Destiny 2: PC graphics benchmark

GoogalyMoogaly · 31 Oct 2017 at 23:14

shankly1985 said:
Again, the FE 10 series cooler is much better. Out of the box it automatically overclocks the GPU much higher than a poor out of the box vega GPU.

How is it fair to have one GPU that will auto boost much higher vs a GPU that will just down clock out of the box?

A simple tweak to the Vega GPU lowering voltage is enough, don't even need to touch the core or memory speed and you will gain much better performance and higher clocks just like a 10 series does.

It's amd biggest downside the stock ref cooler looses them a lot of benches.

But it's how AMD ship it, that's how they intended it (apparently).

I'm sure if you fitted a waterblock to the 1080Ti and flashed a custom BIOS to it you could gain performance out of that too.

Even AMD made it **** out-of-the-box, then that's on them. You wouldn't underclock a factory overclocked card because a factory OC was too good out-of-the-box.

shankly1985 · 31 Oct 2017 at 23:17

Gregster said:
I always run my 1080Ti out the box. I don't have time for messing about tweaking this and that to get it to run normally and comparing a tweaked card to a standard card is just plain wrong.

Mind you, Shankly see's it as fair, which isn't surprising lol

And why should you ? It already does very well and will boost to very high core clocks out of the box. Where VEGA will just throttle right down to 1000 core clock from its rated 1500+ do you guys understand how much performance is lost alone here? and can be gained by simply undervoltage!!

Surely if we comparing Ref GPUs we should show them in there best lime light! Am not talking about Overclocking the core or memory to gain performance I saying reduce the heat and get higher boost clock to least the rated speed of the GPU.

shankly1985 · 31 Oct 2017 at 23:18

GoogalyMoogaly said:
But it's how AMD ship it, that's how they intended it (apparently).

I'm sure if you fitted a waterblock to the 1080Ti and flashed a custom BIOS to it you could gain performance out of that too.

Even AMD made it **** out-of-the-box, then that's on them. You wouldn't underclock a factory overclocked card because a factory OC was too good out-of-the-box.

In typical GM fashion you missing the point and twisting posts!! So I end it here and go to bed!

GoogalyMoogaly · 31 Oct 2017 at 23:34

shankly1985 said:
And why should you ? It already does very well and will boost to very high core clocks out of the box. Where VEGA will just throttle right down to 1000 core clock from its rated 1500+ do you guys understand how much performance is lost alone here? and can be gained by simley undervoltage!!

Surely if we comparing Ref GPUs we should show them in there best lime light! Am not talking about Overclocking the core or memory to gain performance I saying reduce the heat and get higher boost clock to least the rated speed of the GPU.

shankly1985 said:
In typical GM fashion you missing the point and twisting posts!! So I end it here and go to bed!

Missing the point? That comparing results run on a different PC by a different person with a tweak watercooled Vega 64 against a stock 1080Ti FE isn't really a fair comparison?

If we're showing them in the best light surely pick pick a watercooled 1080Ti and then OC both cards as far as you can. You still negate the out-of-the-box issue of the Vega but you also show the 1080Ti in the best light.

5 of my mates bought PCs from OcUK in the last year or so and I doubt any of them has bothered tweaking voltages or power limits or memory clocks or even fan profiles. They just use the GPU as it is. To all the people that just do that tweaking things for a benchmark is giving them a false idea of what to expect from the card the way they'll use it.
Sure, maybe a good number of us on here would tweak things, but a good number of us on here would probably overclock the card too if we're going to the effort of tweaking things.
I can't think who would be willing to tweak settings on a Vega but refuse to do anything like that on an Nvidia card unless they were just trying to be awkward.

I can't see how tweaking only one is fairer than tweaking neither or tweaking both to perform as well as possible.
All this is ignoring that they're not even being run on the same PC! (unlike all of the usual comparison video like that, that that video took the 1080 and 1080Ti footage from)

imginy · 1 Nov 2017 at 10:53

What's this game run like on a 4 core i5?

Psycho Sonny · 1 Nov 2017 at 11:32

imginy said:
What's this game run like on a 4 core i5?

absolutely fine. i was running it on a 7600K but now on an 8400.

the video i posted above is from the 8400. my gtx 1080 is currently underclocked 10% to keep it running cool. trying to get hold of a couple of 120 mm fans to stick underneath the card blowing air directly into the card and it's own blower style fan to force air into into it an over it. i'm also running it on the absolute max settings at 1440p.

if you play at 1080p and 60 hz i imagine this game will run amazingly well on a 1060 6GB and a modern i5. it's a very efficient game. i could probably turn down a few settings slightly and hit 144hz constant.

imginy · 1 Nov 2017 at 11:42

Psycho Sonny said:
absolutely fine. i was running it on a 7600K but now on an 8400.

the video i posted above is from the 8400. my gtx 1080 is currently underclocked 10% to keep it running cool. trying to get hold of a couple of 120 mm fans to stick underneath the card blowing air directly into the card and it's own blower style fan to force air into into it an over it. i'm also running it on the absolute max settings at 1440p.

if you play at 1080p and 60 hz i imagine this game will run amazingly well on a 1060 6GB and a modern i5. it's a very efficient game. i could probably turn down a few settings slightly and hit 144hz constant.

Ok cool thanks I just got sent the code and wasn't sure if I should wait till I get the 8400 before I play.

I will give the game a blast later today.

Psycho Sonny · 1 Nov 2017 at 12:37

imginy said:
Ok cool thanks I just got sent the code and wasn't sure if I should wait till I get the 8400 before I play.

