NVIDIA GTC 2019 Kicks Off Later Today, New GPU Architecture Tease Expected

[medi01]

At what time do we expect stuff mentioned in OP not to be mentioned by the Lisa Su's uncle at GTC?

That was a QC issue

No shit Watson.

there weren't any more failed cards than any other launch

Citation needed.

 

[THU31]

Hopefully they will announce the Titan 7. A proper monster like the first one, not those jokes which are 2% faster than a x080 Ti.

 

[DeathtoGnomes]

Card_Y was obsolete from the time the first benchmarks launched. Way overpriced with little performance benefit over Card_X. Not to mention that it's noisy, power hungry and can't consistently match the product it competes against.

Fixed.

Fixed.

seen this generic canned statement once too often.

 

[Xuper]

from Business point : a Ray-tracing card + GPU card is better than packed Ray-tracing + GPU cores like RTX cards.Hope NV/AMD /Intel follows this strategy.

 

[cucker tarlson]

from Business point : a Ray-tracing card + GPU card is better than packed Ray-tracing + GPU cores like RTX cards.Hope NV/AMD /Intel follows this strategy.

Can you elaborate on that?

 

[jabbadap]

from Business point : a Ray-tracing card + GPU card is better than packed Ray-tracing + GPU cores like RTX cards.Hope NV/AMD /Intel follows this strategy.

Not sure people would buy those though. PhysX cards weren't actually very popular at their prime and I doubt discrete RT cards would age any better than those. Maybe some different chip on the same card or even some MCM RT block would work(i.e. Navi+RT on MCM), but it would still drive price higher.

 

[Vayra86]

Can you elaborate on that?

You get to sell and market two products instead of one. You also get to show improvements on two product stacks instead of one. Its the same idea as with SLI, really. You also present customers with additional options and a degree of freedom: not forcing RTRT but opening it up in a nicer way, without a mandatory price bump.

Not sure people would buy those though. PhysX cards weren't actually very popular at their prime and I doubt discrete RT cards would age any better than those. Maybe some different chip on the same card or even some MCM RT block would work(i.e. Navi+RT on MCM), but it would still drive price higher.

Content sells. PhysX suffered the same problem as RT does today. People forget that in terms of gaming, the actual product is not the GPU, but what it can do.

 

[jabbadap]

Starting now, Anandtech has live blog for those interested.

Spoiler: GTC 2019 Keynote

https://www.ustream.tv/gpu-technology-conference

 

[HwGeek]

So pascal getting DXR next month , Let's see what AMD gonna do with all their compute power on Polaris and Vega vs Pascal.

 

[wolf]

Your comment is cute.

Aww thank you! It's such a mood boost when the younger members compliment me

In actual fact my comment really was cute, having a fun little poke to show another side of the coin. Also amusing is you picking the only part of what I said that you could 'disprove', and in doing so you managed to perhaps inadvertently achieve even more than you intended.

 

[ratirt]

You get to sell and market two products instead of one. You also get to show improvements on two product stacks instead of one. Its the same idea as with SLI, really. You also present customers with additional options and a degree of freedom: not forcing RTRT but opening it up in a nicer way, without a mandatory price bump.

Come to think of it, This isn't a bad idea. If Nvidia does something like that they can only win with this. The problem is how would they do that? SLI isn't so reliable and games developers don't want to spend resources to support it. Unless the cards would render stuff separately somehow but that isn't possible. Is this just your wishful thinking or is there any information that this might happen?

 

[Slizzo]

At what time do we expect stuff mentioned in OP not to be mentioned by the Lisa Su's uncle at GTC?


No shit Watson.


Citation needed.

How about from this very website/forum?

https://www.techpowerup.com/249077/...with-turing-based-cards-its-not-a-broad-issue

Would you like to rescind your vitriolic response?

 

[EarthDog]

How about from this very website/forum?

https://www.techpowerup.com/249077/...with-turing-based-cards-its-not-a-broad-issue

Would you like to rescind your vitriolic response?

LOL, like he will believe that or cite his own source of where they were notably higher... that glass house... those stones.

 

[Vayra86]

Come to think of it, This isn't a bad idea. If Nvidia does something like that they can only win with this. The problem is how would they do that? SLI isn't so reliable and games developers don't want to spend resources to support it. Unless the cards would render stuff separately somehow but that isn't possible. Is this just your wishful thinking or is there any information that this might happen?

We know Nvidia for its proprietary solutions. They have NVLink which is latency wise, fast enough for this. They have their RT/Tensor cores. I have no information, and its not wishful thinking, but rather one of the possibilities. Another one, which I deem far more realistic, is much deeper integration in the GPU / shader itself. Either way, there are quite a few options on the table.

 

[efikkan]

So RT/Tensor cores being proprietary? Well of course they are. When was the last time AMD published verilog files for their shader processors? As far as I know all major GPU microachitectures are proprietary. The only documentation I've seen is architectural overviews and native APIs.

Ray tracing support is implemented through Vulkan, an open standard, and DXR in DirectX, a proprietary Microsoft standard, and neither of these are locked to Nvidia hardware.

