I'm not sure what "contact factory" means (other than the obvious, "if you want to buy some, call us to ask for details"), but it looks like the 10 Gbps parts that the GTX 1080 will use are ahead of the 11 Gbps and 12 Gbps parts that are still only sampling. They're probably just different bins of chips from the same wafer. It's something less than "production" listed on several GDDR5 parts, though.
Perhaps yields improved faster than Micron expected. Or perhaps Nvidia simply offered to pay Micron more to get them to start mass production sooner. For example, you don't normally want to start mass production if only half of the chips work, but you will if a customer offers to pay you double what you were expecting for each chip. That's just my speculation, though.
It's probable that the GTX 1080 gets nearly all of the world's GDDR5X production in the near future. This is kind of like how the Radeon HD 4870 got nearly all of the early GDDR5 production. Eventually, there will be other cards that use it, of course.
Now, you can't start mass production and have cards that use it show up at retail a month later. SDRAM is probably simpler than things like CPUs or GPUs, but for the latter, it can easily be, you start a wafer today and finish it two months later. And then from "chips exist at the fab" to "chips are part of completed video cards on shelves" takes a few more months. Nvidia will undoubtedly try to hurry production along, but there's really only so much that you can do.
It will nominally be launched on May 27, but requires memory that just went into mass production either today or very, very recently. So most of the thread is about when the cards will actually be available to buy and get shipped immediately, as opposed to pre-order and who knows when you'll get it.
It will nominally be launched on May 27, but requires memory that just went into mass production either today or very, very recently. So most of the thread is about when the cards will actually be available to buy and get shipped immediately, as opposed to pre-order and who knows when you'll get it.
ah I see.
well I would assume before the year is over. I personally would love it if its an option mid July when I will likely be getting a new machine but we shall see
Please do not respond to me, even if I ask you a question, its rhetorical.
It will nominally be launched on May 27, but requires memory that just went into mass production either today or very, very recently. So most of the thread is about when the cards will actually be available to buy and get shipped immediately, as opposed to pre-order and who knows when you'll get it.
ah I see.
well I would assume before the year is over. I personally would love it if its an option mid July when I will likely be getting a new machine but we shall see
Before the end of the year is very probable, but I'd bet against the GTX 1080 being easy to find at MSRP in mid-July. The GTX 1070 and Polaris 10 both have a decent chance at being readily available in mid July, though.
It will nominally be launched on May 27, but requires memory that just went into mass production either today or very, very recently. So most of the thread is about when the cards will actually be available to buy and get shipped immediately, as opposed to pre-order and who knows when you'll get it.
ah I see.
well I would assume before the year is over. I personally would love it if its an option mid July when I will likely be getting a new machine but we shall see
Before the end of the year is very probable, but I'd bet against the GTX 1080 being easy to find at MSRP in mid-July. The GTX 1070 and Polaris 10 both have a decent chance at being readily available in mid July, though.
and this (as a general rule) is why I decided to wait to get a new machine until the Rift is on its way to the house.
Please do not respond to me, even if I ask you a question, its rhetorical.
So, not to rub in it in quiz, but it appears I may have been right ;-).
In all seriousness though, I did post a link from February from Micron that said their early samples were producing very good clock rates which was a good indicator that they were already having excellent yields on their production samples. That was 3 months ago, so it's not unreasonable to suggest that it going into mass production today as being unrealistic.
That being said, my main argument up to this point is it would have been extremely stupid for NVidia to "hard" launch a card knowing full well they would have supply issues 4-5months after release, that's just not good business practice. It was/is much more likely that they were in communications with Micron and knew full well when they would reasonably expect to have large quantities of gddr5x in their hands for production.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
So, not to rub in it in quiz, but it appears I may have been right ;-).
In all seriousness though, I did post a link from February from Micron that said their early samples were producing very good clock rates which was a good indicator that they were already having excellent yields on their production samples. That was 3 months ago, so it's not unreasonable to suggest that it going into mass production today as being unrealistic.
