency can make such a difference over a 1st gen i5. But, I'm glad this upgrade fixed all my frame issues.

EDIT: There IS a monumental difference in Gflops though--1st gen i5 at 4Ghz was giving me 56 GFlops on Intel Burn Test, while a 4th gen i5 @ 4.7Ghz is giving me 121 Gflops on Intel Burn Test. That's a crazy difference.

 

[Deders]

It's very strange that with such similar hardware we have such a different experience

 

[Ryanrenesis]

It's very strange that with such similar hardware we have such a different experience

And seeing you only have 8Gb RAM and I had 16Gb RAM, I should have a smoother experience eh?

Strange indeed.

EDIT: Did you install Shadow of Mordor on an SSD? Mine is installed on a typical 3TB harddrive.

 

[beginner99]

As stated in my OP, my previous mobo was the MSI P55-CD53.

I would stutter in certain areas, stutter when entering Wraith state, stutter when executing people, though most of the time it was fairly smooth.

Since this upgrade, all of that is gone, and average FPS has gone up, I'm now hovering 80-120FPS instead of 50-100FPS with occasional dips to 15 FPS.

I can't imagine how a 4th gen i5 with 700Mhz lead in frequency can make such a difference over a 1st gen i5. But, I'm glad this upgrade fixed all my frame issues.

EDIT: There IS a monumental difference in Gflops though--1st gen i5 at 4Ghz was giving me 56 GFlops on Intel Burn Test, while a 4th gen i5 @ 4.7Ghz is giving me 121 Gflops on Intel Burn Test. That's a crazy difference.

This difference is due to new instructions, namley AVX. Note that the biggets jump was from Lynnfield (eg. i5-760) to Sandybridge. And then add some more. All in all haswell is easily 40% faster at same clock.

And seeing you only have 8Gb RAM and I had 16Gb RAM, I should have a smoother experience eh?

Strange indeed.

EDIT: Did you install Shadow of Mordor on an SSD? Mine is installed on a typical 3TB harddrive.

RAM amount shouldn't really matter for in game experience. 8 is enough and 16 is overkill for gaming.

Some people are more susceptible to stutter than other could also be the reason.

 

[Ryanrenesis]

This difference is due to new instructions, namley AVX. Note that the biggets jump was from Lynnfield (eg. i5-760) to Sandybridge. And then add some more. All in all haswell is easily 40% faster at same clock.



RAM amount shouldn't really matter for in game experience. 8 is enough and 16 is overkill for gaming.

Some people are more susceptible to stutter than other could also be the reason.

I used to think that too... Until I saw how much RAM Shadow of Mordor uses: all 3078MB of my GPU's RAM + 10456MB of my RAM, as recorded on MSI Afterburner.

If I only had 8Gb RAM, I wouldn't be able to play this game on Ultra textures.

In this case, luckily I went overkill for my RAM

 

[Techhog]

Wait for Skylake or Cannonlake. Maybe Broadwell-K if you really just can't wait that long.

 

[Seba]

Wait for Skylake or Cannonlake. Maybe Broadwell-K if you really just can't wait that long.

He already bought the new components.

 

[Ryanrenesis]

Wait for Skylake or Cannonlake. Maybe Broadwell-K if you really just can't wait that long.

I'm interested in knowing why you suggest waiting? I notice you also have a Haswell processor.

But yes, waiting till Summer 2015 is too long for me.

 

[Arachnotronic]

I'm interested in knowing why you suggest waiting? I notice you also have a Haswell processor.

But yes, waiting till Summer 2015 is too long for me.

Broadwell-K is probably not going to be worth the wait. Minimal increase in ST perf/clock and it's not clear if BDW will clock any better than HSW. Also, they're sticking GT3e GPU on there, so it may be more expensive than the current models.

You made the right choice, OP.

 

[sm625]

It probably stuttered because you had so much RAM. Shrug. Just because it used 10GB doesnt mean that it needed to. If you only had 8 it would probably only use 5 and be perfectly happy with it. At any rate there is clearly something wrong with your old setup. There is no reason anyone should expect more than 20% minimum fps increase going from a 4GHz nehalem to a 4.5 GHz haswell.

 

[Ryanrenesis]

It probably stuttered because you had so much RAM. Shrug. Just because it used 10GB doesnt mean that it needed to. If you only had 8 it would probably only use 5 and be perfectly happy with it. At any rate there is clearly something wrong with your old setup. There is no reason anyone should expect more than 20% minimum fps increase going from a 4GHz nehalem to a 4.5 GHz haswell.

I'm wondering three things:

1. Where did you get the idea or info that having more RAM will cause stutters and significant FPS drops in games? Because I've never come across this piece of information or idea before. A source would be helpful.
2. How does a game go from using 10Gb RAM to only 5Gb RAM without any reduction in performance or graphics?
3. Have you actually travelled the same upgrade route or merely speculating about the Nehalem to Haswell performance increase?

I'm not diminishing your claims by any means but it's quite surprising especially with the RAM claims.

EDIT: In regards to the 10 -> 5Gb RAM claim, I know Shadow of Mordor uses memory in this order: GPU RAM -> System RAM -> Windows Pagefile. As each tier gets full, additional memory goes to a lower tier. So I can understand that you could probably run Shadow of Mordor on only 8Gb RAM using Ultra textures, but something needs to give: either it uses Windows Pagefile, where performance will start getting severely hindered, OR it automatically uses lower textures as RAM gets full.

Either way, I don't see how a game can go from using 10Gb to 5Gb without any reduction in performance or graphics.

 

[kimmel]

I'm wondering two things:

1. Where did you get the idea or info that having more RAM will cause stutters and significant FPS drops in games? Because I've never come across this piece of information or idea before.
2. Have you actually travelled the same upgrade route or merely speculating about the Nehalem to Haswell performance increase?

I'm not diminishing your claims by any means but it's quite surprising especially with the RAM claim.

There are also people on the internet that claim 30fps is better than 60fps.

 

[Lepton87]

There are also people on the internet that claim 30fps is better than 60fps.

Some say 24fps is even better due to more pronounced cinematic feel.

 

[Termie]

OK, since I've actually done this exact upgrade and have also done lots of testing at 8GB and 16GB, I'll chime in here:

(1) Without a doubt, the OP got much more than a 20% increase in minimums, probably around 40%. The averages probably only increased about 20%. For any game that averaged under 60fps, this will have a dramatic impact, as the minimums likely had been noticeably low.

(2) RAM has absolutely nothing to do with the increase. 16GB provides zero benefit over 8GB. It's just caching. I added another 8GB a few weeks ago to one of my systems and everything uses more memory, but performance hasn't changed at all. Pagefile usage is lower, however, but this doesn't affect frames per second.