Bought a Vega Liquid - thread #2 (last one closed)

File: gpuZ-vega-17-8-17 - Copy.jpg (217KB, 605x759px) Image search: [Google]

217KB, 605x759px

>>61944046

Can you run fallout 4 in 1080p please

File: 3DMark-Firestrike-1stRun-Stock-win7Driver-win8-1---17-8-17.jpg (605KB, 2532x1293px) Image search: [Google]

605KB, 2532x1293px

>>61944290

>>61944436
that puts it in between a titan x (22,600)and 1080ti FE (28,800) in graphics score. not bad

Why you bought it?

>>61944350

fallout 4 is capped at 60fps

the engine is tied to the framerate, so if you unlock the framerate, the speed of the game increases accordingly (above 60fps). what you're left with is NPCs being finished with their animations before the sound has even played. it's funny, but sucks when you're trying to enjoy the game.

>>61944489
Titan X Maxwell?

>>61944553
>flat 120hz master race
Seems like you need to tweak some stuff

>>61944566

what did you tweak? and do you have steam framerate indicators? i would like to see proof, or at least what you fixed.

File: Unigine-Heaven-1stRun-Stock-win7Driver-win8-1---17-8-17 - Copy.jpg (138KB, 881x1008px) Image search: [Google]

138KB, 881x1008px

>>61944436

>>61944553
>enjoying Fallout 4

File: Unigine-Heaven-1stRun1080p-Stock-win7Driver-win8-1---17-8-17 - Copy.jpg (146KB, 1017x1021px) Image search: [Google]

146KB, 1017x1021px

>>61944596

@1080p

>>61944350

Which area do you want me to test? AFAIK there is no benchmark tool for it, so I'll just walk around and records fps..

>>61944782
He said trying to enjoy, not succeeding.

>>61944867
Fair enough.

>>61944847
>K CPU
>not overclocked
Anon pls

>>61944912
>K CPU
>It's a 6-core
>not overclocked, even though there is no cheaper 6-core at 3.6GHz and it actually makes sense if all he wanted was that.

>>61944940
Isn't there a non-k version?

>>61944953
http://lmgtfy.com/?q=i7-6850

File: Unigine-Valley1stRun1440p-Stock-win7Driver-win8-1---17-8-17 - Copy.jpg (141KB, 1012x1008px) Image search: [Google]

141KB, 1012x1008px

>>61944847

Unigine VALLEY @ 1440p

File: Unigine-Valley1stRun1080p-Stock-win7Driver-win8-1---17-8-17 - Copy.jpg (139KB, 1066x942px) Image search: [Google]

139KB, 1066x942px

>>61945013

Unigine VALLEY @ 1080p

File: IMG_0442 - Copy.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945029

Vega installed

File: IMG_0468 - Copy.jpg (3MB, 3024x4032px) Image search: [Google]

3MB, 3024x4032px

>>61945488

File: IMG_0473 - Copy.jpg (3MB, 3024x4032px) Image search: [Google]

3MB, 3024x4032px

>>61945513

>>61945488
>>61945513
>>61945521
how long are those pipes?

It has such a beautiful shroud

>>61945527

60cm

>>61945534
Gj Mr shroud

>>61945552
that's cool, did you notice any pump noise? the radiator is 120mm right?

>>61944489
You should say, puts it around titan x, you imply it falls in the middle.

>>61944596
Not trying to shit it up too bad in here, but for reference scores with same settings from my 980ti and 1080ti, will run test and post scores if wanted
Running on r5 1600 3200mhz CAS 16 1x8gb ram
980ti overclocked - 1440p=1500, 1080p=2200
1080ti stock with slightly more aggressive fan curve - 1440p=2300, 1080p=3500
Shit isn't looking good boys

>>61945567

The radiator is 140mm and the fan is 120mm.

When benchmarking I could hear a louder noise from the pump - though in some scenes it would drop the noise level down towards idle noise.

