AMD VEGA confirmed for 2017 H1

[SKILL]

Nothing theoretical about it, they are NVidia's own figures.

If anyone knows how to measure performance of their cards it will be NVidia.

Ok, to spell it out further:

The number is purely <clock speed> * <cuda cores> * 2 = tflops.

Memory bus, bandwidth, latency, none of that is even considered, nor is it a benchmark result (although I'm sure it's quite possible to write an arbitrary benchmark to return this number), they've just done a very simple calculation to provide the theoretical maximum flops. Nothing more, nothing less.

Using that number to then argue that GDDR5X is superior to HBM2 is like saying the Sun is colder than the Earth because Saturn has a ring system. It's completely nonsensical.

 

[TheRealDeal]

FuryX is half the speed (or less!) of a TXP dude....

Replying before reading. You do know the difference is in the manufacturing process. Nvidia do have a lead though.

 

[TheRealDeal]

Pretty sure Vega will have higher performance/watt than Fiji so who cares. Titan X> FuryX in performance/watt but at a cost everyone knows this.

Titan X is impressive but it's also for those that like to pay a lot. I remember buying a 1900xtx for £350 not so long ago which was the same kind of card. It's to much for my needs these days and i won't be swayed to pay from either company. If more people had my thoughts perhaps we would have more in the bank.

 

[JediFragger]

Best to read the thread before posting. Griffildur implied AMD cards required 'small nuclear generators' to function. I corrected him by demonstrating that AMD's previous flagship (FuryX) had vey similar power consumption to the 980ti/TitanX/TitanXP.

Literally no one mentioned performance, as even a small child has the intelligence to know that the 980ti, TitanX, 1070, 1080, TitanXP all obliterate the 4GB FuryX in 99% of all games and benchmarks.

As I recall, weren't some 390X's absolutely abysmal in this regard? Not to mention the super-hungry 295x2! But we're not talking mGPU's I know.

Probably where he got the impression from...

 

[Kaapstad]

Ok, to spell it out further:

The number is purely <clock speed> * <cuda cores> * 2 = tflops.

Memory bus, bandwidth, latency, none of that is even considered, nor is it a benchmark result (although I'm sure it's quite possible to write an arbitrary benchmark to return this number), they've just done a very simple calculation to provide the theoretical maximum flops. Nothing more, nothing less.

Using that number to then argue that GDDR5X is superior to HBM2 is like saying the Sun is colder than the Earth because Saturn has a ring system. It's completely nonsensical.

It does not matter how it is done, the bottom line is NVidia are saying the non HBM2 card is the faster.

It is often rammed down peoples throats on these forums that HBM(2) will drastically increase the performance of a GPU and here we have NVidia saying this is not the case.

 

[Dave2150]

It does not matter how it is done, the bottom line is NVidia are saying the non HBM2 card is the faster.

It is often rammed down peoples throats on these forums that HBM(2) will drastically increase the performance of a GPU and here we have NVidia saying this is not the case.

NVIDIA use HBM2 on their fastest ever GPU, the GP100. They wouldn't do that if GDDR5X was faster. HBM2 is very expensive, meaning they'll only use it if it affords a tangible benefit. A little thought goes a long way.

 

[Kaapstad]

NVIDIA use HBM2 on their fastest ever GPU, the GP100. They wouldn't do that if GDDR5X was faster. HBM2 is very expensive, meaning they'll only use it if it affords a tangible benefit. A little thought goes a long way.

But that is just it,

NVidia are saying that the Pascal Titan is the faster card based on FP32.

Professional cards like the GP100 don't need to be the fastest, they just need to do what it says on the box without fail.

 

[maonayze]

It works both ways and anything to do with NVidia is seen as bad by some users but those who favour a particular brand only ever see negatives pertaining to the brand they prefer and hence post like you have just done. Those who don't really care see good and bad in both AMD and NVidia and not fussed what they have in their machine.

Well Yes, I agree it does work both ways, however i do not see much evidence of AMD users specifically going into Nvidia threads and launching a hail of negativity at anything Nvidia do just because it's Nvidia. It is weighted by quite a margin IMHO, logically and probably due to the marketshare percentages of owners of the cards.

Not sure that part about only ever seeing negatives pertaining to the brand they prefer.....I mean shouldn't that be the other way around? Why would you only see negatives to the brand that you prefer

I do not go into any thread and throw negativity around just because I favour one card over another, so I am not sure what you mean about "Hence post like you have just done".

