• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

GPU Bench Overview Table

Phixsator

Phixsator

Phixsator

Soldato
Joined
10 Oct 2012
Posts
4,842
Location
Denmark
Ive put together a table with data from 3 bench sites on the most relavent(objectively) GPUs for me to purchase to help me make up my mind and i thought i would share it incase it might help someone else. This is only for 1080p gamers.

15zsdar.png

Original document uploaded to my google drive -> Download <-

Hope it might be useful to someone else than just me.
 
Good work Phix and a nice idea.

Excuse me now, but i need to rant a bit...

The 7950 results are as usual with 99% of official 7950 benchmarks, useless and inaccurate though because of the well documented throttling problem. In some instances dramatically so.

Actual scores of a 7950 at stock in Bioshock

UUtCzDP.jpg


84fps

Vs techpowerup

77fps

Vs Legion 81fps (better must have used +20% power line)

Actual scores of a 7950 at stock in Hitman

E7OSLBQ.jpg

48fps

Vs Techpowerup

43 fps

Vs

Legion

33 fps (absolutely no idea how they got fps this low)


Battlefield 3 results look way off based off my personal benchmarks. Some of the Tomb Raider results look dubious as well but they vary so much from Techpowerup to Legion that im not sure what settings they used. Its a crying shame that either the AIB's or AMD could not have got this problem solved and would show the 7950 to be the sleeping giant it is in a better light.
 
Last edited:
Well im planning on expanding the spreadsheet further. But there is just so much info to sort through. Im not surprised if some of the review results are off. Next im thinking of finding of Toms review(even though i dont like Toms) and hardwarecanucks.

And Ltmatt you have to also consider the CPU clock of your system versus theirs.. you are running with an average of 600mhz higher.. im not saying its the sole reason for the differences but im sure it will give you a few extra on top.
 
Nice idea and very nice of you to share, I like people like you :p

@LtMatt Off topic, but what OSD is that, I know it's via Riva but Precision doesn't have CPU options for me? :(
 
Hyburnate said:
Nice idea and very nice of you to share, I like people like you :p

@LtMatt Off topic, but what OSD is that, I know it's via Riva but Precision doesn't have CPU options for me? :(
Click to expand...

Afterburner OSD and CPU OSD of temps and thread usage is provided by HWINFO64. :)
 
Good work!!

670 will sit above a 760 below 680, depending what 670 & 680's were used the difference could be minimal at best.

This is why I'm finding it hard to pick my next gpu, 780 is far over priced and the rest all sit in the same perf bracket, then have to weigh up slower bus and 3dvison vs higher bus and no 3d...
 
theone8181 said:
Cause its a 760:)
Click to expand...

Maybe the inner workings of the card, but if you check the reviews, It goes:

760->670->680->770.

Also some OC'd 670's trade blows with a 680 easily, so it's not very nice of you to skip such a powerful card with an excellent price to performance ratio.

Edit: Now that we are at it, someone mentioned a throttling problem on cards. However, you should either compair cards at stock with stock drivers, or OC'd with tweaks. If it only performs as OC'd or tweaked (From what I read, you need to use afterburner and/or registry tweaks to fix the throttling) it has no place in a non-OC table of things.
 
Last edited:
Make said:
Maybe the inner workings of the card, but if you check the reviews, It goes:

760->670->680->770.

Also some OC'd 670's trade blows with a 680 easily, so it's not very nice of you to skip such a powerful card with an excellent price to performance ratio.

Edit: Now that we are at it, someone mentioned a throttling problem on cards. However, you should either compair cards at stock with stock drivers, or OC'd with tweaks. If it only performs as OC'd or tweaked (From what I read, you need to use afterburner and/or registry tweaks to fix the throttling) it has no place in a non-OC table of things.
Click to expand...

You do realize right as i already wrote in the OP that this table was for myself and i WOULD NEVER ever consider a 600 series card. The 680 is only there for reference. Second these numbers are taken from various sites so i have to work with what i have and therefor cannot fairydust some results into existence.

The 7950 in this test is the reference with the boost 925 core bios and i have never seen that throttle. Only the broken sapphire edition(with a default vddc of 1225!!!!!) have i seen have throttle issues.
 
And after the OP you wrote you are planning to expand the table, hence the expansion suggestions and ideas.

And no, I did not know you would NEVER EVER consider a series 600 card if you had one in the table already.

Calm down sir.
 
Make said:
And after the OP you wrote you are planning to expand the table, hence the expansion suggestions and ideas.

And no, I did not know you would NEVER EVER consider a series 600 card if you had one in the table already.

Calm down sir.
Click to expand...

Im very calm *starts humming*

Note: the expansion was towards more review sites not more cards as it makes little sense considering every OC model out there since this is not about brands but models etc 760 vs 770 not gigabyte OC vs msi lightning. Hope it makes sense and i understand that it isnt very clear from the spreadsheet itself at this point.

pgi947 said:
Good work!!

670 will sit above a 760 below 680, depending what 670 & 680's were used the difference could be minimal at best.

This is why I'm finding it hard to pick my next gpu, 780 is far over priced and the rest all sit in the same perf bracket, then have to weigh up slower bus and 3dvison vs higher bus and no 3d...
Click to expand...

I know im going to say something very terrible here but ive actually been using TriDef a bit and in some games the depth is actually so much better than 3dvision, like in skyrim. But at this point i dont care about 3d due to all the issues it has (cross talk being the most annoying of all). VR is the thing to go for when its ready :)
 
Last edited:
Only thing with tridef is the other monitor I'd have to track down and then add to the cost. If amd's cards are too fast/cheap to turn down I may try some form of hybrid setup and just fire up my 670 when I want to try 3d.

Its never easy spending money is it :(

Edit: if you want a wider choice of cards on certain benches check out anandtech's gpu bench, though I couldn't confirm if they are still showing amd results pre 12.11
 
pgi947 said:
Only thing with tridef is the other monitor I'd have to track down and then add to the cost. If amd's cards are too fast/cheap to turn down I may try some form of hybrid setup and just fire up my 670 when I want to try 3d.

Its never easy spending money is it :(

Edit: if you want a wider choice of cards on certain benches check out anandtech's gpu bench, though I couldn't confirm if they are still showing amd results pre 12.11
Click to expand...

Thats the major problem when looking at these benchmarks sites. A lot of them doesnt update their scores when a new major driver comes out. Its a nightmare sorting through all that info. I sometimes dream about winning the lottery(like every other guy) and then do some proper benching of current and last gen hardware.
 
Good idea although I would like to see the same table for 1440p setups :D
Thinking of getting a GTX770 and was wondering whether the extra performance out weighs the memory bandwidth on the HD7950 at 1440p :o
 
You must log in or register to reply here.
Back
Top Bottom