Does fast gpu make up for a slow cpu.. ??

[trog100]

i am bit bored so i am fiddling..

people say there is no point in having a fast gpu with a slow cpu.. without making any other changes i have just dropped my multiplyer from 45 to 30..

it does knock by 1440 firestrike extreme score down to 12000 odd from 14000 odd.. still pretty quick though..

windows seems the same.. i cant detect any difference and Just Cause 3 plays the same i cant detect any difference..

my system is in my specs.. i might go down to a multiplyer of 22 a tad less than half of what i normally run at..

i strongly suspect i still wont be able to detect any difference without running a benchmark..

what am i trying to prove.. buggered if i know.. blame it on too much time on my hands.. he he

trog

ps.. maybe an I7 at 3 gig aint slow.. to me it is though..

my vcore shows .918 at 3 gig.. i bet it would go lower..

my cpu tempts running wprime are 35 C..

 

[GoldenX]

Well, a 2,2GHz Haswell is still a very fast processor, try running it at 2,2 but with just 2 or 1 core and HT disabled, so it seems similar to a cheap or old CPU.

 

[trog100]

i recon i would notice a difference with only two cores running at 2.2.. from what i read Just Cause 3 is said to need a fast cpu..

it does tend to suggest that the gpu is king.. given plenty of gpu power the cpu takes a major second place..

but i can try going lower and switching some of the cores off..

trog

 

[BarbaricSoul]

Your 4790k at 3 GHz is about equal to the 3.3 GHz 3770 listed in the recommended system specs for Just Cause 3 (Haswell is about 10% faster than Ivy Bridge). You won't notice a difference as long as you stay within the recommended performance gap set by the recommended system specs for the game. Cut your 4790k down to a non-HT dual core, or down clock it to the 2.5 GHz range and see what happens. Even tri-SLI GTX 980ti won't help if you don't have enough CPU power.

 

[EarthDog]

1. FS has a CPU test in it... bad test as inevitably the score will lower because you are dropping cores and core speed which the tests rely on.
2. Techrspot? Tests games and different clockspeeds of Intel and AMD processors.... go look there for your results....(this is tested on each game they review).
Example: http://www.techspot.com/review/734-battlefield-4-benchmarks/page6.html
3. Stop trying to 'detect' slow FPS and actually measure it... you know, scientific method instead of this seat of your pants stuff.
4. What does Wprime have to do with this? You talking games or what?

In short, it will vary by game you test as to how many, if any, FPS you drop...however
this isn't a new thought and has been tested before.

 

[SnakeDoctor]

Pop in a G3250 or an i3 and compare the difference or will need to run something at 4k that cause not great fps in the first place then downclock after
Not too much point if already maxing out a game with stupid amounts of fps in the first place or compare without the sli maybe, seeing as your cpu and gpu's are beefy

 

[Bill_Bright]

Does fast gpu make up for a slow cpu.. ??

"Make up for" is a pretty generalized statement. A slow CPU can bottleneck just about everything. But with graphics intensive tasks, much of the CPU's job is to hand off tasks to the GPU and it really does not take a lot of CPU horsepower to hand off tasks. And from there, a faster GPU can speed up graphics processing. But some games and many programs are very CPU intensive so a faster GPU may toss the results up on the screen faster (once it gets them from the CPU) but probably not something you would notice.

So people who say there's no point in having a faster GPU with a slow CPU are really being narrow minded. It really depends on the tasks being performed. And of course, other factors factor in too. For example, if you have a small amount of RAM, that can bottleneck your system too - again, depending on the tasks being performed. And a slow hard drive can cause bottlenecks with disk intensive tasks. But you have 32GB and SSDs so no problems there.

 

[Flow]

Most games wont loose much fps if you drop the multi of a modern i7 to defaults.
As for a faster gpu making up for a slower cpu, I pressume you're talking games that need a fast cpu to start with?
A fast gpu wont make up for the slow cpu if a game demands it. You can play at higher settings though, without loosing or gaining fps.
But as nowadays games rely more on the gpu, and most cpu's provide enough speed, you wont encounter that problem soon.
I do know from the past, it was for fps games and multiplayer shooters better to have a fast cpu first, and play with lower graphics until one could upgrade the card.

At present I run into no problems playing games like witcher 3 and fallout 4 with my aging cpu.

 

[trog100]

1. FS has a CPU test in it... bad test as inevitably the score will lower because you are dropping cores and core speed which the tests rely on.
2. Techrspot? Tests games and different clockspeeds of Intel and AMD processors.... go look there for your results....(this is tested on each game they review).
Example: http://www.techspot.com/review/734-battlefield-4-benchmarks/page6.html
3. Stop trying to 'detect' slow FPS and actually measure it... you know, scientific method instead of this seat of your pants stuff.
4. What does Wprime have to do with this? You talking games or what?

