AMD's Vega architecture previewed at ve.ga

[CostcoSamples]

4 hours ago, LAwLz said:

Sorry, didn't feel like reading though over 150 replies.

Seems like you're right. For DX11 they are very even. Both cards win a few games, lose in some, and are even in the rest.

 

 

AdoredTV is one of the biggest AMD fanboys I have ever seen. How anyone trusts anything he says is beyond me. I will gladly look at other reviews showing the same results though.

The 18% improvement HardwareCanucks got sounds way more reasonable than 40%.

 

When one reviewer sees an 18% improvement, and someone else claims 40%, then something fishy is going on (and my guess is that AdoredTV is the one being deliberately misleading).

I find AdoredTV generally presents very thorough analysis.  While a lot of his content is based on AMD, I don't think fanboy is an accurate description.  

 

As I said, the % increase varies depending what part of the game you are using because there is no built in benchmark run.  I did a quick check at other sites and it seems most are around 30%.  HWC and AdoredTV are both outliers compared to the sites below:

 

GamerNexus 30% increase

PC Gamer 30% increase

EuroGamer 29% increase

 

30% increase is a pretty massive increase just from using a different API !!

[NumLock21]

[niofalpha]

I really wish AMD Didn't spend 3/5ths their income on their advertisements. 

[Citadelen]

12 hours ago, lots of unexplainable lag said:

-snip-

Ah okay.

[kiwibacon]

16 hours ago, MadyTehWolfie said:

Hate on SLI is mostly justified but it also depends on the games you play. At most your only going to get 50% increase and that's the best case scenario. Some games become unplayable with SLI....Cough cough dues ex mankind divided. In my experience most games see a Meh improvement with SLI and very few get the full 50% increase. It's always better to just buy a card from next gen then to double up. A trend with Nvidia is that their xx70 cards usually pack the same punch as last year's Titan model. Seen it with the 970 and with the 1070 and Volta will be more of the same if not slightly better.

I Was Playing Dues Ex Other Day Unplayable Without Sli. Getting 30fps with 980ti or 45 with sli. Had to turn gsync on too. Still pretty borderline.l

[Citadelen]

50 minutes ago, niofalpha said:

-snip-

They don't...

[Xsilent(X)]

1 hour ago, niofalpha said:

I really wish AMD Didn't spend 3/5ths their income on their advertisements. 

 

They only advertise on their social media accounts.

[leadeater]

1 hour ago, NumLock21 said:

-snip-

The text layout on that actually makes it horrible to read, well more annoying but still. Where did you get that from btw? All the ones I've seen of that are grey backgrounds and even worse to read.

[MadyTehWolfie]

1 hour ago, kiwibacon said:

I Was Playing Dues Ex Other Day Unplayable Without Sli. Getting 30fps with 980ti or 45 with sli. Had to turn gsync on too. Still pretty borderline.l

Exactly what I was talking about

[MadyTehWolfie]

7 hours ago, App4that said:

So where is the reason to resort to personal attacks, if you have an argument? Seems to me mate you're the one making assumptions based off beliefs. 

 

What you're doing is saying manual transmission cars aren't worth it, because you never learned to drive one. 

Except I have had sli more then three different times from both companies. You fail to grasp this. If anyone is more experienced with sli it is me based on the number sli and crossfire configs I have. I have seen the difference first hand and bench marked in the past to get real world sli numbers. Most of the time the performance in most games wasnt even 50% increase in performance being somewhere around the 30% mark. You also ignore the overwhelming amount of evidence from other sli users and fail to even google dues ex sli issues. You are speaking non sense and your fanboyism is showing.

 

I dont know how you think your in the right when Linus someone who is more experienced then you or me with sli even says that a single more powerful card is better then a sli setup. Please educate yourself a little more before trying to defend something your making yourself look bad.

[NumLock21]

2 hours ago, leadeater said:

The text layout on that actually makes it horrible to read, well more annoying but still. Where did you get that from btw? All the ones I've seen of that are grey backgrounds and even worse to read.

