The OcUK ComputeMark Leaderboard

shankly1985 · 30 Apr 2014 at 02:09

humbug said:
I'm going to add that Stock bench you talked about later or probably tomorrow.

Thanks will give us all better idea how GPU perform from out the box. Overclocking can be miss leading. People on here have super, duper water cooling us guys have no chance

humbug · 30 Apr 2014 at 12:47

R9 290 @ 1165/1300
Phenom II X6 @ 4.1Ghz
Score 8357

humbug · 30 Apr 2014 at 13:05

*GPU Comparison*

Powercolor PCS+ R9 290 (1040/1350)

humbug · 30 Apr 2014 at 13:06

Out of the box comparison table here

http://forums.overclockers.co.uk/showpost.php?p=26225041&postcount=2

please only use the Standard benchmark mode.

For 'Out of the box' comparisons we have a stock clock results table here http://forums.overclockers.co.uk/sho...41&postcount=2
This is to compare the performance between different brands of GPU without any user overclocking, for this please add the make, model, standard clock speeds and the tag *GPU Comparison* to your results page.

MyBrainz · 30 Apr 2014 at 13:18

So my 7950 is a factory overclocked version @ 950c/1250m and i usually underclock to a stock setting @ 800c/1250m when doing a stock run. Does anybody know the original stock clocks for a 7950? I see 880 core for some but need some verification.

Thanks

LtMatt · 30 Apr 2014 at 13:24

Sorry Humbug i can't remember what clocks i used now. It was either 975/1250 or 1050/1250. I will run it again at some point. Its a pain in the ass for me as i have to change my magnifier settings every time i want to run this restart and change it again so can't see myself running this more than a couple of times.

GoogalyMoogaly · 30 Apr 2014 at 13:28

humbug said:
Stock Clock Results - Standard

1, Humbug - Powercolor PCS+ R9 290 (1040/1350) Score 7771 - Link

I'm confused, you say 'stock clocks' and then show a factory overclock of 93Mhz on the core?

humbug · 30 Apr 2014 at 13:30

MyBrains said:
So my 7950 is a factory overclocked version @ 950c/1250m and i usually underclock to a stock setting @ 800c/1250m when doing a stock run. Does anybody know the original stock clocks for a 7950? I see 880 core for some but need some verification.

Thanks

The original 7950 is 800/1250.

If your adding to the GPU comparisons then you need to run it at whatever speed its at out of the box.

humbug · 30 Apr 2014 at 13:32

LtMatt said:
Sorry Humbug i can't remember what clocks i used now. It was either 975/1250 or 1050/1250. I will run it again at some point. Its a pain in the ass for me as i have to change my magnifier settings every time i want to run this restart and change it again so can't see myself running this more than a couple of times.

Press reset on MSI AB, then you will know, its 947/1250, BTW

GoogalyMoogaly said:
I'm confused, you say 'stock clocks' and then show a factory overclock of 93Mhz on the core?

I will change the wording.

If this is to difficult i'm not going to do it, it should be really simple.

shankly1985 · 30 Apr 2014 at 13:32

I think factory Overclock here will be fine. its when we start seeing 1200+ things kinder get pointless, when trying to find out real world performance.

MyBrainz · 30 Apr 2014 at 13:34

humbug said:
The original 7950 is 800/1250.

If your adding to the GPU comparisons then you need to run it at whatever speed its at out of the box.

Cheers.. will do a run later

Kaapstad · 30 Apr 2014 at 13:35

If I could get this to do CF I would give it a go.

LtMatt · 30 Apr 2014 at 13:35

GoogalyMoogaly said:
I'm confused, you say 'stock clocks' and then show a factory overclock of 93Mhz on the core?

That's stock for his card though. I know what you're saying but its easy enough to find out what are stock clocks on any particular card. AIB cards are almost always going to have higher stock clocks than our reference cards.

At least my reference cards are clocked 28mhz higher than yours though.

shankly1985 · 30 Apr 2014 at 13:39

Kaapstad said:
If I could get this to do CF I would give it a go.

Crossfire worked for me when selecting full screen.

GoogalyMoogaly · 30 Apr 2014 at 13:44

LtMatt said:
That's stock for his card though. I know what you're saying but its easy enough to find out what are stock clocks on any particular card. AIB cards are almost always going to have higher stock clocks than our reference cards.

At least my reference cards are clocked 28mhz higher than yours though.

I would argue they are 'default' clock for his card. As his card has a factory OC I don't think they are 'stock' clocks.

I don't see the point in the chart using default clocks as this will just vary based on what card you've got and are no more useful than comparing OC clocks. If it used 'stock' clocks, i.e. 947/1250 for a 290, 1000/1250 for a 290X then you could see how the cards compare at the same clocks to see if there's anything else that might make a difference (i.e. VRAM make).
Using 'default' clocks just seems like it punishes peopel that bought earlier non-OC'd cards.

Another thing that could be interesting if finding clocks that a lot of cards can do and clock them all to that. For example say 800/1250 and then have everyone clock to that whether they have a 290, 290X, 7970, 280X, 780Ti...

humbug · 30 Apr 2014 at 13:49

GoogalyMoogaly said:
I would argue they are 'default' clock for his card. As his card has a factory OC I don't think they are 'stock' clocks.

I don't see the point in the chart using default clocks as this will just vary based on what card you've got and are no more useful than comparing OC clocks. If it used 'stock' clocks, i.e. 947/1250 for a 290, 1000/1250 for a 290X then you could see how the cards compare at the same clocks to see if there's anything else that might make a difference (i.e. VRAM make).
Using 'default' clocks just seems like it punishes peopel that bought earlier non-OC'd cards.

Another thing that could be interesting if finding clocks that a lot of cards can do and clock them all to that. For example say 800/1250 and then have everyone clock to that whether they have a 290, 290X, 7970, 280X, 780Ti...

Your over thinking it Googaly, it punishes no one more than if you don't have a water cooled lightning on the main bench table, its just so you can see what the performance is in reality and the performance difference you get out of the box between different brands.

GoogalyMoogaly · 30 Apr 2014 at 13:53

humbug said:
Your over thinking it Googaly, it punishes no one more than if you don't have a water cooled lightning on the main bench table, its just so you can see what the performance is in reality and the performance difference you get out of the box between different brands.

I suspect that the cards with a faster factory OC will be faster.
Interested to see if i'm right and what else it will tell us.

Kaapstad · 30 Apr 2014 at 13:53

shankly1985 said:
Crossfire worked for me when selecting full screen.

Just tried again with full screen and I scored a lot less than a single HD 7950.

humbug · 30 Apr 2014 at 13:54

Kaapstad said:
Just tried again with full screen and I scored a lot less than a single HD 7950.

Just go into CCC and turn the other 3 off for crying out loud

layte · 30 Apr 2014 at 14:03

Does a benchmark exist that can compare all the different GPU compute languages. I'd be very interested to see how CUDA, OpenCL, Direct Compute and stuff using GLSL stack up.

Competitor rules

***The OcUK ComputeMark Leaderboard***

The OcUK ComputeMark Leaderboard