Ashes of the Singularity Bench

Kaapstad · 7 Sep 2015 at 00:05

ALXAndy said:
Still not posted mine. I posted and told you my GPU clocks were bone stock and my 5820k was clocked to 4.4ghz yet you seem to have missed it.

Third time lucky and on the scoreboard.

bloodkil · 7 Sep 2015 at 18:52

And here's mine:

290X @ 1200/1500
4790K @ 4.7GHz
4K

Kaapstad · 7 Sep 2015 at 19:43

Scoreboard updated.

bloodkil · 7 Sep 2015 at 20:20

Kaapstad said:
Scoreboard updated.

Cheers, you've put it as a 290P btw, it's a 290X

Kaapstad · 7 Sep 2015 at 20:46

bloodkil said:
Cheers, you've put it as a 290P btw, it's a 290X

Thanks and sorted.

andybird123 · 8 Sep 2015 at 13:09

5820k @ 4.625
980ti 1517 / 2100

1080p

1440p

NOT AN ENTRY - DSR 4K

andybird123 · 8 Sep 2015 at 13:27

5820k @ 4.625
980ti 1497 / 2052

1440p DX11

Kaapstad · 8 Sep 2015 at 16:13

Scoreboard updated.

Zeed · 8 Sep 2015 at 21:09

Well looking at those results I am still waiting to see Overclockers dream Fury X that will destroy everything in DX12.... As here i see 980ti as faster card :/
How can i force myself to buy FX over TI

Ian Evey · 8 Sep 2015 at 21:40

Lord_ZED said:
Well looking at those results I am still waiting to see Overclockers dream Fury X that will destroy everything in DX12.... As here i see 980ti as faster card :/
How can i force myself to buy FX over TI

Being patient is the best option, Far too many unknowns for dx12 yet, Amd might pull a rabbit out of the hat with a driver boost, Or Nvidia might get more performance from async working.

fs123 · 8 Sep 2015 at 22:37

Ian Evey said:
Being patient is the best option, Far too many unknowns for dx12 yet, Amd might pull a rabbit out of the hat with a driver boost, Or Nvidia might get more performance from async working.

Yes definitely something holding the Fury back. Look at the 2160P result. A 290X OC is only 1fps behind the FuryX. On Paper the Fiji core has 50% more SP's and a lot more bandwidth so theoretically should be a fair bit faster. Clock speed could be it but surely the extra bandwidth should have a benefit.

JediFragger · 8 Sep 2015 at 22:49

fs123 said:
Yes definitely something holding the Fury back. Look at the 2160P result. A 290X OC is only 1fps behind the FuryX. On Paper the Fiji core has 50 more SP's and a lot more bandwidth so theoretically should be a fair bit faster. Clock speed could be it but surely the extra bandwidth should have a benefit.

Lack of ROPs throttling it maybe?

andybird123 · 9 Sep 2015 at 08:17

Muziqaz would like all of his bench thread scores deleted

muziqaz said:
And as I mentioned earlier, anyone who are owners of benchmark threads I am participating are more than welcome to delete my entries, since they all are achieved with AMD VSR,

KillBoY_UK · 9 Sep 2015 at 08:27

JediFragger said:
Lack of ROPs throttling it maybe?

+1

fs123 · 9 Sep 2015 at 08:44

andybird123 said:
Muziqaz would like all of his bench thread scores deleted

Maybe a note should be put at the side of his scores since VSR will actually reduce fps a bit. This is interesting because his FuryX is overclocked a bit so should really be more than a couple of fps faster than the stock FuryX result.

Gregster · 9 Sep 2015 at 08:49

fs123 said:
Maybe a note should be put at the side of his scores since VSR will actually reduce fps a bit. This is interesting because his FuryX is overclocked a bit so should really be more than a couple of fps faster than the stock FuryX result.

According to AMDMatt, VSR gives slightly more fps over native resolution. I don't think it is right to have VSR scores but that's not up to me and down to the thread starter/updater.

Kaapstad · 9 Sep 2015 at 09:00

Gregster said:
According to AMDMatt, VSR gives slightly more fps over native resolution. I don't think it is right to have VSR scores but that's not up to me and down to the thread starter/updater.

No VSR or DSR scores for the scoreboard please.

Using VSR or DSR does not give the same performance as the native resolution so can not be used sorry.

fs123 · 9 Sep 2015 at 09:01

Gregster said:
According to AMDMatt, VSR gives slightly more fps over native resolution. I don't think it is right to have VSR scores but that's not up to me and down to the thread starter/updater.

If it is indeed VSR, Musiqaz's 4K result is only 0.6 fps faster so until we have more results it's difficult to determine if VSR is much faster or not. I was under the impression that VSR would take more processing power since the GPU would have to render at 4K then perform scaling to a lower resolution and add whatever postprocessing is needed to make it look good. On true 4K screens the 4K frame buffer is output straight the display.

Kaapstad · 9 Sep 2015 at 09:10

fs123 said:
If it is indeed VSR, Musiqaz's 4K result is only 0.6 fps faster so until we have more results it's difficult to determine if VSR is much faster or not. I was under the impression that VSR would take more processing power since the GPU would have to render at 4K then perform scaling to a lower resolution and add whatever postprocessing is needed to make it look good. On true 4K screens the 4K frame buffer is output straight the display.

Most of the VSR and DSR results I have seen in various benches return slightly higher scores than using the native resolution.

More importantly than that we are comparing DX12 on various cards and allowing VSR and DSR would add another variable that would only serve to confuse the results making the thread pointless.

Gregster · 9 Sep 2015 at 09:22

fs123 said:
If it is indeed VSR, Musiqaz's 4K result is only 0.6 fps faster so until we have more results it's difficult to determine if VSR is much faster or not. I was under the impression that VSR would take more processing power since the GPU would have to render at 4K then perform scaling to a lower resolution and add whatever postprocessing is needed to make it look good. On true 4K screens the 4K frame buffer is output straight the display.

I know that DSR gives me slightly more frames than native for some reason but no idea why. I would trust AMDMatt on this as well, as he has done both.