In short, it will vary by game you test as to how many, if any, FPS you drop...however
this isn't a new thought and has been tested before.

dude as i said i am bored.. i have already read tons sh-t.. most of it bollocks.. i am simply amusing myself.. read the thread or dont but dont tell me what i can and cant do please.. he he

i mention just cause cos it happen to be the game i am currently playing..

i am now going to move down to 2 gig 4 cores with hyper threading .. just to see.. and cos i can..

trog

ps.. 20 x multiplyer 2 gig.. vcore reads .801

cpu temps running wprime 32 C..

now to see if it will play 1440 just cause 3 okay..

 

[Frick]

I have some suggestions if you want some good reading. Start here.

Otherwise you can do some benches and make some nice graphs, that'll take some time too.

 

[chuck216]

A slow CPU can definitely help hamper your GPU. Taking Firestrike as an example I score 6681 with my cpu at stock speeds (3.5 Ghz) and with a 4.5 ghz OC it scores 7539. A significant difference considering I only have a $200 GPU and a $130 CPU.

 

[trog100]

i dont dispute the fact that slow a cpu will hamper a gpu.. that goes without saying..

what i am more interesting in find out for myself is just how much a fast gpu makes up for the slow cpu..

so far i am down to 2 gig.. pretty slow i think.. it certainly does hamper my pair of 980 TI cards no question of that..

but Just Cause 3 was still perfectly playable with the cpu at 2 gig.. 1440 frame rates mostly around 80 with highs around 100 and lows around 50..

which does kind of suggest that if you chuck enough gpu power at any game even a half assed cpu will do a good enough job..

it also kind of suggests (and not everybody knows this) that anybody with an older generation I5 or I 7 cpu aint gonna suffer that much because of this.. without a doubt the gpu is king..

trog

 

[Bill_Bright]

what i am more interesting in find out for myself is just how much a fast gpu makes up for the slow cpu..

And again, it depends on the tasks you throw at them. I certainly would not base your decision on Just Cause 3 alone.

 

[Frick]

i dont dispute the fact that slow a cpu will hamper a gpu.. that goes without saying..

what i am more interesting in find out for myself is just how much a fast gpu makes up for the slow cpu..

so far i am down to 2 gig.. pretty slow i think.. it certainly does hamper my pair of 980 TI cards no question of that..

but Just Cause 3 was still perfectly playable with the cpu at 2 gig.. 1440 frame rates mostly around 80 with highs around 100 and lows around 50..

which does kind of suggest that if you chuck enough gpu power at any game even a half assed cpu will do a good enough job..

it also kind of suggests (and not everybody knows this) that anybody with an older generation I5 or I 7 cpu aint gonna suffer that much because of this.. without a doubt the gpu is king..

trog

Beeeenncchheeeees and graphs.

 

[chuck216]

Well for a GPU vs GPU comparison we have this: Exact same system only difference swapped GPUs. In fact the R9 380 was a free RMA replacement for the 270X after one of the fans failed on the 270X.

 

[trog100]

witcher had has a mention.. just to see i had quick run around in it.. cpu at 2 gig.. witcher fully maxed out at 1440.. frames rates around 70 fps..

witcher is very gpu dependant but not very cpu dependant.. the 1440 fps wasnt a deal different than with the cpu at 4.5 gig..

i am still at 2 gigs just to see how windows in general runs on a much slower cpu..

trog

 

[chuck216]

The main problem with most reviews comparing high end graphics cards... and even mid range ones in benchmarks and even real world tests is that they compare them using the most powerful CPU they have at the time. Nobody testing the newest GTX 980ti is going to run it on a motherboard with a Pentium III or Athlon 64.

Also CPU performance depends on a number of factors, even slowing the processor down you still have to consider number of cores, hyperthreading, and if the particular piece of software being tested uses single or multiple threads as 8 threads at 1 ghz is still faster than 1 thread at 5 gz

 

[trog100]

The main problem with most reviews comparing high end graphics cards... and even mid range ones in benchmarks and even real world tests is that they compare them using the most powerful CPU they have at the time. Nobody testing the newest GTX 980ti is going to run it on a motherboard with a Pentium III or Athlon 64.

Also CPU performance depends on a number of factors, even slowing the processor down you still have to consider number of cores, hyperthreading, and if the particular piece of software being tested uses single or multiple threads as 8 threads at 1 ghz is still faster than 1 thread at 5 gz

mostly it isnt and more so with games.. nothing much except cpu benchmarks and the odd specialist video software will use them..

the real problem with most reviews is they are all trying to exaggerate the differences.. they create an artificial scenario which does just that.. exaggerate the differences..

i just had another play with Just Cause 3 with my normal 70-ish frame rate cap.. with a g-sync monitor 70 fps seems to produce very good very smooth results without my PC turning into a room heater.. i just played the game in my normal way and liberated a town.. this time not looking at a frame rate counter.. the game played exactly the same with the cpu at 2 gig as it does with it at 4.5 gig.. nice and smooth without any problems..

we aint talking small differences here in cpu power.. we are talking f-cking huge ones..