From here

http://videocardz.com/65288/amd-vega-architectural-upgrades-revealed

 

And... this was in the comment section
 

"May the Radeon be with you... lmfao"

[Princess Luna]

http://www.digitaltrends.com/computing/ibm-power9-server-processor-architecture-revealed-hot-chips-28/

[App4that]

1 hour ago, MadyTehWolfie said:

Except I have had sli more then three different times from both companies. You fail to grasp this. If anyone is more experienced with sli it is me based on the number sli and crossfire configs I have. I have seen the difference first hand and bench marked in the past to get real world sli numbers. Most of the time the performance in most games wasnt even 50% increase in performance being somewhere around the 30% mark. You also ignore the overwhelming amount of evidence from other sli users and fail to even google dues ex sli issues. You are speaking non sense and your fanboyism is showing.

 

I dont know how you think your in the right when Linus someone who is more experienced then you or me with sli even says that a single more powerful card is better then a sli setup. Please educate yourself a little more before trying to defend something your making yourself look bad.

And you're proving why you're on this forum...

 

Remember Linus also said RAM speed doesn't matter in gaming, and depending on the game it's been proven to. Linus is a salesman,  not a bad thing just how it is.

 

SLI works. I get way over 50% scaling in most my games. I get over 50% scaling in Deus Ex, the game YOU said had problems. Further evidence the problem is you, not SLI.

[Dabombinable]

15 minutes ago, App4that said:

And you're proving why you're on this forum...

 

Remember Linus also said RAM speed doesn't matter in gaming, and depending on the game it's been proven to. Linus is a salesman,  not a bad thing just how it is.

 

SLI works. I get way over 50% scaling in most my games. I get over 50% scaling in Deus Ex, the game YOU said had problems. Further evidence the problem is you, not SLI.

RAM speed does matter if you have a CPU bottleneck. Just look at how much an i3 6100 improves with fast kits of DDR4 over standard kits. And its also the same for DDR2 with older Core 2 uads (though with that you need a motherboard that isn't finicky AF the way my P5 Deluxe is-you need memory off the ualified vendors list to run 4 sticks of DDR2 1066 or higher regardless of the timings), with DDR2 1066 5-6-6-16 providing a noticeable reduction in stuttering (as well as increased bandwidth over DDR2 800 4-4-4-12, there is also significantly decreased latency).

And if only SLI scaling with Nvidia's current cards was as good as it was with the Voodoo2-you get extremely close to double the frame rates with Voodoo2 in SLI.

[MadyTehWolfie]

20 minutes ago, App4that said:

And you're proving why you're on this forum...

 

Remember Linus also said RAM speed doesn't matter in gaming, and depending on the game it's been proven to. Linus is a salesman,  not a bad thing just how it is.

 

SLI works. I get way over 50% scaling in most my games. I get over 50% scaling in Deus Ex, the game YOU said had problems. Further evidence the problem is you, not SLI.

If he's a salesman then he wouldn't be for buying multiple cards not against it. Me thinks you don't know how math works if you think your getting a more then 50% in most games with SLI. You believing that makes me laugh never in my years of owning two Titan blacks, Titan X, 1080, and the r9 295x2 with a 4970K have I ever gotten more then in the 50% range. Please troll somewhere else and learn basic math and percentages.

[Humbug]

5 hours ago, CostcoSamples said:

I find AdoredTV generally presents very thorough analysis.  While a lot of his content is based on AMD, I don't think fanboy is an accurate description.  

hmmm ya adoredTV may choose to do more AMD coverage, and I think at this point that's what his viewership too wants from him. But he doesn't lie or deliberately mislead which is what matters... He seems to dispense some useful information.

