rumor has it nVidia will begin mass production of Pascal GPUs before 2nd half of this year

[Notional]

How many times are you going to bring this up? I've completely lost count now. Nvidia released drivers in june that fixed most of the performance issues with the 700 series. Stop trying to cause a shit storm.

 

From November:

http://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_Waterforce/23.html *fixed link*

 

A 290 beating a 780ti in 4k, 290x beating the 780ti in every resolution. Planned obsolescence from NVidia wasn't only in their drivers losing focus (as NVidia's solutions are based more on optimized drivers, than powerful hardware). No the entire 700 series was vram starved, which is more and more obvious, even though the GPU's are still powerful enough to use ultra textures and all the vram.

Maybe NVidia made a mistake, that cost too much performance in a driver, that was later patched, but the result still stands: Now that the 700 series is not in focus, the performance is plummeting compared to AMD. It will be interesting to see how much Maxwell will crash and burn once Pascal releases.

 

no they wont. Nvidia has blocked the last .5GB with drivers. Not all applications are allowed to see it.

Some applications can see it, but cannot access it.

 

yes, this was confirmed quite some time ago, but it was never properly broadcasted. Yes with todays drivers, the 3.5GB shouldnt affect anything, because you litterally have only 3.5GB, the last .5GB is blocked in most games.

 

Non gaming applications can freely use it from what i know

 

So the card literally is only 3.5GB now for games? WOW. It always leaves me dumbfounded when a joke surpasses reality.

[marldorthegreat]

no, it does not

but the same deal goes for AMD too

But amd could have them out before, as there is a rumour they are using two/three different companies. Samsung 14nm, gloflo 14nm, TSCM

[Prysin]

From November:

http://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_Waterforce/23.html​

 

A 290 beating a 780ti in 4k, 290x beating the 780ti in every resolution. Planned obsolescence from NVidia wasn't only in their drivers losing focus (as NVidia's solutions are based more on optimized drivers, than powerful hardware). No the entire 700 series was vram starved, which is more and more obvious, even though the GPU's are still powerful enough to use ultra textures and all the vram.

Maybe NVidia made a mistake, that cost too much performance in a driver, that was later patched, but the result still stands: Now that the 700 series is not in focus, the performance is plummeting compared to AMD. It will be interesting to see how much Maxwell will crash and burn once Pascal releases.

 

 

So the card literally is only 3.5GB now for games? WOW. It always leaves me dumbfounded when a joke surpasses reality.

Nvidia had few options here.

they fucked up, got called out on it and the witch-hunt to find out how badly it affects things started.

To prevent anyone from seeing HOW catastrophic this could be, they litterally blocked access to the .5GB rather quickly. This caused an issue though. Because some games kept insisting on using it before dumping data over to system memory. This is probably the tests that showed the "gpu tanking"... because in effect, you ran out of ALL memory due to this bug.

After that, they blocked visibility to programs that didnt want to play ball.

[eLucid]

From November:

http://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_Waterforce/23.html​

 

A 290 beating a 780ti in 4k, 290x beating the 780ti in every resolution. Planned obsolescence from NVidia wasn't only in their drivers losing focus (as NVidia's solutions are based more on optimized drivers, than powerful hardware). No the entire 700 series was vram starved, which is more and more obvious, even though the GPU's are still powerful enough to use ultra textures and all the vram.

Maybe NVidia made a mistake, that cost too much performance in a driver, that was later patched, but the result still stands: Now that the 700 series is not in focus, the performance is plummeting compared to AMD. It will be interesting to see how much Maxwell will crash and burn once Pascal releases.

 

 

So the card literally is only 3.5GB now for games? WOW. It always leaves me dumbfounded when a joke surpasses reality.

 

Says 404 not found. And yea, planned obsolescence is right... :rolleyes:

 

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/70125-gtx-780-ti-vs-r9-290x-rematch-8.html

 

You should change your member title to AMD4Life or something. Grow up.

