NVIDIA Pascal officially called PK100 and PK104 w/ HBM2

[sof006]

All I can say is "Yaaaaaaaaay!"

 

Omg I can't wait, will probably buy one of these upon release GTX-X70 or GTX-X80 (X is whatever number nvidia sticks at the front, probably GTX 1070 or 1080 lol)

[Blackie Sheen]

PK... why not GP I wonder...

 

GF = Geforce Fermi

GK = Geforce Kepler

GM = Geforce Maxwell

PK = Pascal (Kepler?)...

PK...Suban

 

Pascal's Kingdom?

[Khvarrioiren]

Please share some oficial quotes with confirmation of the tech being used in the upcoming GPU, I would love to get a similar time spam like this claim from Jen ... wich is probably a 12 months period from release. We, the public, got confirmation on the HBM in 2015GPUs 4 days ago, in May... or show us the roadmap in 2014 claiming it for 2015.

 

I would appreciate it because I can't find any.

 

 

 

So I talk about a 3 to 4 years period and you talk about something they did last week after months of hiding information of their to-be released GPUs? And OEM wich are the most predictable cards (specially the ones announced, the low tier ones)... the 2015 GPU lineup announced in Q2 2015 lol... no veil there

They haven't announced jack shit until we received confirmation 4 days ago on HBM presence in the new GPUs. All the rest was vague comments and logical assumptions - nothing clear and was in no roadmap as far as we know. On thing is assumption, other thing is the confirmation from a official representative, specially a CEO.

 

AMD had HBM on their roadmaps as early as 2010/2011 if I remember correctly. Nvidia had something about HBM/HMC around the same time too. Going to scour some places and try to find it.

[RagnarokDel]

PK... why not GP I wonder...

 

GF = Geforce Fermi

GK = Geforce Kepler

GM = Geforce Maxwell

PK = Pascal (Kepler?)...

This is what PK stands for

 

 

Marchand stands for AMD

[NateGSR117]

This super cool to hear about. I want more!

[Crazy Ginger]

Sounds like something Nes would say in Sm4sh

[osbios]

it does make sense

 

if LTT is anything to go by (lol) then more RAM= Nvidia is a step in front of AMD, and there is a bigger number there= better

 

Yes, and it does not even matter how well all the parts of the ram are connected to the GPU!

 

Seriously. 32GiB just screams scientific workload. Because the bus connection is not so much of a bottleneck when you just can upload all your data to the GPU at once and don't need to stream it.

And for the same reason this is actually not that interesting for most of us. Because this will be extremely pricy "professional" cards that don't have any real benefits except specific workload.

 

And to the people saying there is no need for more bandwidth... why do you think AMD and Nvidia switching to this kind of technologies then?

[colonel_mortis]

Reminder: Stay on topic. This is a thread about the fact that nVidia's Pascal GPUs' codenames will be PK100 and PK104, and that they will have HBM2. Discussing the other parts of the OP is also acceptable. However, arguing about whether Intel or AMD invented this feature or that, or whether the AMD R9 295X2's cooler was good or not, doesn't count as being on topic.

[s3ns3]

whether the AMD R9 295X2's cooler was good or not, doesn't count as being on topic.

why because we talk about future products based on what we see in current and previous gens and what changes we expect .

 

thats like not talking about haswell/bulldozer in skylake/zen topics.

 

anyway,just my opinion

[TroubleKlef]

Glad I bought a GTX 960. I'm crossing my fingers for cheaper 4K monitors to coincide with Nvidia's (hopefully) beast Pascal cards.

[colonel_mortis]

why because we talk about future products based on what we see in current and previous gens and what changes we expect .

 

thats like not talking about haswell/bulldozer in skylake/zen topics.

 

anyway,just my opinion

There's a difference between discussing another product and how it compares to this new product, and arguing about whether the cooler on that product was too loud or not. If you're discussing another product, but it's still all related to the original topic at hand, it's fine, but the argument that I was referring to was not.

[carolkarine]

But these aren't out yet... So they're still behind.

is pascal out? your point is moot.

[zappian]

Glad I bought a GTX 960. I'm crossing my fingers for cheaper 4K monitors to coincide with Nvidia's (hopefully) beast Pascal cards.