[Dabombinable]

5 minutes ago, MadyTehWolfie said:

If he's a salesman then he wouldn't be for buying multiple cards not against it. Me thinks you don't know how math works if you think your getting a more then 50% in most games with SLI. You believing that makes me laugh never in my years of owning two Titan blacks, Titan X, 1080, and the r9 295x2 with a 4970K have I ever gotten more then in the 50% range. Please troll somewhere else and learn basic math and percentages.

It all depends on the:

• Driver version
• How well the game is optimised by the devs for SLI or Crossfire (the majorityt barely optimise at all
• The motherboard+chipset
• The generation of the PCIe slots, and the number of PCIe lanes if your using Crossfire
• The CPU your using

SLI and Crossfire scaling is very much dependant on the individual person's rig. You obviously are one of those who's only had grief from multi GPU setups.

[Dabombinable]

Turns out that the majority of Voodoo2 SLI benchmarks were done on sub 400MHz CPU, not a 1.2GHz Pentium III. *sigh* So not many people have seen the full potential (though Glide is far better on weak CPU than OpenGL and DirectX ever have been-shifting to glide reduces bottlenecks)

1 minute ago, MadyTehWolfie said:

So you can't even post a link to a site? Lol that requires a quick Google search if you want I'll post SLI benchmarks and we can see if they have your magic 60-80% fps boosts

I'll be blunt-I think we need to see higher bandwidth (and also dedicated) communication between graphics cards in SLI/Crossfire as well as far stronger CPU (which is a long time away)..

[leadeater]

That's enough, has nothing to do with Vega release.

[Dabombinable]

8 minutes ago, leadeater said:

That's enough, has nothing to do with Vega release.

But I had a lengthy wall of text+reliable data that would have ended the discussion 

[leadeater]

2 minutes ago, Dabombinable said:

But I had a lengthy wall of text+reliable data that would have ended the discussion 

As much as I'd actually like to see that, as a crossfire user myself, best put somewhere else.

[Dabombinable]

3 minutes ago, leadeater said:

As much as I'd actually like to see that, as a crossfire user myself, best put somewhere else.

I have now-my status updates (and looking up X model card+SLI/crossfire review is all that's needed)

 

Edit: Well, I moved part of it.

[MadyTehWolfie]

@App4that @Dabombinable

 

After looking at your status update I admit I was wrong and thank you for not only providing proof for your argument but also giving a good read on the subject.

 

Getting back on topic I think AMD shot themselves in the foot. They put themselves in a place where they either have to match performance for less money or have a good bump in performance more then Nvidia's offerings at the same price to win over Nvidia's already established market share of the high end GPU market.

[RagnarokDel]

On 02/01/2017 at 1:00 PM, EOZ said:

I wish all the best to AMD Vega but reality is reality i think GTX 1080Ti (GDDR5X) will easily knockout AMD Vega GPU with (HBM2) Memory.

 

there's no way to tell until benchmarks, it's a new architecture.

[captain cactus]

3 hours ago, MadyTehWolfie said:

@App4that @Dabombinable

 

After looking at your status update I admit I was wrong and thank you for not only providing proof for your argument but also giving a good read on the subject.

 

Getting back on topic I think AMD shot themselves in the foot. They put themselves in a place where they either have to match performance for less money or have a good bump in performance more then Nvidia's offerings at the same price to win over Nvidia's already established market share of the high end GPU market.

The new architecture would allow AMD's GPUs to perform much closer to its TFLOPs number. If we were in a situation where TFLOPs equals a given fps a Fury X would beat a GTX1080 (8.6 TFLOPs vs 8.2 TFLOPs). From what I've read left and right (including in that Reddit thread which has some really genuinely interesting posts by people who know their stuff) the new NCU architecture will do just that. Right now only compute applications benefit from AMD's higher TFLOPs count. AMD wants games to do that to.

 

AMD won't put a poster in an advertisement for Vega that says Poor Volta if they knew Vega would only match or slightly exceed Pascal. They want to turn the tide around where nVidia plays catchup to AMD rather than the opposite.