I5 4670k vs fx-8150

so i have an fx 8150 oc to about 4.6ghz(just running on liquid) paired with a 7990.

i also just built a intel i5 4670k (haven't oced yet/waiting on cooler) paired with a sapphire 7970.

here's my dillema, i5 is supposed to run circles around the fx-8150 according to all the benchmarks/reviews i've seen. but its not even coming close in the bench marks. it scored below a "gaming machine" (3dmark rates that as being a i7-3670k paired with a titan) where my fx is beating the score by a good enough margin that i'm pretty dam happy. is it just the oc that is allowing this? is it the graphics card? can i disable half my 7990 to get a more fair comparison? would a ssd vs a hdd make that much of a difference?

i'm honestly trying to bench this to get a true comparison.

i will post score of the 2 tomorrow afternoon unless someone comes up with a solution.

Not overclocked and a lesser GPU? Seems about right. For a true test put both CPUs at stock and use the same video card on both.

yeah but isn't the point of getting a fx processor to overclock it? and besides you could end up spending a little more combined on cooling and processor and out perform the intel. i just think its very skewed the way most people are looking at the processors performance ratings.

All depends what are you testing but when I had FX8320 then Intel rig was still much better in about everything what I was doing.
You won't see that in most games as they're using mainly GPU power , not CPU and if you have 120 or 100 FPS then it doesn't really matter. Recently I've switched i5 4670K to i3 4330 and I feel no difference. There is probably less FPS but I just can't see it.
SSD makes big difference in some games but not really in benchmarks.
I got rid of FX platform mainly because of much more heat. Cheaper was to sell it and move to Intel than change all my cooling in daily PC ... and I already had Intel anyway.

What does the SSD do for games because games loads from HDD or SSD into system memory first.

If games used SSD or HDD for the graphics you would not need system memory.

maesterdraconas said:

yeah but isn't the point of getting a fx processor to overclock it? and besides you could end up spending a little more combined on cooling and processor and out perform the intel. i just think its very skewed the way most people are looking at the processors performance ratings.

Click to expand...

Just like the point of the unlocked multiplier on the intel chip? Your comparison is very skewed, the GPU in your FX machine has potentially twice the horsepower (scaling would actually put it at ~80% or so on average, depending on game) and the CPU overclock certainly won't help.

I'm not sure where you get running circles around the FX chip from either - that will only apply in single threaded tests, multithreaded will be much closer with the FX sometimes able to pull ahead with no HT on the intel (both at stock). In GPU benchmarks the difference can be smaller if it is the GPU bottlenecking and not the CPU, and in most games both can offer over 60FPS. The overclock you have probably brings your two CPUs to be roughly equal in gaming performance (if you had identical hardware in each).

Put the intel at the same clock and then you will see it "run circles round the FX" its a shame but AMD just can't match intel no more, no matter how much I wish they could.

even at 4.2Ghz the Intel chip will smoke the AMD chip in games id actually bet stock the Intel chip will give you more FPS with the same GPU installed.

Just the efficiency alone from the Intel architecture will yield you better fps.

that said with the AMD chip oced as much as you have it you should be close to no bottleneck's.

Now in benchmarks this may be different since alot of benchmarks award points for more cores. for instance a 2600k will score thousands points higher then a 2500k at the same clock speeds in 3dmark 11 even though the fps is similar/same.

Basically every game available does not reward fps for multicore performance past 2 cores. so single core performance is key for current gaming.

However with mantel and the coming DX12 things may change for multi core processor performance contribution.

If i was you id get a half decent cooler for the 4670k get at least a 4.3-4.4ghz overclock on it and then run it with the 7990 and i would guess your going to see a sizable fps gain esp for your minimum fps levels.

You have to remember the Intel chip clock for clock is a-lot higher performance and more efficient. the AMD 8350 is closer to sandy bridge performance and still even slightly slower clock for clock then sandy bridge. the haswell chip is about 30% on average faster then sandy bridge.

If however Multiple cores become a more important factor having the 8 cores vs the 4 cores on the Intel cpu should really close that gap up alot.

but right now id use the oc'ed Intel cpu like i said for your main rig.

Well, here is a Sb vs IB vs Haswell per clock comparison.

http://www.pcper.com/reviews/Proces...erformance-and-Architecture/Clock-Clock-Sandy

We're more on 10/15% from SB to Haswell.