 

[Prima.Vera]

NVIDIA could leverage 7 nm to increase transistor densities, and bring its RTX technology to even more affordable price-points.

'Affordable', meaning at least at HALF of the existing callous and ridiculous prices??

 

[medi01]

How about from this very website/forum?

"On this very website" nVidia says, "problem exists', but "it is not a big issue.
Are you freaking kidding me?

So RT/Tensor cores being proprietary?

Tensor cores aren't (I mean it's a generic idea of hardware accelerating a trivial operation).
However RT cores are (figuring what and how they do to do it so quickly is not trivial)

 

[efikkan]

"On this very website" nVidia says, "problem exists', but "it is not a big issue.

Are you freaking kidding me?

What matters is the rate of defective cards. RMA rates of 2-5% is completely normal, Turing has been below average so far.
Most cases of reported problems were bugs in the launch driver, totally unrelated to the few defective cards.

Tensor cores aren't (I mean it's a generic idea of hardware accelerating a trivial operation).
However RT cores are (figuring what and how they do to do it so quickly is not trivial)

What they do on an API level is clearly defined.
What they do on a transistor level is proprietary, and shouldn't matter to end-users or programmers. This is no different from any other units inside GPUs and CPUs.

 

[ratirt]

What matters is the rate of defective cards. RMA rates of 2-5% is completely normal, Turing has been below average so far.
Most cases of reported problems were bugs in the launch driver, totally unrelated to the few defective cards.

You are wrong. It isn't normal. Never was. The companies feed you with this explanation "it is normal" that the cards can be defective and can malfunction and you take it. The fact is they shouldn't cause any problems. All cards are tested and they need to pass the test before they are shipped to the markets. Of course bad cards can happen but not on such scale. This bad cards should be a per mil of the production. I'm not talking only about NV but AMD and other producers of any kind of equipment. Stop going for this malarkey explanation companies give you. 2-5 % of all cards is few to you and how many 2000 series cards NV shipped last year? Gosh didn't know that "few" has been redefined for thousands.

 

[efikkan]

You are wrong. It isn't normal. Never was. The companies feed you with this explanation "it is normal" that the cards can be defective and can malfunction and you take it. The fact is they shouldn't cause any problems. All cards are tested and they need to pass the test before they are shipped to the markets. Of course bad cards can happen but not on such scale. This bad cards should be a per mil of the production. I'm not talking only about NV but AMD and other producers of any kind of equipment. Stop going for this malarkey explanation companies give you. 2-5 % of all cards is few to you and how many 2000 series cards NV shipped last year? Gosh didn't know that "few" has been redefined for thousands.

You are either clueless or misguided.
Failure rates of ~2% for DOA/almost immediate failures and upwards to 5% during the warranty is completely normal for consumer electronics. There is a reason why manufacturers provide warranty after all. Failure rates of 1 per million (0.0001%) is unheard of for consumer electronics, ask any retailer or manufacturer. Personally I've RMAed a lot of stuff over the years, including many hard drives, several graphics cards, laptops and even two identical server boards which were DOA.

Consumer electronics are tested during production, and serious makers do have a basic QA procedure on every sample, usually a test run of a few minutes and a visual inspection. They usually also do a lengthy test of a small random selection of each batch. But they can't run every sample on a 2 week stress test before shipping, not only would it cost too much, but it will also cut like 3 months of the product's life expectancy, doing too much testing might actually hurt RMA rates. Most DOA products are products that have failed between the factory and the end user. Transporting products around the world is not risk free, rapid changes in temperature and pressure will cause weak parts to fail, and you really can't completely eliminate this from the factory side.

So far, there is no evidence supporting that Turing have higher failure rates than normal, In fact, according to Steve at Gamers Nexus, several board partners are claiming that Turing so far have lower than normal failure rates. So until you have evidence proving he is lying, stop spreading your FUD.

 

[ratirt]

You are either clueless or misguided.
Failure rates of ~2% for DOA/almost immediate failures and upwards to 5% during the warranty is completely normal for consumer electronics. There is a reason why manufacturers provide warranty after all. Failure rates of 1 per million (0.0001%) is unheard of for consumer electronics, ask any retailer or manufacturer. Personally I've RMAed a lot of stuff over the years, including many hard drives, several graphics cards, laptops and even two identical server boards which were DOA.

Consumer electronics are tested during production, and serious makers do have a basic QA procedure on every sample, usually a test run of a few minutes and a visual inspection. They usually also do a lengthy test of a small random selection of each batch. But they can't run every sample on a 2 week stress test before shipping, not only would it cost too much, but it will also cut like 3 months of the product's life expectancy, doing too much testing might actually hurt RMA rates. Most DOA products are products that have failed between the factory and the end user. Transporting products around the world is not risk free, rapid changes in temperature and pressure will cause weak parts to fail, and you really can't completely eliminate this from the factory side.

So far, there is no evidence supporting that Turing have higher failure rates than normal, In fact, according to Steve at Gamers Nexus, several board partners are claiming that Turing so far have lower than normal failure rates. So until you have evidence proving he is lying, stop spreading your FUD.