i have five different profiles set up in my bios.. 4.2.. 4.4.. 4.5.. 4.6.. and 4.7.. al finely tuned with minimum stable vcore voltages.. this is the first time i have tried silly low clock speed figures.. i must admit that i am a tad surprised that a cpu intensive game like Just Cause 3 still plays okay with my cpu as low as 2 gig..

i did say i was just bored and filling in time..fiddling about.. no fucking way am i gonna produce charts.. the would simply earn me much flack.. cos they would not be saying what most folks want to here..

you dont often get to hear what a pair of 980 ti cards behave like with a slower cpu.. all i can say is pretty good.. way better than most folks would think.. he he

trog

 

[yogurt_21]

Fast car on slow traffic ridden road = slow going
slow car on fast open road = slow acceleration but likely faster than above.

see even your 3 cylinder Geo Metro is eventually going to get up to speed on that open road. Whereas even a world record setting rocket car is going to be stuck in traffic on the other road.

The question as Bill Bright elaborated on is what are we doing. In most games the cpu is the car and the gpu is the road. So a slow cpu + fast gpu will still cut it once the settings are right. But certain simulators require more from the cpu than the gpu in which case the cpu is the road and the gpu is the car. So in that scenario the slow cpu is the problem since its handling most of the load.

bottlenecks are a just as much a battle of "what am I doing" as they are "what am I using to do this task."

 

[chuck216]

Here's an idea... try the very same game Just Cause with g-sync/vsync turned off so you're not artificially limiting the frame rate your dual 980Ti's are capable of. Then you might see a difference in framerate, with 2 $1000s GPU any system had better be able to get 70 FPS no matter what cpu they are using. It's how high (or low) the framerate can go without being limited by artificial means that matters.

 

[JunkBear]

NO. It's like if your eyes are super fast and they can see lot images at same time. But on the other side your brain can't compute all this visual information.

 

[xorbe]

JC3 seems pretty CPU intensive, in my testing. My other A10-6700 system just doesn't crank out the frames even when I use a GTX 960 at low resolution / low quality settings. ie, I can set the 960 to low or high settings, and the fps barely changes. CPU bottleneck.

 

[trog100]

i did run an unlimited test and when running a benchmark did see a difference.. just not a big enough one to to make any of the games i have unplayable..

the point is if 70 fps is enough and i would say it is.. my pair of 980 ti cards will still play any game i have chucked at them even with my cpu clocked right down to 2 gig.. good enough not to see a difference unless a frame rate counter is active.. i have tried three games so far.. all at 1440.. mad max.. just cause 3 and witcher.. they all played fluidly with no problems..

i have also run firestike extreme with my cpu at 2 gig.. i score just over 10000 pretty much the same as what they have listed as 4k gaming machine.. still pretty a good performance even if limited with a slowed down cpu..

firestike extreme.. cpu at 4.5 gig just over 14000.. cpu at 3 gig the score is 12000.. cpu at 2 gig the score is 10000.. 10000 is what pair of 980 cards with a good cpu would score..

running benchmarks will show a difference i have never said or implied anything else.. playing a game for real so far dosnt show any noticeable difference.. pretty much like normal windows use dosnt.. i am still on the 2 gig setting giving it a fair test..

it certainly suggests any amd cpu (slower then intel or not) wont have the slight problem playing any of todays games..

it does tend to suggest that overclocking a higher end cpu wont show any gaming gains ether.. not at real life resolutions.. plus given decent gpu power any benchmark gains wont translate into real life gains..



trog

xorbe.. JC3 is both cpu and gpu intensive.. not easy on the hardware thats for sure..

but i am starting a new day running a nice under volted and cool 4.4 gig with a 1.2 vcore safe in the knowledge i aint missing much.. my little underclocking experiment did serve some purpose..

trog

 

[CAPSLOCKSTUCK]

I think the most pertinent quote is

good enough not to see a difference unless a frame rate counter is active

 

[Bill_Bright]

The main problem with most reviews comparing high end graphics cards... and even mid range ones in benchmarks and even real world tests is that they compare them using the most powerful CPU they have at the time.

That's not a problem - that's a good thing! They should use the most powerful/fastest CPU, motherboard, RAM and drives possible.

By using the most powerful CPU available, that removes the CPU from the equation (test results) by ensuring the CPU is not creating a bottleneck and skewing the results.

As I noted above, for many graphics tasks, the results are dependent on how fast the CPU can hand off those tasks to the GPU. If the GPUs are sitting there in wait states until a slow CPU can hand off the next task, the full potentials of the GPUs will never be realized.

You can't determine how much water a 6" drain pipe can drain in one minute if you are feeding water to it from a 1/2" garden hose.

So, if the comparative review is to be considered unbiased and legitimate, they must use the most powerful CPU, fastest and gobs of RAM, a motherboard with the fastest bus speeds, and drives with the fastest access and reads times to properly, fairly, and accurately test and compare any graphics card, from high-end to entry level.