Agreed on everything else though SF101

OP also has a fx 8150 not a 8350 which was a fair bit faster and oc's better im actually impressed he got 4.6ghz out of his 8150 the one i oc'ed for my boss was throttling @ 4.3ghz at 65 or so *C and it was a real pain.

maesterdraconas said:

so i have an fx 8150 oc to about 4.6ghz(just running on liquid) paired with a 7990.

i also just built a intel i5 4670k (haven't oced yet/waiting on cooler) paired with a sapphire 7970.

here's my dillema, i5 is supposed to run circles around the fx-8150 according to all the benchmarks/reviews i've seen. but its not even coming close in the bench marks. it scored below a "gaming machine" (3dmark rates that as being a i7-3670k paired with a titan) where my fx is beating the score by a good enough margin that i'm pretty dam happy. is it just the oc that is allowing this? is it the graphics card? can i disable half my 7990 to get a more fair comparison? would a ssd vs a hdd make that much of a difference?

i'm honestly trying to bench this to get a true comparison.

i will post score of the 2 tomorrow afternoon unless someone comes up with a solution.

Click to expand...

So... which 3DMark did you use to test? Some 3DMark tests, such as the latest one, 11, and Vantage, take into account your CPU score. So if you are testing the GPU, those benchmarks are not an apples to apples comparison as the number of cores matter SIGNIFICANTLY in those 3DMarks I mentioned. Notice the CPU Physics score...since the 8150 has '8' cores and the 4670K 4 cores, that is the reason why you are seeing a significant difference in that type of testing. If you want to isolate the GPU, run a test like Uningine Heaven or Unigine Valley and compare the scores.

Though these things were mentioned before, they deserve another mention. And SSD vs HDD in gaming, as far as FPS does not matter in the least. Now, to LOAD the game, certainly, but they have nothing to do with FPS.

Your testing methodology is inherently flawed for the data you are trying to get.

well then i'll just have to wait for the cooler to come in, then test again with the same gfx card in both machines. i really do appreciate all the feed back, as well as i'm seeing from most of the comments that the intel is not directly much faster per say, just on a single core basis. i'll let you know what i find out in the next few days.

as for the 3dmark, its the one from steam v1.2.362 first strike/cloud gate/ice storm i know there are many versions, but this isn't listing it really anywhere

So............... which 3DMark did you use?

As far as on a single core basis... not really. the 4670K is a quad core, no HT, the 8150 is an '8' core CPU. So up to 4 cores, performance should be the same (assuming zero use of the rest of the 8150's cores). AFTER 4 cores the 8150 will pull ahead.

4670k+7990 3dmark firestrike results:
http://www.3dmark.com/search#/?mode...ntel Core i5-4670K&gpuName=AMD Radeon HD 7990

8350FX+7990 3dmark firestrike results:
http://www.3dmark.com/search#/?mode...puName=AMD FX-8350&gpuName=AMD Radeon HD 7990

If your 4670k scores less than your 8350, there is something really wrong in your Intel setup...

Edit: and 8150+7990 firestrike results:
http://www.3dmark.com/search#/?mode...puName=AMD FX-8150&gpuName=AMD Radeon HD 7990

The reason it scores better is because of better CFx scaling from the CPU pushing them. Single card, things would be a bit different I would imagine.

1 GPU, 8150:
http://www.3dmark.com/search#/?mode...puName=AMD FX-8150&gpuName=AMD Radeon HD 7970

1GPU 4670k:
http://www.3dmark.com/search#/?mode...ntel Core i5-4670K&gpuName=AMD Radeon HD 7970

Even bigger gap...

Dude, I am not searching through all those results for matching clock speeds on both items...LOL! That said, in thinking, and seeing, perhaps FS is not the best example of this, LOL! 3d11 shows some big differences. I forgot, though there is a physics tests, it doesn't matter nearly as much in the overall score as 11 and vantage.

It would sure help if the OP answered questions asked of him (which bench!!) and posts the scores instead of trudging forward rather aimlessly...

Yes, you are! Lol!

I will run a couple of FS later on tonight with my 3770k@stock speed and 2 280x's to get accurate results.

Don't waste your time as I do not think that will show a thing... let's get the OP to clarify and point him in the right direction.

See this is what happens when you put benchmarks as your compass of performance.

Those 3D mark benches reward points for more cores even when the fps difference isn't their.

The physics scores vastly improve.

A bench like heaven was a much better benchmark to judge gaming performance from as it gave a more real world performance view and didn't just hand out free points for having more cpu cores.

A good example of that is like i said before 2500k vs 2600k @ same clock speeds. in 3D mark 11 you will get way more points with a 2600k go play a game and you will get the exact same fps because they are basically the same thing except one has hyper threading enabled.

So when you have a 8 core AMD processor that's oc'ed to 4.6 vs a 4 core processor running stock clocks @ 3.4-3.8 GHz it ends up being a skewed result that does not really reflect real world performance.