Poll: Do you think AMD will be able to compete with Nvidia again during the next few years?

[bru]

So a few pages back when I said that

Well that is one opinion, another would be.

Well AMD have prettier driver interface, a comparable variable refresh rate experience, very similar image quality, but saying all that they are dearer, consume more power and were much latter into the market.

​

You have all jumped to defend Vega, even though you now admit that the default settings are not good. the bottom line is that Vega is power hungry and it was late to the market, or do you want to argue that point as well.

Hold on a bit on the power consumption front. I had until yesterday a GTX1080Ti Xtreme, with no manual overclock just the normal boost and was consuming 70W more than the V64 Nitro+ I have today. Dont take the feeble FE power consumption to make a case for it. Take the factory overclocked versions which some are bordering to the old mighty 295X2 in that front.

You want to argue the power consumption point by using a 1080Ti extreme, vs the card you have after you have tweaked it, how is that a fair comparison.

Point out these loads of people.

They might not be all great overclockers, but, they all can be tweaked to use less power and perform better with a few simple steps because the default settings are rubbish.

 

[melmac]

So a few pages back when I said that

​

You have all jumped to defend Vega, even though you now admit that the default settings are not good. the bottom line is that Vega is power hungry and it was late to the market, or do you want to argue that point as well.



You want to argue the power consumption point by using a 1080Ti extreme, vs the card you have after you have tweaked it, how is that a fair comparison.

I am not arguing over Vega's default settings. I am asking you to point out the loads of people who can't tweak Vega, that's the point I quoted you on. So argue with me based on what I said.

 

[GoogalyMoogaly]

So a few pages back when I said that

​

You have all jumped to defend Vega, even though you now admit that the default settings are not good. the bottom line is that Vega is power hungry and it was late to the market, or do you want to argue that point as well.



You want to argue the power consumption point by using a 1080Ti extreme, vs the card you have after you have tweaked it, how is that a fair comparison.

Well, that depends. The Nvidia 900 series was competing with the AMD 300 series (980Ti vs Fury X), so the Nvidia 10 series would be competing with the AMD 400 series. That means the AMD 500 series and Vega has come out WAY before Nvidia's successor to the 10 series!

 

[tyler_jrb]

No.

 

[bru]

I am not arguing over Vega's default settings. I am asking you to point out the loads of people who can't tweak Vega, that's the point I quoted you on. So argue with me based on what I said.

Really…….

Go and read all of these I'm sure you will find loads amongst the 1.1 Million that cannot.

Cannot tweak vega

 

[bru]

Not sure where some people are getting their figures from but reference Vega 64 uses a lot more power than any of the high end reference NVidia cards like the 1080 Ti FE, Vitan Xp or Titan V.

Non reference cards can not be used for a fair comparison as they are not what either NVidia or AMD think the card should perform like.

Having said that as zophiel said above the cost in using either brand adds up to very little over a year.​

Quite agree, anyone really trying to say that Vega uses less power than Pascal, needs to pass the bong around.

 

[melmac]

Really…….

Go and read all of these I'm sure you will find loads amongst the 1.1 Million that cannot.

Cannot tweak vega

LOL where are the 1.1 million that cannot? the link you posted brings me to google search and the first results are from reddit and then there are videos showing how to tweak Vega. Even those first results show people having great success with reducing the power use of Vega.

Show me those people that "have had not much luck tweaking theirs" with regard to making their Vega more power efficient. Loads of people according to you.

 

[LoadsaMoney]

AMD GPU Rumors: Polaris 30 12nm GPUs in Q4 2018, Navi GPU Based Mid-Range Cards Arrive Before High-End HBM2 Variants

https://wccftech.com/amd-gpu-rumors-12nm-polaris-30-7nm-navi-7nm-vega-20-radeon-rx-gaming/

 

[FoxEye]

AMD GPU Rumors: Polaris 30 12nm GPUs in Q4 2018, Navi GPU Based Mid-Range Cards Arrive Before High-End HBM2 Variants

https://wccftech.com/amd-gpu-rumors-12nm-polaris-30-7nm-navi-7nm-vega-20-radeon-rx-gaming/

Why would anyone want another Polaris generation that's 15% faster only? /boggle

 

[bru]

LOL where are the 1.1 million that cannot? the link you posted brings me to google search and the first results are from reddit and then there are videos showing how to tweak Vega. Even those first results show people having great success with reducing the power use of Vega.

