[SOLVED] Little to no performance increase after CPU upgrade

[stevewp8427]

I upgraded from an i3-4150 to an i5-4590 (Yes, I know my system is a bit outdated, I'm running a pre built cyberpower that I've upgraded over time) and ran a couple bench marks before and after the upgrade. For example I ran Heaven benchmark, userbenchmark.com, and a couple steam game benchmarks like hitman and tomb raider. The odd thing was there was little to no increase in performance FPS wise, I got almost the exact same min, max and avg. I don't know a whole lot about PCs so is there something I forgot to do? I made sure all drivers are up to date. I am also running a GTX 1050 Ti SSC with 8gb ram and a 600w psu.

Edit: Also, I did check system info and device manager, everything shows up the way it should.

 

[logainofhades]

That GPU is pretty weak, so that might be part of why you saw no real change.

 

[SamirD]

The reason you didn't see any difference is because the single thread performance between the 4150 and 4590 is nearly identical. If you want a real performance bump, you need to get a 4790k if your motherboard supports it:
https://www.cpubenchmark.net/compar...el-i5-4590-vs-Intel-i7-4790K/2252vs2234vs2275

 

[logainofhades]

A 4790k isn't going to help, with such a weak GPU either, though I agree an i7 would have been a better choice, for long term use. 4c/4t gaming is dead.

 

[SamirD]

A 4790k isn't going to help, with such a weak GPU either, though I agree an i7 would have been a better choice, for long term use. 4c/4t gaming is dead.

It believe it will as it's at least 10% faster. And I don't think the 1050ti SSC is weak depending on what you are making it do.

 

[logainofhades]

Considering even a Pentium G4560 can handle a 1050ti, without an issue, yes, that card is weak.

 

[Karadjgne]

It would help to clarify fps.
The cpu pre-renders a frame. It'll do that according to the game code, it's IPC and clock speeds. It takes a certain amount of time to place every object, quantify, give dimensions, AI, interactivity etc. The amount of times it can complete a frame in 1 second is the fps limit.

That info gets sent to the gpu, which then finish renders the frame into a picture, giving color, depth, shadows, 3d etc. It'll do that according to detail levels, post processing affects and resolution. That determines the fps you get onscreen, in the counter.

So if the cpu can do 100fps, and the gpu is set for ultra, it'll either succeed or fail. If it fails, lowering details, changing resolution or lowering post processing can raise onscreen fps, but is still blocked by the 100fps limit from the cpu.

You changed cpus in the same family. Nothing else. So the only thing done is raise the IPC slightly and clock speeds slightly. How well that translates is determined by the game and game code, which might be no gain at all to a few fps. But it won't be a drastic change on the limit. If detail settings etc are high enough that you are not reaching the fps limits, then you'll see no change basically as the cpus fps output isn't much different with such similar cpus.

The i7 can make a difference, but mostly in multi-player online games or single player games with mods. This'll raise the available thread count beyond 4, allowing for higher fps. You'll not see much difference with simpler, single thread heavy games as they are optimised for 4 threads or less.

Speeds don't usually translate in a 1:1 ratio. A 10% faster cpu may or may not average an extra 2-3 fps, and with as dynamic as game code is and the changes in view and details, 2-3fps will get lost in the translation. You'll only see such a minor change in a benchmark with a 0.1% low/max numbers in a side by side comparison.

 

[SamirD]

Considering even a Pentium G4560 can handle a 1050ti, without an issue, yes, that card is weak.

That doesn't make the card 'weak'. It is just sub-optimal.

 

[stevewp8427]

That GPU is pretty weak, so that might be part of why you saw no real change.

Yea i got it a couple years ago so i understand that, I just figured since the i5 had 2 more physical cores it would have helped a bit.

 

[stevewp8427]

It believe it will as it's at least 10% faster. And I don't think the 1050ti SSC is weak depending on what you are making it do.

I never really try maxing out games, im pretty casual so ill usually play games medium/high settings, nothing too extreme

 

[stevewp8427]

Considering even a Pentium G4560 can handle a 1050ti, without an issue, yes, that card is weak.

since the card is basically 4 years old now, what would be a good option as a replacement around the same original price? Maybe like $200ish or less

 

[stevewp8427]

It would help to clarify fps.
The cpu pre-renders a frame. It'll do that according to the game code, it's IPC and clock speeds. It takes a certain amount of time to place every object, quantify, give dimensions, AI, interactivity etc. The amount of times it can complete a frame in 1 second is the fps limit.

That info gets sent to the gpu, which then finish renders the frame into a picture, giving color, depth, shadows, 3d etc. It'll do that according to detail levels, post processing affects and resolution. That determines the fps you get onscreen, in the counter.

So if the cpu can do 100fps, and the gpu is set for ultra, it'll either succeed or fail. If it fails, lowering details, changing resolution or lowering post processing can raise onscreen fps, but is still blocked by the 100fps limit from the cpu.

You changed cpus in the same family. Nothing else. So the only thing done is raise the IPC slightly and clock speeds slightly. How well that translates is determined by the game and game code, which might be no gain at all to a few fps. But it won't be a drastic change on the limit. If detail settings etc are high enough that you are not reaching the fps limits, then you'll see no change basically as the cpus fps output isn't much different with such similar cpus.

The i7 can make a difference, but mostly in multi-player online games or single player games with mods. This'll raise the available thread count beyond 4, allowing for higher fps. You'll not see much difference with simpler, single thread heavy games as they are optimised for 4 threads or less.

Speeds don't usually translate in a 1:1 ratio. A 10% faster cpu may or may not average an extra 2-3 fps, and with as dynamic as game code is and the changes in view and details, 2-3fps will get lost in the translation. You'll only see such a minor change in a benchmark with a 0.1% low/max numbers in a side by side comparison.

thanks for the in dept reply, so i realize the upgrade isn't as substantial as i expected it to be, that is okay though i will eventually have to upgrade the gpu anyway. Any recommendations for a decent card that's a good match for the i5 $200ish?

 

[SamirD]

I never really try maxing out games, im pretty casual so ill usually play games medium/high settings, nothing too extreme

And that's why I think you'd see an improvement. The 4790k is a significantly faster processor than anything else lga1150.