>>61945582
Valley scores:
980ti - 1440p=2500, 1080p=3971
1080ti - 1440p=3700, 1080p=5100
Justctrying to add some context here, with my real world testing on a dirty windows install with a similar 6 core processor

>>61945595
dang, why wouldn't they put a 140mm fan on it?

>>61945604
There are very few effective 140mm static pressure fans, and there are none that out lass something like say a gentle typhoon. You're better off with a gasket or adapter and a 120mm
Same idea behind why 200mm isn't popular for water cooling even given the massive airflow and space that would allow with virtually no noise, static pressure a shit

>>61945604

I'm think of replacing it with an led fan actually...

>>61945582
primitive shaders are turned off anon, if RTG fixes their shit up we'll see at the bare minimum a 30% perf improvement

>>61945616
fair enough, if it cools the gpu properly there's no problem though
>>61945617
wouldn't you have to take it apart to do that, voiding the warranty in the process?

>>61945630
im an amd fan, i really am. the only reason i even have these cards are because i got them for incredible prices, but come on now anon

File: IMG_0478.jpg (3MB, 3024x4032px) Image search: [Google]

3MB, 3024x4032px

>>61945029

Total War: Warhammer @ 1080p (Std Benchmark)

Settings: Dx11 All Ultra, AAx8

>>61945029
>>61945013
>>61944847
>>61944596
>>61944046
op, >>61945582
>>61945598 guy here
is there any way you could overclock your cpu and give us some results with overclocking? that could be limiting the card
also, if others are interested, I can redo tests with the 1080ti with my r5 at 3.6 to match op

the fact that you already found one and bought it is a miracle...
amd already stated that they had stocked 3 million chips and they are down on 30% after 4 days

>>61945652
the cards are bandwidth starved because of all the fucktons of extra triangles they're having to draw, just like previous GCN cards, the primitive shaders remove all the triangles that are outside the player's FoV, helping a fuckton on the bandwidth, and allowing TBR to do its work properly

Here's an interview explaining it in more detail
https://www.youtube.com/watch?v=XC5Dy_b-kE8

File: IMG_0480.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945682

File: IMG_0481.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945697

>>61944046
Run the thing featured in https://www.youtube.com/watch?v=Nc6R1hwXhL8 and tell us if Vega's tile-based stuff even works or if it's yet another thing AMD failed on for release.

>>61945630
>>61945696
so when will they be turned on?

>>61945652
its true
https://www.computerbase.de/2017-08/radeon-rx-vega-64-56-test/7/#diagramm-deus-ex-mankind-divided-hbcc-aktiv-16-gb
nothing from the new features is working on the gaming driver while hbcc is working on the workstation driver..

i think we can all agree that amd just didnt care much about gaming or most probably they didnt had enough money to develop at once workstation and gaming drivers

File: gtx1080.png (31KB, 744x618px) Image search: [Google]

31KB, 744x618px

>>61945029
This can't be real, right? 80 FPS average.
My pic is what I get with a GTX 1080.

File: IMG_0482.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945716

As above 1080p with FXAA instead of 8xAA

File: IMG_0483.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945736

Is this the housefire thread?

>>61945696
There will be a difference for sure, but still, you pulled that 30% perf improvement out from your ass. For some reason you all always overhype yourself into eventual disappointment.

>>61945728
no one knows, it works just fine on the pro drivers, which were the ones RTG gave their attention first.
Maybe they manage to enable it for when the AiBs start getting released in September

File: IMG_0484.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945747

Again 1080p with FXAA but with DoF OFF and Screen Space Reflections OFF (both quite costly)

>>61945696
no im familiar with how it works, but why isnt it enabled already? what is stopping them from working everyone they have to get this obviously killer feature working?
polaris was bandwidth starved too, but obviously there wasnt an arch feature to fix that, but still. 30% is a lot anon, a LOT.
>>61945730
its not that they arent in the gaming driver, HBCC is just disabled by default, because very little currently takes advantage of it

File: IMG_0489.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945776

>>61945735
this lines up with my scores with my 1080ti and the known and tested performance difference between the 1080 and 1080ti
so yeah it would appear so

>>61945777
what you mean very little? everything can take advantage of it considering what he does its program agnostic