[Notional]

Says 404 not found. And yea, planned obsolescence is right... :rolleyes:

 

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/70125-gtx-780-ti-vs-r9-290x-rematch-8.html

 

You should change your member title to AMD4Life or something. Grow up.

 

Ugh sometimes it ads a bunch of crap after .html. Just delete that and it works.

 

Great link you got there: 4 of 10 of the games are GameWorks games (that seems to be activated), they use a STOCK 290x, which we know can't OC or even maintain base speeds.

 

According to your own link, 290x has gained 3% points of performance on the 780ti. From 9% to 6%. So even with all the GameWorks and using a stock cooled 290x, the 780ti is still losing performance to both the 980 and the 290x. What was your point again?

 

Why? Because all empirics supports my claims? But we all know when people run out of arguments, they move on to ad hominem.

[MonkeyFace]

What's the big deal? If Samsung is producing HBM2 also.

 

I know AMD had an exclusivity clause with Hynix for HBM. But I highly doubt they have that same deal with HBM2.

[zMeul]

But amd could have them out before, as there is a rumour they are using two/three different companies. Samsung 14nm, gloflo 14nm, TSCM

they can't assemble video cards without the GPU+VRAM package - make the PCB, yes they can (nVidia too)

outside Samsung neither of the other two companies even hinted at being able to produce HBM(2) DRAM

[zMeul]

"AMD only confirmed to be sourcing HBM2 from Hynix" does not imply "AMD only uses HBM2 from Hynix"

"there's also the possibility that nVidia has priority over HBM2 production from Samsung" is entirely your *cough* biased *cough* speculation...

oh ... so one speculation is OK, the other isn't  <_<

[VagabondWraith]

AMD can optimize their drivers better because they have the same fucking architecture 3 generations in a row. They compile drivers for the HD/200/300 series. It's all the same with worthless improvements and duck taped VRAM that a card can't make use of. If AMD actually released new architectures instead of straight rebrands, let's see them keep up with older architectures. Hint: they wouldn't waste time and resources. nVidia will spend SOME time on older architecture, but their FOCUS is Maxwell.

[marldorthegreat]

they can't assemble video cards without the GPU+VRAM package - make the PCB, yes they can (nVidia too)

outside Samsung neither of the other two companies even hinted at being able to produce HBM(2) DRAM

The high end gpus need the hbm, but amd will probably release the lower end first.

[Kinda Bottlenecked]

AMD can optimize their drivers better because they have the same fucking architecture 3 generations in a row. They compile drivers for the HD/200/300 series. It's all the same with worthless improvements and duck taped VRAM that a card can't make use of. If AMD actually released new architectures instead of straight rebrands, let's see them keep up with older architectures. Hint: they wouldn't waste time and resources. nVidia will spend SOME time on older architecture, but their FOCUS is Maxwell.

 

Are 1.0, 1.1 and 1.2 the same?  :huh:

[Trixanity]

AMD can optimize their drivers better because they have the same fucking architecture 3 generations in a row. They compile drivers for the HD/200/300 series. It's all the same with worthless improvements and duck taped VRAM that a card can't make use of. If AMD actually released new architectures instead of straight rebrands, let's see them keep up with older architectures. Hint: they wouldn't waste time and resources. nVidia will spend SOME time on older architecture, but their FOCUS is Maxwell.

 

It's not like Nvidia throws out an old architecture and make a new one from scratch in every new release. Like AMD, they build upon the old. Being GCN doesn't mean it's the same. It's multiple generations, even if many things are similar - perhaps more similar than Nvidia's older architectures. AMD just don't see the need to use fancy names in their marketing for improved architectures (although perhaps they should since no one seems to understand it). AMD uses generation to differentiate their products like Intel does, Nvidia is the exception.

 

It's costly to make a radical shift in architecture and it cannot be done every year. Just look at Intel. The difference in architecture since Sandy Bridge is not particularly big but they obviously still improve and add things. Just like Nvidia and AMD do.

 

In regards to HBM. It's not like all graphics card from either team will feature HBM exclusively. The majority will still use GDDR5(X), so it's pretty early to start prepping for the apocalypse. Even if HBM were to be late, it's not like they can't release products without it.

