What should I upgrade GPU or CPU

ravs316

Distinguished
Apr 14, 2011
29
0
18,530
Hi all. I wan't to upgrade my PC to improve the gaming performance what I have now is:

e8400 3.4 ghz
MSI P35 NEO F
EVGA GTX 275 FTW
8GB KINGSTON RAM
1 TB SEAGATE BARRACUDA 7200.11

Considering this:

■ I play on 1920x1080 resolution
■ I play all kind of games (from: windows solitaire (LOL) to crysis 2)
■ Lately I played Dragon Age 2 I felt a minimum lag but not big deal and also crysis 2 but i know that this game is CPU intensive

The budget that I have is 500$ so I was thinking in 2 options

■ i5 2500K + (Asus P8P67 B3 LE or MSI P67A-GD65 B3) + DDR3 4GB (2x2GB) 1600MHz Doninator, CORSAIR ~ 500$ (in my country)
OR
■ GTX 560 ti ~ 250$ (in my country) and save the other 250$ for future

I would like your opinion about what should I upgrade first? GPU or CPU (CPU + MD + RAM)

Thanks in advance,
RAVS

 
The lag in Dragon Age is likely due to the GPU. The lag in Crysis is definitely the GPU.

I would upgrade the GPU first for sure though. Personally, I'd be looking at a 6950 1GB for the same price as the 560ti due to better performance at higher resolutions. It for the most part matches or beats the 560ti in benchmarks. But still, pricing is a regional thing and the 560ti is an excellent card, especially for overclocking from what I have heard.
 

ravs316

Distinguished
Apr 14, 2011
29
0
18,530
Thank you so much for the answer :) now that I decided to upgrade GPU I have the following brand options:
■ 560 ti : GALAXY, MSI, ASUS
■ 6950: MSI, XFX, DIAMOND
For my experience. I like the EVGA brand but unfortunately it's not in the store. I had an old 9800gt ASUS but after a year the cooler suddenly died.

My questions are:
■ Which brand from the mentioned above for each video card?
■ In general which brand from your experience would you recommend?

Thanks again for your time
 

ravs316

Distinguished
Apr 14, 2011
29
0
18,530


I just checked again and for the 560 ti I have a lot of options almost all brands including the MSI twin frozr that Uther39 mentioned. Anyone else has an opinion related to the brand?
 

ravs316

Distinguished
Apr 14, 2011
29
0
18,530


Thanks for you answer. I decided to go for the GPU upgrade first, I suppose that at least for this year the latest games won't be taking advantage of quads. what do you think?
 

irsoccer05

Distinguished
Mar 9, 2011
40
0
18,540


Updating the GPU first is generally a good idea. However, I believe Dragon Age II does take advantage of quad cores. Not sure about Crysis 2 though.
 
I guess if you are ready to compromise on the eye candy a bit, your current hardware is a keeper for some more time to come.. However, if you are dead fixed on making an upgrade, route 1 gets my recommendation (i.e. upgrading the entire platform).. You might be able to sell your current hardware (and add a few bucks if required) to get a 560TI also..
 

Dude, seriously? The guy used a GTX590 dual card to test. Of COURSE in this case the CPU would be insufficient.

As for the dragon age benchmark you linked, for all we know he could be using a core i5 2500k, which would be more than sufficient.

The point I am making is that in his case, with his setup, the issues I mentioned are the issues.
 

I agree. Has there been an update or something?
Correction: My mistake, Dragon Age is the CPU hungry one I read about. I have no idea about DAII. So please disregard my first post about this.

Could Ct please post his source rather than just picking benchmark pictures and posting them?
Nevermind, found it, at least, the dragon age II one. Not sure where he got the crysis from:
http://www.winmatrix.com/forums/index.php?/topic/31119-dragon-age-ii-gpu-cpu-benchmark-performance-test/

560 ti : GALAXY, MSI, ASUS
6950: MSI, XFX, DIAMOND


Anyway, I recommend the 6950 1GB over the 560ti. I quite enjoy the XXX one I got on sale from NCIX a while back.
The 560ti apparently has better overclocking headroom but the 6950 performs better at higher resolutions and more often maintains a better minimum framerate. Those are my observations from multiple website reviews. It can also depend on the game though, so research into what games you want to play.
As for power consumption, I've run into multiple contradicting sites. Some bring the 6950 as more efficient and some the GTX560ti. So just assume they are about the same.
http://www.guru3d.com/article/radeon-hd-6950-1gb-vs-geforce-gtx-560-ti-review/4
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/40165-amd-radeon-hd-6950-1gb-review-15.html
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/40119-nvidia-geforce-gtx-560-ti-1gb-review.html
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/40165-amd-radeon-hd-6950-1gb-review.html
Both are excellent cards in the end though at close price ranges so neither choice would be bad.

XFX has and excellent double lifetime warranty - but only if you live in the USA. Live anywhere else and you will get charged all shipping and handling. I don't know much about other brands.
 
I am well aware of the effect of resolution on the CPU load as well as the reasoning for using a high end video card in a test benchmark system.
The high end video card in the test system is the exact reason why I believed the test to not apply, since he does not need to drive such a high end dual video card with his current CPU.
 
he has a point. mounting such high end gpu's on low end cpu's will cause bottlenecks.
the gpu will just overwhelm the cpu to the point it cant cope.
thus slanting the results. not every site is unbias... and often do it to get sponsorship.

if you look at other benches that use a balanced approach to cpu/gpu setup you will find the fps is much better even with lower end cards.
 
When testing a component, the other parts are selected with the intention of eliminating any possible bottlenecks and thus extracting the best performance the component under testing is capable of delivering.. As such, it is quite common to find the most expensive CPU being used when testing a high end video card and vice versa.. The core focus is to find out the tested component's highest capability.. IMO nothing wrong with such approach..
 

That doesn't make sense. It sounds like you are saying a balanced system with the same GPU as a system that has an overkill CPU will get better FPS. I don't see the logic.