>>61945777
they didn't give a shit for the gaming drivers, they were more worried with the professional/enterprise stuff first, because those are the ones that will sell better

File: IMG_0490.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945785

>>61945792
not him, but maybe because outside of really high res textures on modded games, most of the time the dataset is small enough to fit on VRAM entirely

>>61945792
look at the performance difference in the reviews where it is tested. little to no difference as we have seen so far
>>61945795
i believe it, especially given that they've already released a mining driver before fixing gaming drivers. I just dont see 30% coming out of this arch. i definitely see 10 to maybe 15%, maaybe 20% in a few years or in dx12 or vulkan, but not 30%

File: IMG_0491.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945797

More Total War: Warhammer @ 1440p

File: warhammer.png (491KB, 884x597px) Image search: [Google]

491KB, 884x597px

>>61945682
>>61945716
This is what I got with my GTX 1080.

I didn't even close stuff like browsers when running these. How is the liquid cooled Vega getting such poor performance? Shouldn't it be better than that?

Isn't it supposed to be more or less equal to a GTX 1080? Warhammer is an AMD Gaming Evolved title even.

>>61945807
exactly. the only title i could think of would maybe be shadow of mordor at 4k+ with the ludicrous texture pack on there, or maybe heavily modded skyrim

File: IMG_0492.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945814

All settings @ ultra, 8xAA, DoF ON, SSR ON

>>61945823
>>61945814
this is interesting. vega anon, can we see an afterburner chart of clockspeeds, temps, power draw, etc.? interested to see where the card is really running

File: IMG_0496.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945830
>>>61945807

i really feel like the cpu is a limiting factor right now, especially for 1080p. OP is running around 1ghz slower than most other intel chips used for benchmarking as of now, which is probably contributing to at least a 10% drop in performance

If you're benching shit, try with AA off and high settings rather than ultra 16xAA that reviewers use. Especially if you can do 4k even if it's with a virtual resolution

File: IMG_0497.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945841

File: IMG_0498.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945855

FXAA, DoF ON, SSR ON

File: IMG_009889.jpg (2MB, 3264x1836px) Image search: [Google]

2MB, 3264x1836px

HOUSEFIRE BRIGADE REPORTING IN

File: IMG_0503.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945874

File: IMG_0504.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945885

File: IMG_0505.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945897

Ok last one for Total War: Warhammer

FXAA, DoF OFF and SSR OFF (most favourable)

File: IMG_0511.jpg (3MB, 4032x3024px) Image search: [Google]

3MB, 4032x3024px

>>61945912

>>61945912
DX11 or 12?

File: IMG_0514.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61945925

>>61945927

Warhammer done on dx11, since I'm on win8.1 using win7 VEGA drivers.

>>61945844
I think it very well could be. I run an i5-6600k clocked to 4.4 and a gtx 1080 gaming x on a 144hz monitor (but only 1080 for now) and the CPU definitely can be the limiting factor in a handful of games when I check on afterburner. Gta5 is the obvious one but there are some others.

>>61945955

I plan to undervolt the vega according to the advice from the prev thread then re-bench - with MSI Afterburner running.

I will then try to oc the cpu + ram and test again.

>>61945985
>ask op for afterburner
>ask op for cpu oc
>op is aware of undervolting helping
>op plans to deliver
op is somewhat based

>>61944530

My previous card was a Sapphire rx480 8GB Nitro+, decent but for 1440p I need more power.

>>61945998
So why not a 1080ti ??

>>61945985
Nice. Based op.
To be fair, I think my i5 being the limiting factor is mainly due to only 4 cores, which of course isn't the case for you. But I do wonder if in poorly threaded games that a lower clockspeed can be a bottleneck.

>>61945998
I have the exact same card and it struggles at arma 3. I hope vega does better so I can consider a nano later.

>>61945029
>>61944847

>valley 1080p extreme hd
my 4790k + 1080ti scored 5700
fps etc idk, screenshot didn't save.