That being said, my main argument up to this point is it would have been extremely stupid for NVidia to "hard" launch a card knowing full well they would have supply issues 4-5months after release, that's just not good business practice. It was/is much more likely that they were in communications with Micron and knew full well when they would reasonably expect to have large quantities of gddr5x in their hands for production.
that possible story reminds me of Oculus and pre-orders.
but to be fair Oculus is not Nvidia and hasnt been doing basically the same pattern every 18 and 36 months nearly like clockwork for more than a decade
Please do not respond to me, even if I ask you a question, its rhetorical.
So, not to rub in it in quiz, but it appears I may have been right ;-).
Honestly, I hope you are right, and that the rest of us skeptics are wrong.
I still remain pessimistic about it. Yes, it's possible they were in communication with Micron and had this whole thing planned down to the day of May 27. But I still think the more likely story is that it's a Wall Street tactic aimed at bumping share prices, and that while cards may indeed be marginally available on that date, it will be a slow trickle of supply until GDDR5X catches up.
But I really do hope the cards come out, and that they are awesome, and that Polaris ships as well and is awesome too - because that raises the bar for hardware, and that's a good thing.
From the article, so you can see how press can get spun on Wall Street:
Patrick Moorhead, analyst at Moor Insights & Strategy, said in an email, “Nvidia again crushed it this quarter and we’re finally seeing the revenue and profit dollars come from their investments in deep neural networks for cognitive computing and artificial intelligence. These cards for the datacenter just started shipping and I believe we will see even more growth in future quarters."
Wall Street trades more in expectations and fairy tales than in anything grounded in reality, and JSH (actually Silicon Valley in general) definitely knows how to play to that crowd. Patrick is referring to GP100, which isn't scheduled for broad release until 2017. But because nVidia could say this:
NVIDIA server partners have revealed the first actual Tesla P100 graphics cards that will be shipping this year.
That translates into Wall Street speak as "Now".
Pretty similar to what they have said for the 1080. I guess May 27 will be the tell all.
Well, NVidia itself said that first low volume shipments (to select Universities and such) of P100 will start in June. Other point is availability in Q12017 and some undisclosed number will be shipped to select customers in between.
From the article, so you can see how press can get spun on Wall Street:
Patrick Moorhead, analyst at Moor Insights & Strategy, said in an email, “Nvidia again crushed it this quarter and we’re finally seeing the revenue and profit dollars come from their investments in deep neural networks for cognitive computing and artificial intelligence. These cards for the datacenter just started shipping and I believe we will see even more growth in future quarters."
Wall Street trades more in expectations and fairy tales than in anything grounded in reality, and JSH (actually Silicon Valley in general) definitely knows how to play to that crowd. Patrick is referring to GP100, which isn't scheduled for broad release until 2017. But because nVidia could say this:
NVIDIA server partners have revealed the first actual Tesla P100 graphics cards that will be shipping this year.
That translates into Wall Street speak as "Now".
Pretty similar to what they have said for the 1080. I guess May 27 will be the tell all.
when was the last time you have ever seen Ninvidia do something as you are suggesting here? I am just curious, the company has been around for a long time and they tend to do things like clockwork each iteration.
Please do not respond to me, even if I ask you a question, its rhetorical.
Quizzical said:
Perhaps yields improved faster than Micron expected.
..or more likely yields improved faster than You expected.
The joys of your water based theorycrafting.
In February, Micron said that they expected to start volume production of GDDR5X this summer. Now they said they've already started and it's clearly not summer. Clearly, something went better than they were willing to publicly project as of February. If in February, they knew that they would start volume production in May, they'd have said they'd start it in May.
As was said earlier, Nvidia was presumably in contact with Micron and knew of--and possibly influenced--Micron entering volume production of GDDR5X sooner. This still doesn't make a hard launch of the GTX 1080 practical in May, but it might mean widespread availability of GTX 1080s in August rather than November.
From the article, so you can see how press can get spun on Wall Street:
Patrick Moorhead, analyst at Moor Insights & Strategy, said in an email, “Nvidia again crushed it this quarter and we’re finally seeing the revenue and profit dollars come from their investments in deep neural networks for cognitive computing and artificial intelligence. These cards for the datacenter just started shipping and I believe we will see even more growth in future quarters."
Wall Street trades more in expectations and fairy tales than in anything grounded in reality, and JSH (actually Silicon Valley in general) definitely knows how to play to that crowd. Patrick is referring to GP100, which isn't scheduled for broad release until 2017. But because nVidia could say this:
NVIDIA server partners have revealed the first actual Tesla P100 graphics cards that will be shipping this year.
