nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Child Item

Nvidia RTX Discussion

K.O.Kiki

A new era in 3D graphics stands before us.

First-gen Raytracing GPUs:

Nvidia "Turing" - Quadro RTX 5000 / 6000 / 8000 (US$ 2'300 / 6'300 / 10'000)

https://www.anandtech.com/show/13214/nvidia-reveals-next-gen-turing-gpu-architecture
https://www.anandtech.com/show/13217/nvidia-announces-turing-powered-quadro-rtx-family

Find more posts tagged with

Quick Links

Comments

RobertKK

I don’t want a problem for anyone who bought the new GPU. I was just pointing out an issue. I bought a 1080 recently, would I want there to be a problem? Well obviously no and I wouldn’t wish it on anyone else either. One expects the same standard of quality for all.
Imagine if you designed your dream PC and the GPU sent it all up in flames.
It is better to be aware of a problem that arose elsewhere rather than being ignorant that there was a fire issue elsewhere.
I don’t want anyone to have bought a faulty one, going on fire is a potentially very serious issue. Thankfully it looks to be rare at the moment and hopefully it stays that way.

Genevieve Disgusting Self-improvement wrote: »

I stand by my post although I admit I do assume the motivation of the poster.

He's not a board regular Fitz (unless I'm very mistaken), and he's bringing interesting information - so maybe take him at face value rather than assuming he wants to bonfire your RTX to dance by the warmth of the flames.

Different note, I'll dig up the link when I have a little more time, but it looks like Overclockers.co.uk are quoting around a 3.5% return rate - about half of which is failure and half is 14 day returns.

Not sure what to make of that though it does fit nicely with the earlier percentages we got.

Also of interest, it looks like there's a dud batch of founder edition RTX 2080 Tis. I'm very curious what the failure rate for that batch is, and what the general failure rate excluding it is.

tuxy

K.O.Kiki wrote: »

4K with AA is dumb.

AA with 1440p was also considered really stupid if anyone thinks back to when the first few people were making the switch to that resolution.

BloodBath

L wrote: »

He's not a board regular Fitz (unless I'm very mistaken), and he's bringing interesting information - so maybe take him at face value rather than assuming he wants to bonfire your RTX to dance by the warmth of the flames.

Different note, I'll dig up the link when I have a little more time, but it looks like Overclockers.co.uk are quoting around a 3.5% return rate - about half of which is failure and half is 14 day returns.

Not sure what to make of that though it does fit nicely with the earlier percentages we got.

Also of interest, it looks like there's a dud batch of founder edition RTX 2080 Tis. I'm very curious what the failure rate for that batch is, and what the general failure rate excluding it is.

Again though until you can prove the numbers are higher than normal compared to any electronics failure rate, which you haven't, then it's not even worth bringing up.

It just comes across as a desperate attempt to piss on these cards even more. Like anti Tesla people latching on to a car fire/death despite the fact the cars are far safer than normal cars.

Revoker88

How is bringing up an issue which could potentially affect people who have bought a 2080 ti an attempt to piss on the cards?

Would you not want to know if there was potentially something wrong with a very expensive card?

Even if the chance of anyone here getting a problem card is tiny,personally id rather know and try find out if my card was one of the affected so i could RMA.

BloodBath wrote: »

Again though until you can prove the numbers are higher than normal compared to any electronics failure rate, which you haven't, then it's not even worth bringing up.

It just comes across as a desperate attempt to piss on these cards even more. Like anti Tesla people latching on to a car fire/death despite the fact the cars are far safer than normal cars.

...Sure. Being interested in understanding what's going on with card failures clearly means I hate RTX. Come on...

BloodBath

Did you have the same interest in card failure rates in the past or are you latching on to any negative media surrounding these cards because of confirmation bias?

Where is the proof that the failure rates are above average to make this a news story in the first place?

Yes actually I did. Comes from having expertise in manufacturing.

Now, if you go back and read my post you'll notice that I imply I'm interested in the higher failure rate of the bad batch and the *lower* rate of failure that the other cards may have as a result.