As for people who are not fussed what they have in their machine and don't really care....well I guess you must be talking about someone else and not yourself, as all of this "Swings both Ways" talk is so easily seen through, it's transparent.
And before you scream that you have had cards of both manufacturers......well surprise, surprise your not the only one. Many of us do and have done over the years, but we don't put up a neon sign and broadcast it and then use it to seemingly sit on a very rickety fence when it suits us.

 

[maonayze]

I've found it to be quite the opposite overall. People here seem to always hype up whatever AMD have coming to extraordinary levels, while anything Nvidia potentially have coming up is downplayed.


Sure. But if you ignore that AMD users here, or AMD 'fans', dont jump on Nvidia for the most miniscule perceived slight all the time as well, then you're just as one-sighted as anybody.

The way this whole Nvidia vs AMD fanboy brigade plays out on this site is hilarious. AMD fans make up a significant percent of the posting in these types of things, yet still act like they are some persecuted minority.

Being honest, I think it's more hope than hype. We hope that AMD can pull themselves out of the pit they have dug and compete more in each sector as it will be better for ALL concerned. We all know this to be true regardless of which side you are on or prefer.

Hey come on, when the power thru the PCie slot debacle came to light, Nvidia users on these forums ran rampant ****ging off the 480's which were sorted within a few days by a driver update. Compare this to the EVGA VRMGate where lots of people on here were in total downplay mode even though it wasnt strictly an Nvidia issue and a board partner issue.

If one side does it then the other is going to retaliate the first chance they get....human nature after all. If no-one did it then we would all have a quieter life but surely much less excitement.

If all the talk and evidence is right about market share (and it is..) then how come there are a significant percent of AMD fans "posting in these types of things, yet still act like they are some persecuted minority"

I thought 70/30 was a minority the last time I looked. I don't see it as being part of a persecuted minority, I see it as loving the underdog.

 

[Seanspeed]

If one side does it then the other is going to retaliate the first chance they get....human nature after all. If no-one did it then we would all have a quieter life but surely much less excitement.

Something is wrong with you if you think GPU fanboy wars are 'exciting'. I'd pay money to see this place rid of all the fanboys so those of us actually interested in hardware discussion can have a place to do so without agendas influencing half the comments.

If all the talk and evidence is right about market share (and it is..) then how come there are a significant percent of AMD fans "posting in these types of things, yet still act like they are some persecuted minority"

I thought 70/30 was a minority the last time I looked. I don't see it as being part of a persecuted minority, I see it as loving the underdog.

You're assuming this place is 100% representative of the average PC gamer?

 

[Geck0]

Being honest, I think it's more hope than hype. We hope that AMD can pull themselves out of the pit they have dug and compete more in each sector as it will be better for ALL concerned. We all know this to be true regardless of which side you are on or prefer.

Hey come on, when the power thru the PCie slot debacle came to light, Nvidia users on these forums ran rampant ****ging off the 480's which were sorted within a few days by a driver update. Compare this to the EVGA VRMGate where lots of people on here were in total downplay mode even though it wasnt strictly an Nvidia issue and a board partner issue.

If one side does it then the other is going to retaliate the first chance they get....human nature after all. If no-one did it then we would all have a quieter life but surely much less excitement.

If all the talk and evidence is right about market share (and it is..) then how come there are a significant percent of AMD fans "posting in these types of things, yet still act like they are some persecuted minority"

I thought 70/30 was a minority the last time I looked. I don't see it as being part of a persecuted minority, I see it as loving the underdog.

Personally I wouldn't have a clue what goes on in the nVidia threads because I don't read them because they don't interest me. I'm AMD biased and freely admit it, but try to keep my comments in the threads for the audience they are intended for - that's the purpose of the forum, isn't it? My reasons are to me valid and I won't rake over that here again.

There are people here and all over the net that thread**** AMD subjects that categorically state that they will never buy the product, these people fall into two categories as far as I am concerned;

a) Have a financial or ego driven directive to attempt to disrupt

b) Are just sociopaths (trolls)

Then you have the second type that say or do buy both brands then tend to pan AMD whilst pre-qualifying their "impartiality" & talking up their "expert assessment" for the "good of all". These are the types you should be red (green) flagging most of all as they more than likely have youtube presences and no doubt receive product and or cash for comment. These are the shills.

Just my .02

 

[maonayze]

Something is wrong with you if you think GPU fanboy wars are 'exciting'. I'd pay money to see this place rid of all the fanboys so those of us actually interested in hardware discussion can have a place to do so without agendas influencing half the comments.


You're assuming this place is 100% representative of the average PC gamer?

I actually didnt mean that it was exciting to see people argue etc... more like it would be so quiet that it would be boring if no-one posted at all.