That translates into Wall Street speak as "Now".
Pretty similar to what they have said for the 1080. I guess May 27 will be the tell all.
when was the last time you have ever seen Ninvidia do something as you are suggesting here? I am just curious, the company has been around for a long time and they tend to do things like clockwork each iteration.
That's a product that won't even enter closed testing trials until 2017. But it got a lot of investor attention. Autonomous cars, AI, cloud, ... lots of buzz words in there.
2015 saw a large push for Tegra and ARM processors, Shield mobile and mini-console, Android, and cloud rendering (remember GRID?).
See a lot of buzz words in there - that's partially to try to expand into new markets, but it's also because it gets the attention of Wall Street - even if the product itself is a failure, if it serves to boost the stock price enough, it was a success.
From the article, so you can see how press can get spun on Wall Street:
Patrick Moorhead, analyst at Moor Insights & Strategy, said in an email, “Nvidia again crushed it this quarter and we’re finally seeing the revenue and profit dollars come from their investments in deep neural networks for cognitive computing and artificial intelligence. These cards for the datacenter just started shipping and I believe we will see even more growth in future quarters."
Wall Street trades more in expectations and fairy tales than in anything grounded in reality, and JSH (actually Silicon Valley in general) definitely knows how to play to that crowd. Patrick is referring to GP100, which isn't scheduled for broad release until 2017. But because nVidia could say this:
NVIDIA server partners have revealed the first actual Tesla P100 graphics cards that will be shipping this year.
That translates into Wall Street speak as "Now".
Pretty similar to what they have said for the 1080. I guess May 27 will be the tell all.
when was the last time you have ever seen Ninvidia do something as you are suggesting here? I am just curious, the company has been around for a long time and they tend to do things like clockwork each iteration.
That's a product that won't even enter closed testing trials until 2017. But it got a lot of investor attention. Autonomous cars, AI, cloud, ... lots of buzz words in there.
2015 saw a large push for Tegra and ARM processors, Shield mobile and mini-console, Android, and cloud rendering (remember GRID?).
See a lot of buzz words in there - that's partially to try to expand into new markets, but it's also because it gets the attention of Wall Street - even if the product itself is a failure, if it serves to boost the stock price enough, it was a success.
what I meant to ask is in the video card line.
So basically, if this happens will it be the first time in the history of Ninvida cards
Please do not respond to me, even if I ask you a question, its rhetorical.
This is the scenario I think will play out with the 1080, just swap TSMC for Micron.
so basically as I see it setting aside news being fed to us on Wallstreets behaf, its unlikely to be any different. You know...last min 'oh my god at the last second we made the home run like we did in the last 30 superbowls just in time!' kinda play...ha
Please do not respond to me, even if I ask you a question, its rhetorical.
Tech Report says that Nvidia didn't give them a review sample of a GTX 1080. That's one of the indispensable video cards review sites, too, as they pioneered measuring frame pacing rather than just average frame rates. Sometimes review samples don't get distributed that widely for a minor launch (e.g., exactly the same CPU as before, except clocked 100 MHz higher), but the GTX 1080 is not a minor launch.
This is the scenario I think will play out with the 1080, just swap TSMC for Micron.
so basically as I see it setting aside news being fed to us on Wallstreets behaf, its unlikely to be any different. You know...last min 'oh my god at the last second we made the home run like we did in the last 30 superbowls just in time!' kinda play...ha
You totally lost me there big guy.
Most companies will announce when they have something concrete to announce - the card is ready and is shipping today/tomorrow/this week. That's a typical announcement. One we've seen both AMD and nVidia make a lot of these types of announcements over the years.
When a company makes a huge announcement "We have this technology! It's great! ....... (it's shipping next year, we hope)"... or, cleverly stages a product announcement for a product that isn't quite ready ("It's shipping in a few weeks!") just before the competition makes a major announcement or product release - that's clearly a statement just for the fanboys and Wall Street.