 

Thats a weird thing to be glad about.

Such a underwhelming gpu . Even a lot of nvidia fanatics  were disappointed  by its performance .

GTX 970 is where its at from nvidia , amazing card still.

[mr moose]

Glad I bought a GTX 960. I'm crossing my fingers for cheaper 4K monitors to coincide with Nvidia's (hopefully) beast Pascal cards.

 

I'm not glad at all about any of this,  I am due to upgrade my GPU next year and even if both companies only half come good on their hype I am going to have choice panic.

[TroubleKlef]

Thats a weird thing to be glad about.

Such a underwhelming gpu . Even a lot of nvidia fanatics  were disappointed  by its performance .

GTX 970 is where its at from nvidia , amazing card still.

 

970 is nearly $150 more with only 3.5gb of ram and it's more of a 1440p card. I'd rather cut my losses with a cheap 1080p card and spend the bank with 4K and Pascal.

 

I'm not glad at all about any of this,  I am due to upgrade my GPU next year and even if both companies only half come good on their hype I am going to have choice panic.

 

Hm... hopefully there will be a price war to help you choose but if there isn't then GL lmao.

[Dabombinable]

970 is nearly $150 more with only 3.5gb of ram and it's more of a 1440p card. I'd rather cut my losses with a cheap 1080p card and spend the bank with 4K and Pascal.

That is extremely wrong, it has 4GB of vRAM, its just that the last 512MB isn't much faster than system memory.

 

Hm... hopefully there will be a price war to help you choose but if there isn't then GL lmao.

[PlayModeOn]

Money reasons. AMD is a small company compared to Intel...

They have Global Foundries, AMD just has to design.

[Admire]

Thats a weird thing to be glad about.

Such a underwhelming gpu . Even a lot of nvidia fanatics  were disappointed  by its performance .

GTX 970 is where its at from nvidia , amazing card still.

 

He got a new GPU.....he isn't even allowed to be happy about his purchase? No matter where it lies on the performance to dollar scale......

Do you have to become a total ***** to everybody as soon as there is a discussion with AMD? He said nothing about AMD, nothing about anything, and you just had to stomp on whatever joy he might have had....

 

Stay classy Zappian, Stay classy.

[ivan134]

DX 10.1 adoption was practically non-existent until Nvidia jumped aboard. It's called reading the market momentum. 400 & 500 were not disasters, and I currently run off a 570 in my school build. up until very recently 3GB was more than enough. Suddenly memory management in games went to pot.

 

XDMA XFire still has pretty prolific microstutter issues. Frankly it's not a superior solution right now, and Nvidia understands that. There's also the fact going bridgeless means potentially creating a PCIe bottleneck where there wasn't one before. Using a bridge cable allows mitigation of this. Nvidia pulled ahead on Power Consumption with Kepler, or else IBM wouldn't be using Kepler in their supercomputers. They'd have picked AMD's FirePro chips.

 

Fully programmable shaders, HDMI 2.0, DP 1.2, physics engines, real-time rendering and simulation, GPGPU AI programming, and a whole slew more. Nvidia is still a step ahead, but Nvidia also knows which steps to take when. Some of you may find Nvidia's slowness to move as complacency or arrogance, but I'd pose this question: is it arrogance if you really are the best in the room? It seems most consumers recognize Nvidia's superiority, and I can guarantee you it's not primarily marketing.

The history revision here is amazing.

 

Dx 10.1 adoption was low because of Nvidia, not the other way around. Nvidia had clout with so many devs back then and if Nvidia cards didn't support a feature, they wouldn't bother with it.

 

Did you really just say there was nothing wrong with Fermi? Let me help you jog your memory.

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/30321-nvidia-geforce-gtx-470-review.html

Heat issues were well known, but pay special attention to power consumption in relation to performance.

 

Nvidia pulled ahead in power consumption with Kepler?

 

Yea, I can definitely see how you came to the conclusion that Nvidia has always been one step ahead.

[patrickjp93]

The history revision here is amazing.

Dx 10.1 adoption was low because of Nvidia, not the other way around. Nvidia had clout with so many devs back then and if Nvidia cards didn't support a feature, they wouldn't bother with it.