I, like you much prefer a positive forum where people help each other more often than not. Unfortunately there are some who only post negativity and nothing much more than that, only because they can. my apologies if you took that bit the wrong way.

This place is representative of PC Gamers, but obviously not 100% and not "average" by any means. However, it isnt full of people who can all afford the top end enthusiast stuff either. I think there is a fair mix though.

 

[SKILL]

It does not matter how it is done, the bottom line is NVidia are saying the non HBM2 card is the faster.

It is often rammed down peoples throats on these forums that HBM(2) will drastically increase the performance of a GPU and here we have NVidia saying this is not the case.

Except it does matter. But I give up, you're clearly either unwilling or unable to comprehend that the TFLOPS figure has nothing at all to do with the memory bus, or even real world performance, so I'm out.

 

[Gregster]

Well Yes, I agree it does work both ways, however i do not see much evidence of AMD users specifically going into Nvidia threads and launching a hail of negativity at anything Nvidia do just because it's Nvidia. It is weighted by quite a margin IMHO, logically and probably due to the marketshare percentages of owners of the cards.

Not sure that part about only ever seeing negatives pertaining to the brand they prefer.....I mean shouldn't that be the other way around? Why would you only see negatives to the brand that you prefer

I do not go into any thread and throw negativity around just because I favour one card over another, so I am not sure what you mean about "Hence post like you have just done".

As for people who are not fussed what they have in their machine and don't really care....well I guess you must be talking about someone else and not yourself, as all of this "Swings both Ways" talk is so easily seen through, it's transparent.
And before you scream that you have had cards of both manufacturers......well surprise, surprise your not the only one. Many of us do and have done over the years, but we don't put up a neon sign and broadcast it and then use it to seemingly sit on a very rickety fence when it suits us.

My point was people who are massively biased really feel the need to put EVERYONE else in a user group. You are doing it right now and then going on a massive hyperbole with comments such as "Before you scream!"

For the record, I am hoping Vega is a stroke of genius and regardless of cost, if it delivers, I will get it (so long as not silly cost of course) but I am sure you will look past that and only see an NVidia user posting in an AMD thread and that is forbidden in your eyes or you just accuse everyone of being negative towards your precious!

Don't bother replying, as I am done with your silly comments.

 

[Bonjour]

I'm not massively biased.

I hate you all equally

 

[Stanners]

**Let's keep this on topic, Vega news. Anymore bickering shall be dealt with**

 

[eddyr]

I think the smaller Vega die is probably going to be similar in shader count to the the GPU in the XBox Scorpio GPU,just like the shader count of Polaris 10 and the PS4 PRO GPU are the same,and the noise is it has 4096 shaders like Fiji does. Now add two generations of uarch improvements,two generations of memory compression tech,more VRAM and hopefully more ROPs and a higher clockspeed,you should be looking at a sub 400MM2 die,with probably great than GTX1080 performance. I kind of expect GTX1080 beating performance,but not as fast as GTX1080TI for the smaller Vega chip.

Sure they may not have necessarily gone for the largest die, but it is very likely big Vega's die will be at least as big as Titan PX. It may not mean having an extra year they will definitely beat Titan PX but my guess is they can and will. Others are being cautious purely in my opinion because they do not want to get their hopes up again and be burned, also you have others who cannot see past their green tinted shades

I agree. If we assume that for small Vega it will be along the lines of; worst case scenario coming in just under GP104 performance; best case scenario slightly better than GP104 performance. Considering that as a baseline, big Vega should easily offer performance around GP102 or above. Any weaker (vs that hypothetical baseline) and there would be little point to developing 2 separate dies.

 

[AllBodies]

I agree. If we assume that for small Vega it will be along the lines of; worst case scenario coming in just under GP104 performance; best case scenario slightly better than GP104 performance. Considering that as a baseline, big Vega should easily offer performance around GP102 or above. Any weaker (vs that hypothetical baseline) and there would be little point to developing 2 separate dies.

And just a more simplistic way of looking at it:

With 1 die, there'd be a 4096 core die and a 2304 core die (P10, albeit a different arch).

That's a very large gap within just their own lineup. So surely they'd want to have something like a 2560, 2880, or 3072 die to plug the gap.

Also no doubt large Vega will have a ~3584 core cut down chip for yields.

 

[Dave2150]

For the record, I am hoping Vega is a stroke of genius and regardless of cost, if it delivers, I will get it (so long as not silly cost of course)

Joining the Vega club, right on

 

[CAT-THE-FIFTH]

Considering the XBox Scorpio uses GDDR5 or GDDR5X apparently,I think the smaller Vega chip will use GDDR5 or GDDR5X too.