[zMeul]

The high end gpus need the hbm, but amd will probably release the lower end first.

that would make no sense whatsoever - why would you buy one knowing the higher ones are around the corner?

[DildorTheDecent]

ew Hynix. 

[Trixanity]

that would make no sense whatsoever - why would you buy one knowing the higher ones are around the corner?

Price. Different needs. Plenty reasons. That's like saying "why buy a laptop with a Core-M SKU when you can get a beastly Core i7 45W TDP gaming laptop". Obvious reasons.

[zMeul]

Price. Different needs. Plenty reasons. That's like saying "why buy a laptop with a Core-M SKU when you can get a beastly Core i7 45W TDP gaming laptop". Obvious reasons.

that's beyond the point of our discussion - GPUs with HBM2

 

if you are looking for cheap, you can already buy cheap Radeons

[-BirdiE-]

"AMD only confirmed to be sourcing HBM2 from Hynix" does not imply "AMD only uses HBM2 from Hynix"

"there's also the possibility that nVidia has priority over HBM2 production from Samsung" is entirely your *cough* biased *cough* speculation...

 

In fact, Nvidia and Samsung have a much more complicated relationship (reason: Nvidia suing Samsung which resulted in Nvidia violating Samsung's patents) than AMD and Samsung (reason: Samsung now manufacturing chips for AMD).

 

Either way, the scope of the Memory market far exceeds just GPUs. The battle between Samsung and Hynik will be about who can quickly access all markets that can use HBM. While Micron is on the side trying to make HMC.

 

If they confirmed that they'd be sourcing from Hinyx, then why wouldn't they also confirm they'd be sourcing from Samsung if they were sourcing from both?

 

Obviously you can't know for sure either way, but it's really all we have to go on. This is all speculation.

 

If all you want is cold, hard truth... then this is not the place for you. Wait until it is released and you'll know.

[dragosudeki]

oh ... so one speculation is OK, the other isn't  <_<

You're right. Sorry, I didn't realize bias is more credible than the facts surrounding the relationships between the companies.  <_<

[zMeul]

You're right. Sorry, I didn't realize bias is more credible than the facts surrounding the relationships between the companies.  <_<

facts .. what facts?!

is there any proof AMD is actually sourcing HBM2 from Samsung?! no! even a teeny-weeny rumor? no!

[HeadsUpHigh]

that would make no sense whatsoever - why would you buy one knowing the higher ones are around the corner?

Maybe because you can't afford the high-end ones ? All gpus above the gtx 940 have a spot on the market. Also, the improved power efficiency that hbm can offer means better performance in the entire line of products. You can only get a small amount of extra performance by increasing transistor count or power consumption and that also increases cost. Nvidia did exacly the same, improving efficiency with maxwell, and ended up with a very appealing line of products.

[zMeul]

Maybe because you can't afford the high-end ones ? All gpus above the gtx 940 have a spot on the market.

they do not, they are very expensive and you could run on integrated video until you manage to buy something actually worth the cash

 

Luke just did a new video on the matter, up on Vessel

[Linusinium]

For smart buyers, it doesn't matter if Pascal is first to market, value proposition simply won't be there. AMD might go for the jugular because they need market share and offer bang for the buck right away but I would still wait for Pascal too before deciding.

 

First Pascal GPU will be a Titan to sweep up those with legit needs but also those with more money than sense. Then it will be the 980 and 970 equivalents of Pascal. Once the 980ti Pascal drops then it will make sense to pick a card.

[azurezeed]

We'll see, rumors come and go nothing is set in stone

[Briggsy]

so that would mean a late 2016/early 2017 release lineup if this rumour has any truth. seems about on track with previous launches.

[eLucid]

so that would mean a late 2016/early 2017 release lineup if this rumour has any truth. seems about on track with previous launches.

 

God I hope not. People have been waiting patiently now for a long time. If amd launches polaris first - I think you'll see a huge migration of die hard nvidia fans making the switch.