AMD has done it as well - Northern Islands Cayman release comes to mind. It was announced in around October 14, same time as the Barts cards (which shipped 2 weeks later) and were announced to ship in mid-November (several weeks after the announcement). It was ultimately pushed back until mid-late December for availability. And the reason that AMD jumped the gun - they wanted to beat the GTX 580 announcement that occurred on Oct 31 (and those cards started shipping the following week). That sounds also like a familiar situation.... if only I could recall where I have recently heard a tale that starts out so similar to this one....
This is the scenario I think will play out with the 1080, just swap TSMC for Micron.
so basically as I see it setting aside news being fed to us on Wallstreets behaf, its unlikely to be any different. You know...last min 'oh my god at the last second we made the home run like we did in the last 30 superbowls just in time!' kinda play...ha
You totally lost me there big guy.
Most companies will announce when they have something concrete to announce - the card is ready and is shipping today/tomorrow/this week. That's a typical announcement. One we've seen both AMD and nVidia make a lot of these types of announcements over the years.
When a company makes a huge announcement "We have this technology! It's great! ....... (it's shipping next year, we hope)"... or, cleverly stages a product announcement for a product that isn't quite ready ("It's shipping in a few weeks!") just before the competition makes a major announcement or product release - that's clearly a statement just for the fanboys and Wall Street.
AMD has done it as well - Northern Islands Cayman release comes to mind. It was announced in around October 14, same time as the Barts cards (which shipped 2 weeks later) and were announced to ship in mid-November (several weeks after the announcement). It was ultimately pushed back until mid-late December for availability. And the reason that AMD jumped the gun - they wanted to beat the GTX 580 announcement that occurred on Oct 31 (and those cards started shipping the following week). That sounds also like a familiar situation.... if only I could recall where I have recently heard a tale that starts out so similar to this one....
what I am saying is past pattern suggests everything is going to go fine with this deployement just like it has the last trillion times for video cards.
Please do not respond to me, even if I ask you a question, its rhetorical.
I'd say Fermi wasn't fine, woodscrews were the least of those problems. I'd say the 680 release was a fiasco, you may not remember that because you were already rocking your $1,500 rig at the time. Videos of cards catching on fire during reviews. The PX 2 "Pascal" really being a Fermi chip. The 970 3.5/4G VRAM issue. And that's just in recent years time...
Just because nVidia survived those doesn't mean everything is always tits and gravy. You may not recall all of those issues several months after they get smoothed over, but they are still issues - especially if you are one of the suckers early-adopters wanting one of those cards.
Comments
"My Fantasy is having two men at once...
One Cooking and One Cleaning!"
---------------------------
"A good man can make you feel sexy,
strong and able to take on the whole world...
oh sorry...that's wine...wine does that..."
http://videocardz.com/59839/micron-gddr5x-memory-enters-mass-production
It has enetered mass production....today
https://www.micron.com/products/dram/gddr/gddr5-part-catalog#/
I'm not sure what "contact factory" means (other than the obvious, "if you want to buy some, call us to ask for details"), but it looks like the 10 Gbps parts that the GTX 1080 will use are ahead of the 11 Gbps and 12 Gbps parts that are still only sampling. They're probably just different bins of chips from the same wafer. It's something less than "production" listed on several GDDR5 parts, though.
Perhaps yields improved faster than Micron expected. Or perhaps Nvidia simply offered to pay Micron more to get them to start mass production sooner. For example, you don't normally want to start mass production if only half of the chips work, but you will if a customer offers to pay you double what you were expecting for each chip. That's just my speculation, though.
It's probable that the GTX 1080 gets nearly all of the world's GDDR5X production in the near future. This is kind of like how the Radeon HD 4870 got nearly all of the early GDDR5 production. Eventually, there will be other cards that use it, of course.
Now, you can't start mass production and have cards that use it show up at retail a month later. SDRAM is probably simpler than things like CPUs or GPUs, but for the latter, it can easily be, you start a wafer today and finish it two months later. And then from "chips exist at the fab" to "chips are part of completed video cards on shelves" takes a few more months. Nvidia will undoubtedly try to hurry production along, but there's really only so much that you can do.
May 27 and June 10 will be the release dates of the NVIDIA GeForce GTX 1080 and GTX 1070, respectively
http://www.ecumenicalnews.com/article/nvidia-geforce-gtx-1080-and-1070-release-dates-revealed/44241.htm
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
well I would assume before the year is over.