I will give the game a blast later today.

you should have no issues hitting similar frames to what I'm getting currently. Make sure you have the latest drivers and you are set. this game doesn't require a lot of power, baring in mind how good it looks and the settings are maxxed out on my rig it's a very efficient game. i might try the customised game profile from geforce experiece for optimised settings. that will probably help me hit the 144hz I want to without having to compromise too much on lowering settings.

shredgodxp · 2 Nov 2017 at 07:23

Playing on an R9 390 with 4690k (not OCd at all) at 1440p with high settings and getting decent performance throughout. No issues what so ever. Game looks awesome and the performance is rather good.

chroniclard · 2 Nov 2017 at 07:44

I tried it on my work PC, 660 gtx, runs 1080p on low at 60fps and actually looks ok.

Martini1991 · 2 Nov 2017 at 08:01

shankly1985 said:
Again, the FE 10 series cooler is much better. Out of the box it automatically overclocks the GPU much higher than a poor out of the box vega GPU.

How is it fair to have one GPU that will auto boost much higher vs a GPU that will just down clock out of the box?

A simple tweak to the Vega GPU lowering voltage is enough, don't even need to touch the core or memory speed and you will gain much better performance and higher clocks just like a 10 series does.

It's amd biggest downside the stock ref cooler looses them a lot of benches.

Out of the box is the fairest comparison for most people. It's also how the majority of people will run them.
Vega's a mess out of the box, that's AMDs decision to ship it that way.

Martini1991 · 2 Nov 2017 at 08:02

shankly1985 said:
And why should you ? It already does very well and will boost to very high core clocks out of the box. Where VEGA will just throttle right down to 1000 core clock from its rated 1500+ do you guys understand how much performance is lost alone here? and can be gained by simply undervoltage!!

Surely if we comparing Ref GPUs we should show them in there best lime light! Am not talking about Overclocking the core or memory to gain performance I saying reduce the heat and get higher boost clock to least the rated speed of the GPU.

But that's how AMD have released it. That's Vega running at officially supported specifications!

Muhammad.Abdelhameed · 15 Nov 2017 at 20:01

NVIDIA just posted a whooping 50% improvement in Destiny 2 with their new drivers:

destiny-2-geforce-game-ready-driver-388-31-whql-3840x2160-performance-improvements-640px.png

destiny-2-geforce-game-ready-driver-388-31-whql-2560x1440-performance-improvements-640px.png

https://www.geforce.com/whats-new/articles/star-wars-battlefront-ii-game-ready-driver

LoadsaMoney · 15 Nov 2017 at 21:15

Thats a big increase.

GoogalyMoogaly · 15 Nov 2017 at 21:34

Muhammad.Abdelhameed said:
NVIDIA just posted a whooping 50% improvement in Destiny 2 with their new drivers:

https://www.geforce.com/whats-new/articles/star-wars-battlefront-ii-game-ready-driver

Waiting for someone to come along and say that Nvidia's graphs are going to be biased and show best case scenario.

So many haters!

shankly1985 · 15 Nov 2017 at 21:45

GoogalyMoogaly said:
Waiting for someone to come along and say that Nvidia's graphs are going to be biased and show best case scenario.

So many haters!

To be honest though dont AMD and Nvidia always over do game benchmarks? Also this results was done when the CPU wasn't a bottleneck most gamers wont have a i7 7820x so it will be interesting in seeing results from say a i7 7700k or 8700k mainstream CPUs..

Anyway well played to Nvidia! But I feel these results are a best case scenario.

At 2560x1440, on our Intel Core i7-7820X system, which eliminates any CPU bottlenecks, our testing revealed performance improvements of up to 40% when using the new Game Ready driver, with an average improvement of 36.4% recorded across 6 different GPU configurations:

Muhammad.Abdelhameed · 15 Nov 2017 at 21:47

GoogalyMoogaly said:
Waiting for someone to come along and say that Nvidia's graphs are going to be biased and show best case scenario.

Independent site testing will soon follow, this situation is the same as Forza 7.

shankly1985 said:
Also this results was done when the CPU wasn't a bottleneck most gamers wont have a i7 7820x so it will be interesting in seeing results from say a i7 7700k or 8700k mainstream CPUs..

Focus on the 4K numbers then, those are GPU limited and are not affected by the CPU. Also a 8700K is faster and better than a 7820X.

TheRealDeal · 15 Nov 2017 at 21:48

GoogalyMoogaly said:
Waiting for someone to come along and say that Nvidia's graphs are going to be biased and show best case scenario.

So many haters!

They are a best case scenario and why wouldn't they be. I would expect AMD to do the same. If the average bump is 40% then it's still very impressive. Damn i am usually impressed by 20% never mind what they have done here. Even more impressive is the fact that they already had good numbers in this game. It is a Nvidia sponsored game though so i suppose there numbers were a little behind where you would expect comparing to AMD.

Gregster · 15 Nov 2017 at 22:07

LOL - Usual response to a boosting driver from the usual AMD owners when it comes to NVidia. I like seeing big improvements for both AMD and NVidia but never understand why some seem to abhor improvements for the opposite GPU manufacturer

TheRealDeal · 15 Nov 2017 at 22:10

Gregster said:
LOL - Usual response to a boosting driver from the usual AMD owners when it comes to NVidia. I like seeing big improvements for both AMD and NVidia but never understand why some seem to abhor improvements for the opposite GPU manufacturer

My response is positive and the truth so i don't get your point. What part of impressive didn't you get.