Show me those people that "have had not much luck tweaking theirs" with regard to making their Vega more power efficient. Loads of people according to you.

I can be facetious just as much as you, they are in that 1.1 million google links, if you look through enough of them you will see that I was right, obviously it may take you a while.

 

[Panos]

Why would anyone want another Polaris generation that's 15% faster only? /boggle

Why would anyone want a second Maxwell refresh? (first one was Pascal).

 

[FoxEye]

Why would anyone want a second Maxwell refresh? (first one was Pascal).

If it's only 15% faster - nobody.

People won't pay silly money (£300+) for 15% improvements.

But thankfully whatever nV's newest cards will be, they'll be a lot more of a jump than that. Because Pascal chips were nice and small and power efficient, so at the very least nV can just up the core count. I'm not sure AMD can realistically do the same. So 15% improvement sounds like the max they can improve Polaris 30 by over and above the 580.

 

[Panos]

If it's only 15% faster - nobody.

People won't pay silly money (£300+) for 15% improvements.

But thankfully whatever nV's newest cards will be, they'll be a lot more of a jump than that. Because Pascal chips were nice and small and power efficient, so at the very least nV can just up the core count. I'm not sure AMD can realistically do the same. So 15% improvement sounds like the max they can improve Polaris 30 by over and above the 580.

Pascal is better than Maxwel because we had die shrink. This time there is no new tech no shrinking either.
it will be 6700K -> 7700K transition for more money because of GDDR6.

 

[D.P.]

Pascal is better than Maxwel because we had die shrink. This time there is no new tech no shrinking either.
it will be 6700K -> 7700K transition for more money because of GDDR6.

Pascal also had a completely re-worked scheduler, pixel-level preemption, stereo multi-projection, improved compression and bandwidth saving techniques, multi-path optimization, and an increase in DX12 feature sets such as order independent transparency and conservative rasterization. Internally the CUDA cores went through a lot of optimization that improves throughput. Cache sizes also changed

 

[CAT-THE-FIFTH]

If it's only 15% faster - nobody.

People won't pay silly money (£300+) for 15% improvements.

But thankfully whatever nV's newest cards will be, they'll be a lot more of a jump than that. Because Pascal chips were nice and small and power efficient, so at the very least nV can just up the core count. I'm not sure AMD can realistically do the same. So 15% improvement sounds like the max they can improve Polaris 30 by over and above the 580.

Its from Wccftech - remember the "beast mode" RX480 cards. LMAO.

 

[The Running Man]

Id question the 'again but of the question.

When was the last time that Amd took real market share from Nvidia?

 

[D.P.]

Id question the 'again but of the question.

When was the last time that Amd took real market share from Nvidia?

radeon HD 5000 https://www.techpowerup.com/228095/amd-and-nvidia-add-in-board-gpu-market-share-from-2002-to-q3-2016

 

[melmac]

I can be facetious just as much as you, they are in that 1.1 million google links, if you look through enough of them you will see that I was right, obviously it may take you a while.

Why should I look through all those links looking for proof of your claim. And as I said, I looked through the first page and they all showed people after tweaking Vega to use less power and also guides to tweaking Vega. Where are all these cards that aren't capable of it as per your post, where are these loads of people who have failed to tweak Vega to be more power efficient? Just a reminder of what you said.

you might have tweaked your card to be relatively power efficient but do not try to kid us that they are all capable of it, as loads of people have not had much luck tweaking theirs.

Those are your claims. There were people looking for help and asking for guides as Vega and Wattman were new. Wattman also had a few teething problems but these got sorted. When Vega came out first people were saying that at least miners wouldn't buy them because they used so much power, but, when people found out how easy they could be tweaked, they got very popular. Those guides became very popular, I can link to one or two if you want.

 

[bru]

So you guarantee that every Vega card can be tweaked, regardless of what case it is in and how shabby the rest of the system is. You say there isn't a single case out there in the world, where that isn't possible, out of the several hundred thousands of cards sold world wide That is certainly what you seem to be saying.

 

[Panos]

So you guarantee that every Vega card can be tweaked, regardless of what case it is in and how shabby the rest of the system is. You say there isn't a single case out there in the world, where that isn't possible, out of the several hundred thousands of cards sold world wide That is certainly what you seem to be saying.

Actually given how many Mac Pros Apple has sold, quite a few million Vega 64s and last 3 quarters more than 20 million AMD GPUs in the Mac, given the sales figures published by Apple.