>heaven 1080p extreme hd
scored 3647, max fps 304, min 36.3

vega is a huge ass flop

>>61946492
* heaven avg fps was 144.8

>>61944553
holy shit, how old is the engine?

>>61944489
Titan X or Titan XP (not to be confused with Titan Xp)?

File: Unigine-Valley1440p-StockVega-OCcpu-win7Driver-win8-1---17-8-17a - Copy.jpg (2MB, 2560x1440px) Image search: [Google]

2MB, 2560x1440px

>>61945945

I ran into a few problems but it seems as though I was successful in overclocking the cpu to 4ghz and getting the ram to run at xmp-3002 (1501Mhzx2).

So far I only successfully re-benched Unigine Valley @ 1440p Ultra

File: Unigine-Valley1440p-StockVega-OCcpuSettings-win7Driver-win8-1---17-8-17 - Copy.jpg (886KB, 2523x1067px) Image search: [Google]

886KB, 2523x1067px

>>61948230

Here are the screencaps of speccy, gpu-z and cpu-z for the above benchmark.

>>61944596
Why do you put windows nt like an autist?

>>61945029
Why does it say 4gb gpu?
Man this thread is so weird I'm getting anxious here

File: Unigine-Valley1440p8xAA-StockVega-OCcpu-win7Driver-win8-1---17-8-17 - Copy.jpg (2MB, 2560x1440px) Image search: [Google]

2MB, 2560x1440px

>>61948254

Valley @ 1140p Ultra 8xAA

>>61948397

meant to say 1440p

>>61948315

The benchmark picks the string to put there.

>>61945582
8xMSAA is not working correctly on Vega at the moment, turning it on in any game massively tanks Vega performance relative to turning it off. That's what's tanking the results in those benchmarks.

File: 3DMark-Firestrike-2ndRun-StockVega-OCcpu-win7Driver-win8-1---17-8-17 - Copy.jpg (694KB, 2559x1385px) Image search: [Google]

694KB, 2559x1385px

>>61948397

3dmark FireStrike1.1 with OC cpu @ 4Ghz and ram @ 3002Mhz (prev 3.6Ghz and 2132Mhz)

File: ARMA3-ScenarioYAAB-StockVEGA-OCcpu-17-8-17 - Copy.jpg (215KB, 2560x1440px) Image search: [Google]

215KB, 2560x1440px

>>61948588

Arma 3 @ 1440p benchmarked with Yet Another Arma Benchmark (YAAB) Scenario from Steam Workshop.

With OC Cpu and Ram.

>>61948717
Could you also run it at 1080p ultra too?

>>61948588
my 1080ti scored 30k

vega is a flop

>>61947020
2002

File: RU0kIdU.jpg (177KB, 1734x1267px) Image search: [Google]

177KB, 1734x1267px

>>61944489

>between a titan x (22,600)and 1080ti FE (28,800)

What did you say, amigo?

>>61948827

Yep sure thing.

Will be redoing warhammer, trying starcitizen and bf1.

How is this thread so friendly

>>61950166
It's because OP is a nice person that listens to suggestions.

>>61946366
Isn't Arma mostly cpu bound? Even a 1800x has trouble with it. An oced i7 69xx is probrably the best cpu out atm for it.

>>61951318
kek, shut up

>>61951318
Mostly but I want to know if a Vega changes that.

>>61944046
>128GB RAM
>2x27" with different native resolutions

WEW LAD

What do you use your rig for?

>>61952874
Arma uses a 2core engine. for that case the best solution would be a high clocked 4core processor

>>61949427
stock founders cards, not your liquid cooled 1080 with max oc. vega partner cards with better performance will come.

File: chrome_2017-08-17_19-52-43.png (106KB, 1009x294px) Image search: [Google]

106KB, 1009x294px

>>61956112

>liquid cooled

I don't think so, friendo.

>>61947020
it's a fork of gamebryo, which bethesda first used in morrowind.