Did you really just say there was nothing wrong with Fermi? Let me help you jog your memory.

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/30321-nvidia-geforce-gtx-470-review.html

Heat issues were well known, but pay special attention to power consumption in relation to performance.

Nvidia pulled ahead in power consumption with Kepler?

-snips-

Yea, I can definitely see how you came to the conclusion that Nvidia has always been one step ahead.

1) no, ATI and Nvidia were neck and neck for market share.

2) look at the Teslas and Quadros where power consumption actually matters and compare to FirePro cards. There are critical clock speeds at which it takes quadratic or greater increases power relative to linear growth in clock speed/performance. For gaming cards where clock is king, both companies push beyond that critical point. Below it Nvidia won out by a fair margin, a margin made only wider by Maxwell, though of course without any Tesla products it's less noteworthy.

3) read again. I said Fermi wasn't a disaster, and it wasn't. Loud and hot were the flavors du jour from both companies. That's hardly a disaster.

[LukaP]

So, can we just fast forward to 2016?  Like--is there a place I can go to just speed up time, or nah?

Either a black hole, or going really fast. Or just sleep near your mom for a night

[ivan134]

1) no, ATI and Nvidia were neck and neck for market share.

2) look at the Teslas and Quadros where power consumption actually matters and compare to FirePro cards. There are critical clock speeds at which it takes quadratic or greater increases power relative to linear growth in clock speed/performance. For gaming cards where clock is king, both companies push beyond that critical point. Below it Nvidia won out by a fair margin, a margin made only wider by Maxwell, though of course without any Tesla products it's less noteworthy.

3) read again. I said Fermi wasn't a disaster, and it wasn't. Loud and hot were the flavors du jour from both companies. That's hardly a disaster.

1. http://www.bit-tech.net/news/hardware/2009/04/30/nvidia-increases-market-share/1

 

2. You have to be trolling.

 

3. Pay special attention to the 5850 vs the 470

 

 

 

 

 

 

 

The gtx 470 was using 100W more to give as little as 2 more fps in some cases. The gtx 480 was a bit better, but still terrible compared to what AMD was offering. It used 20W less than a 5970 while performing an average of 20 fps worse.

 

At this point, I'm convinced you're a troll so don't bother responding anymore.

[patrickjp93]

Either a black hole, or going really fast. Or just sleep near your mom for a night

 

1. http://www.bit-tech.net/news/hardware/2009/04/30/nvidia-increases-market-share/1

 

2. You have to be trolling.

 

3. Pay special attention to the 5850 vs the 470

 

 

 

-so-

-many-

-snippity-

-snip-

-snips-

 

The gtx 470 was using 100W more to give as little as 2 more fps in some cases. The gtx 480 was a bit better, but still terrible compared to what AMD was offering. It used 20W less than a 5970 while performing an average of 20 fps worse.

 

At this point, I'm convinced you're a troll so don't bother responding anymore.

1) incomplete and blatantly false article that only took into account new sales in the previous year, not actual standing marketshare in a rapidly expanding market

2) Not at all. You can see this behavior yourself modeled in Intel and AMD CPUs as well as in GPUs or the Power 8 processors. Everything I've said is true, even if highly inconvenient for you. External proof of this is reflected in the cost structure for GPU-accelerated supercomputers/datacenters. Electricity is the dominant cost over the long term, not equipment replacement. If AMD had a decent power consumption and/or heat density lead, especially with its theoretical performance lead, then it would have vastly outsold the Nvidia Kepler Teslas purely on merit.

 

3) Your point? Also you are quoting launch benchmarks which did narrow in the performance gap over the course of the product lifetime. 

 

Fermi wasn't a disaster. It wasn't Nvidia's best architecture or time period, but it wasn't a disaster by any stretch.

[s3ns3]

1) no, ATI and Nvidia were neck and neck for market share.

 

now its 24% amd and the rest is took by nvidia

[MEC-777]

Lol @ all the people thinking AMD is going to "disappear" or "go bankrupt" or that they are "so far behind" when it comes to GPUs. Oh ye of ignorance and little knowledge...