I personally would love it if its an option mid July when I will likely be getting a new machine but we shall see
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
So, not to rub in it in quiz, but it appears I may have been right ;-).
In all seriousness though, I did post a link from February from Micron that said their early samples were producing very good clock rates which was a good indicator that they were already having excellent yields on their production samples. That was 3 months ago, so it's not unreasonable to suggest that it going into mass production today as being unrealistic.
That being said, my main argument up to this point is it would have been extremely stupid for NVidia to "hard" launch a card knowing full well they would have supply issues 4-5months after release, that's just not good business practice. It was/is much more likely that they were in communications with Micron and knew full well when they would reasonably expect to have large quantities of gddr5x in their hands for production.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
- Friedrich Nietzsche
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
I still remain pessimistic about it. Yes, it's possible they were in communication with Micron and had this whole thing planned down to the day of May 27. But I still think the more likely story is that it's a Wall Street tactic aimed at bumping share prices, and that while cards may indeed be marginally available on that date, it will be a slow trickle of supply until GDDR5X catches up.
But I really do hope the cards come out, and that they are awesome, and that Polaris ships as well and is awesome too - because that raises the bar for hardware, and that's a good thing.
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
http://venturebeat.com/2016/05/12/nvidia-zooms-by-earnings-targets-thanks-to-graphics-chip-demand-from-cars-and-games/
From the article, so you can see how press can get spun on Wall Street:
Wall Street trades more in expectations and fairy tales than in anything grounded in reality, and JSH (actually Silicon Valley in general) definitely knows how to play to that crowd. Patrick is referring to GP100, which isn't scheduled for broad release until 2017. But because nVidia could say this:
That translates into Wall Street speak as "Now".
Pretty similar to what they have said for the 1080. I guess May 27 will be the tell all.
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
As was said earlier, Nvidia was presumably in contact with Micron and knew of--and possibly influenced--Micron entering volume production of GDDR5X sooner. This still doesn't make a hard launch of the GTX 1080 practical in May, but it might mean widespread availability of GTX 1080s in August rather than November.
The most recent:
https://blogs.nvidia.com/blog/2016/01/04/drive-px-ces-recap/
That's a product that won't even enter closed testing trials until 2017. But it got a lot of investor attention. Autonomous cars, AI, cloud, ... lots of buzz words in there.
2015 saw a large push for Tegra and ARM processors, Shield mobile and mini-console, Android, and cloud rendering (remember GRID?).
See a lot of buzz words in there - that's partially to try to expand into new markets, but it's also because it gets the attention of Wall Street - even if the product itself is a failure, if it serves to boost the stock price enough, it was a success.
So basically, if this happens will it be the first time in the history of Ninvida cards
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
https://semiaccurate.com/2012/05/01/why-cant-nvidia-supply-keplergk104gtx680/
This is the scenario I think will play out with the 1080, just swap TSMC for Micron.
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
Most companies will announce when they have something concrete to announce - the card is ready and is shipping today/tomorrow/this week. That's a typical announcement. One we've seen both AMD and nVidia make a lot of these types of announcements over the years.
When a company makes a huge announcement "We have this technology! It's great! ....... (it's shipping next year, we hope)"... or, cleverly stages a product announcement for a product that isn't quite ready ("It's shipping in a few weeks!") just before the competition makes a major announcement or product release - that's clearly a statement just for the fanboys and Wall Street.
AMD has done it as well - Northern Islands Cayman release comes to mind. It was announced in around October 14, same time as the Barts cards (which shipped 2 weeks later) and were announced to ship in mid-November (several weeks after the announcement). It was ultimately pushed back until mid-late December for availability. And the reason that AMD jumped the gun - they wanted to beat the GTX 580 announcement that occurred on Oct 31 (and those cards started shipping the following week). That sounds also like a familiar situation.... if only I could recall where I have recently heard a tale that starts out so similar to this one....
Please do not respond to me, even if I ask you a question, its rhetorical.
Please do not respond to me
Just because nVidia survived those doesn't mean everything is always tits and gravy. You may not recall all of those issues several months after they get smoothed over, but they are still issues - especially if you are one of the suckers early-adopters wanting one of those cards.