Of course, you could go on being the only one in the thread implying I'm skewed on this issue. Your call

BloodBath

I don't think you are the only 1. I'm just waiting for actual proof. If you had expertise in manufacturing then you know that around 5% failure rate of electronics in the first year is the norm.

Now onto where is the proof that it is above 5%?

You might want to reread what I actually wrote there bloodbath. It's really not what you seem to have read or inferred.

On a related note, if you go back earlier in the thread you'll see me commenting that 5% is pretty typical for cards. Good enough?

You could also just quit being painfully combative for no real reason. It's not adding anything

btkm8unsl0w5r4

This is exactly my point. People rightly say "well I would want to know of an issue"....thing is your being told about everything negative, be it issue or not, and everything is being turned into an issue in the absence of evidence and common sense. If somebody cut themselves installing an RTX you would have an article the next day about the sharp edges, and an all too eager cohort of people waiting to spread the "news" that hyper expensive RTX cuts your fingers off. Its becoming irritating at this stage.

Gamers Nexus had a vid yesterday reporting that the RMA numbers are very low, and well withing manufacture expected numbers. So that whole thing was a fart in the wind. But thats not being reported. The initial RMA story was vomited out into the world without context and then abandoned. Not the same on here cause there is a discussion, if that me painfully combative (combat only hurts if your the one loosing

) so be it.

tuxy

RTX cards have a 0% failure rate amongst people partaking in this thread!

Genevieve Disgusting Self-improvement wrote: »

if that me painfully combative (combat only hurts if your the one loosing ) so be it.

It doesn't Fitz - you read what I wrote, and didn't make it weirdly personal.

Also, you're a dentist, don't tell me you've never had to treat someone who "won" a fight.

btkm8unsl0w5r4

L wrote: »

It doesn't Fitz - you read what I wrote, and didn't make it weirdly personal.

Also, you're a dentist, don't tell me you've never had to treat someone who "won" a fight.

Only see people claiming to have won...the real winners dont require my services, or maybe dont get into fights.

Genevieve Disgusting Self-improvement wrote: »

or maybe dont get into fights.

That's the one.

Back on topic, some useful links:

Fairly balanced article setting out the state of play as of last week.

Nvidia's statement about test escapes. Original Nvidia post here.

Steve from Gamersnexus has a new interesting article/video with some more analysis/numbers from the dud cards he got sent in. The TLDR is they figure there's a silicon or board problem showing up but it's pretty uncommon.

For myself, there's two elements here that really interest me.

First, Nvidia's characterization of these as test escapes is significant - it means they're admitting their test team had their test vectors wrong for some of the early RTX batches. I'd really love to see what percentage of those card batches were RMA'd vs the general run of cards.

Second, the RTX RMA rate is still very low (especially since they've such a large die size), but their performance and transistor density is also much lower than I'd expect for a 12nm die shrink (even giving the benefit of the doubt that maybe a third of the die is custom special sauce). I'm really curious if that's down to them speccing a high core count and aggressively cutting back down post-testing to manage their failure rates. There's no way for me to get an answer on that though so I'm stuck with wild speculation.

K.O.Kiki

tuxy wrote: »

AA with 1440p was also considered really stupid if anyone thinks back to when the first few people were making the switch to that resolution.

:pac:

K.O.Kiki

tuxy

Time to buy stock rather than cards?

tuxy wrote: »

Time to buy stock rather than cards?

Still wildly overpriced. Memes aside, this is some of the crypto bubble deflating not RTX related. Nvidia should settle back down somewhere in mid double digits.

tuxy

Yeah, their predicted profits were massive if the crypto thing had kept going and they ramped up production of pascal for it. The share price should stabilise over the net quarter.

btkm8unsl0w5r4

The scale may have been chosen to make a comedic point, but somebody who bought in mid 2016 and has quadrupled their investment might be the one laughing. I posted a few pages back that this drop was well flagged, nows the time to buy, AMD's Navi push will rally when markets reopen potentially.

6034073

Genevieve Disgusting Self-improvement wrote: »

The scale may have been chosen to make a comedic point, but somebody who bought in mid 2016 and has quadrupled their investment might be the one laughing.