>>61949359
>>61956450
i can already imagine how frankenstein the code looks

>>61956450
>fallout4 is built on the morrowwind engine aka gamebryo.

no wounder it chrashes so damn much lol.
every half hour it crashes and auto hard restart on my PC

>>61944553I tried unlocking framerate and using frame rate limiting at 60fps to fix a hitching issue. Broke the animation for entering a vault and I had to fix the config and reload a previous save to fix it. I do like the game though.

>>61956112
YOULL SEE!! THE PRODUCT I BOUGHT BRAND NEW WILL WORK HOW ITS INTENDED MONTHS FOR NOW!!!! FINE WINE!!!!

fuck off. glad i sold my fury x and got a 1080 yesterday.

Raja is a dumb fuck, thank god for based jim keller and Ryzen,

File: vega.jpg (119KB, 1440x1343px) Image search: [Google]

119KB, 1440x1343px

Ayyy OP
I'm just waiting on my other parts to arrive now.

>>61958362
>Paying $600 for an inferior product.
Where your price to performance ratio argument now?

>>61958941
It depends on what you use the card for...

If you're using it for just gaming, you can get a 1080 for cheaper... for now at least.

But if you're using it for productivity, it's still 100 bucks cheaper than a 1080 Ti while having equal - and in some cases, better - performance.

So Vega 64 is a great all-around card, and when prices come back down, it'll be the best bargain out there until Volta comes out.

>>61959586
you dont do productivity. you game and you're a shill.

File: IMG_0519 - Copy.jpg (2MB, 4032x3024px) Image search: [Google]

2MB, 4032x3024px

>>61948717

Photo of the old video cards, rx480 nitro+ and rx470, pretty hard to tell them apart.

Arma3 results coming up, then Warhammer.

After than bf1, titanfall2, then starcitizen.

>>61961342

A small thing, but on the Sapphire Vega Liquid there were no dp/hmdi port/socket covers - the little blue plastic things on the gpu on the right in the pic above.

>>61961361
Gotta shave as much cost as possible on Vega

File: 1470457029422.jpg (41KB, 263x383px) Image search: [Google]

41KB, 263x383px

>>61946049

>>61944046
>redditor target advertising on thechan
wtf mods

File: 1495995943212.jpg (132KB, 1280x720px) Image search: [Google]

132KB, 1280x720px

>>61944046
>broadwell-e
>vega
its time for a housefire

>>61961342

I think I have the Vega Stable with Memory @ 1100Mhz, just doing firestrike now and comparatives for two games.

>>61945572
synthetic benchmarks, the true reality is it wont be anywhere near in games

File: 3DMark-Firestrike-1stRun-VegaOCMem1100-OCcpu-win7Driver-win8-1---18-8-17 - Copy.jpg (1MB, 4480x1440px) Image search: [Google]

1MB, 4480x1440px

>>61965099

>>61965266
Do you have any core clock increases added, or is this result just HMB2 overclock and memory overclock?

>>61965315

This includes the CPU @ 4ghz, System Ram @ 3000Mhz, Vega HBM2 @ 1100Mhz, Vega GPU Clock is stock.

>>61965667
You have the power slider in Waterman at +50% also, right?

>>61965700

Yep

>>61965733

I think it's possible to push the CPU clock to maybe 4.2Ghz, since the vcore is only @ 1.25v from 1.15. I also undervolted the Vega GPU by 20mV, but still experimenting.

>>61965733
Interesting. I've seen graphics scores in FS as high as 25k with the settings you are running, don't know what's up with that.

>>61965908

Are they running win10? Since I'm on 8.1 using the win 7 driver...

File: Qk2WB2M (1).jpg (167KB, 949x1866px) Image search: [Google]

167KB, 949x1866px

>>61965929
Have no idea, there is a guy who has some special way of searching for Vega results on 3DMark and he blurs stuff out of the images he posts to protect his methods I think. Pic related is all I have.

File: CorsairLINK4-ax1500i-power--19-8-17 - Copy.jpg (428KB, 2213x1151px) Image search: [Google]

428KB, 2213x1151px

>>61966008

Thanks for the info.

I have Corsair Link 4 installed which show power usage - even by pcie slot and rail I think.

This might help to see if wattman +50% power level is doing anything.