I understand what you are saying but I guess you don't get what I want to express here.
I'm not talking about turing. The chips are fine so the quality and the production and failure rates make sense and it is below 1%. What you are referring to is the memory failure. Which causes the failure. You are focusing on the GPU as a product which has components from dozen of vendors. I don't and that's the difference. You consider graphics card as a product and that's ok. Maybe you should focus on, for instance chip as a product which will be used in variety of cards form different vendors. Then your 2-5% is ridiculous. That's what I'm talking about.
For instance it has been said that the Turing cards it is mostly due to memory not the chip. So it means that memory modules aren't being examined or tested properly.
So if you have 5% of broken cards and the percentage of these broken cards is for example 95% with mem issue, no it isn't normal.

 

[efikkan]

I'm not talking about turing. The chips are fine so the quality and the production and failure rates make sense and it is below 1%. What you are referring to is the memory failure. Which causes the failure. You are focusing on the GPU as a product which has components from dozen of vendors. I don't and that's the difference. You consider graphics card as a product and that's ok. Maybe you should focus on, for instance chip as a product which will be used in variety of cards form different vendors. Then your 2-5% is ridiculous. That's what I'm talking about.

For instance it has been said that the Turing cards it is mostly due to memory not the chip. So it means that memory modules aren't being examined or tested properly.

So if you have 5% of broken cards and the percentage of these broken cards is for example 95% with mem issue, no it isn't normal.

Failure rates, for any reason, for consumer electronics like graphics cards, motherboards or laptops typically are in the 2-5% range.

The "memory problem" you refer to with Turing is not a specific memory problem, it's in fact all kinds of random problems which happens with any series of graphics cards. The problem with this matter is that all kinds of normal problems are mixed into one giant bag and called one problem. Most of the reported cases had no hardware defects at all, and were experiencing one of the various distinct driver issues. The remaining cards have various hardware defects, which are covered by warranty. And that "space invader" pattern doesn't necessary mean it's a memory problem, just that something causes memory corruption. It could just as well be power delivery, memory controller or even weak traces on the PCB. You need proper debugging tools to find out what's causing it.

 

[ratirt]

Failure rates, for any reason, for consumer electronics like graphics cards, motherboards or laptops typically are in the 2-5% range.

The "memory problem" you refer to with Turing is not a specific memory problem, it's in fact all kinds of random problems which happens with any series of graphics cards. The problem with this matter is that all kinds of normal problems are mixed into one giant bag and called one problem. Most of the reported cases had no hardware defects at all, and were experiencing one of the various distinct driver issues. The remaining cards have various hardware defects, which are covered by warranty. And that "space invader" pattern doesn't necessary mean it's a memory problem, just that something causes memory corruption. It could just as well be power delivery, memory controller or even weak traces on the PCB. You need proper debugging tools to find out what's causing it.

How can you know they didn't have defects when they have failed actually? What kind of thinking this is? No defects and yet they fail. I don't understand this. It makes no sense or has absolutely no logic behind it.
If it fails it is defective for any reason. In terms of turing and memory related issues. What is it exactly if you please enlighten me? If it is not memory specific then what is it? Don't tell me random components. I just don't buy it. Large chunk of the problems was memory related. Maybe it's because it is new technology but doesn't change anything.

 

[efikkan]

How can you know they didn't have defects when they have failed actually? What kind of thinking this is? No defects and yet they fail. I don't understand this. It makes no sense or has absolutely no logic behind it.

I don't understand any of this, so perhaps you misunderstood me?

People reported various symptoms, after a few weeks at least two distinct driver issues were identified and fixed. Most of the reported cases in forums etc. fit one of those buckets. Those that remain are probably some kind of hardware defect.

If it fails it is defective for any reason. In terms of turing and memory related issues. What is it exactly if you please enlighten me? If it is not memory specific then what is it? Don't tell me random components. I just don't buy it. Large chunk of the problems was memory related. Maybe it's because it is new technology but doesn't change anything.

I'm telling you a symptom is a symptom, and doesn't necessary reveal the underlying problem.
For the end user it doesn't matter, since a reproducible hardware defect is a defective product, and they will get a new one.

When it comes to the so-called memory diagnosis from various forum users, who claimed it was one bad batch from one supplier, but it was later debunked as the "problem" was reproduced across two memory vendors and multiple batches of chips. Some of these cards may certainly have failing memory chips, but you can't claim that just based on the crash.

 

[ratirt]

I don't understand any of this, so perhaps you misunderstood me?

When it comes to the so-called memory diagnosis from various forum users, who claimed it was one bad batch from one supplier, but it was later debunked as the "problem" was reproduced across two memory vendors and multiple batches of chips. Some of these cards may certainly have failing memory chips, but you can't claim that just based on the crash.

I'm not claiming that crash is based on the memory. The first findings when the crashing appeared where due to memory issues. It has been told several times that the memory was behind all of this. Maybe it has been fixed on the way but still it happened and it has been confirmed.