Sure, but back in 2016 their share price and fundamentals had a saner relationship. Price/Earnings was sub 20 then, and up until the recent drop it was around 50ish (it's now down around 25ish). Similar for Price/book-value - you were looking at ~4 then, and ~20 until last week (now 11ish).

2016 Nvidia made sense on the face of it. 2018 Nvidia doesn't.

Genevieve Disgusting Self-improvement wrote: »

I posted a few pages back that this drop was well flagged, nows the time to buy,

Oh man, I really disagree on this one. I wouldn't touch Nvidia's stock with a 10 foot pole at the moment. Not because I think they're a bad company, but because I think their stock is still very inflated relative to the rest of their industry and their own historical values. It's long odds of their stock price going back up from here longterm.

btkm8unsl0w5r4

L wrote: »

Sure, but back in 2016 their share price and fundamentals had a saner relationship.

I know, and I knew back then. However the oppertunity for others to buy in 2016 has passed.

L wrote: »

2016 Nvidia made sense on the face of it. 2018 Nvidia doesn't.
.

Nearly all of the analysts disagree with you recommending a buy or a hold. I will add your opinion to the analysis and weight it duly.

https://www.cnbc.com/2018/11/16/heres-what-wall-street-analysts-had-to-say-about-nvidias-earnings.html

https://investorplace.com/2018/11/should-you-buy-nvidia-stock-during-its-crypto-hangover/

Genevieve Disgusting Self-improvement wrote: »

Nearly all of the analysts disagree with you recommending a buy or a hold. I will add your opinion to the analysis and weight it duly.

https://www.cnbc.com/2018/11/16/heres-what-wall-street-analysts-had-to-say-about-nvidias-earnings.html

https://investorplace.com/2018/11/should-you-buy-nvidia-stock-during-its-crypto-hangover/

Interesting. I still wouldn't buy it but I'll have some fun figuring out why they're still recommending it.

Edit: Here we go.

Nvidia’s datacenter revenue grew 83% year-over-year, while its gaming sales jumped 52% and the revenue generated by its professional graphics business climbed 20%. Anyone who calls Nvidia stock a bubble must believe that autonomous cars will not come to fruition, datacenters will dissipate, gaming is a fad and professional graphics will no longer be necessary.

On his list, I don't believe Nvidia will be the industry leader for autonomous cars (when they arrive) or datacentres running forward (they've already had issues here having to artificially restrain datacentres from using their consumer cards), and I don't see their gaming and professional graphics being sufficient to support their current share price alone.

So, yeah, that more or less explains it.

btkm8unsl0w5r4

L wrote: »

Interesting. I still wouldn't buy it but I'll have some fun figuring out why they're still recommending it.

Best to publish your recommendations after you have done your research

...They overtly tell you why, and if you have access to a trading platform you can get much deeper analysis.

The sudden drop was because Nvidia had flagged the lower profit some weeks back but not flagged the extent of the difference between the projections and reality. The market has not fully priced in the figures and as is the way with a spooked and pissed off market it tends to overreact and punish. The core business of Nvidia has not changed and it should return to normal levels soon enough albeit not in the very short term. My opinion is likely Jenson might "step down", if he does it will stay low for a while longer.. although it could rebound, the AMD news is not priced in yet. Still for the medium term investor Nvidia at that price is a good bet.

Genevieve Disgusting Self-improvement wrote: »

Best to publish your recommendations after you have done your research

Publish is such a strong word for a throwaway comment on a forum, and research is such a strong word for "reading analysts recommendations". :P

Genevieve Disgusting Self-improvement wrote: »

The sudden drop was because Nvidia had flagged the lower profit some weeks back but not flagged the extent of the difference between the projections and reality. The market has not fully priced in the figures and as is the way with a spooked and pissed off market it tends to overreact and punish. The core business of Nvidia has not changed and it should return to normal levels soon enough albeit not in the very short term. My opinion is likely Jenson might "step down", if he does it will stay low for a while longer.. although it could rebound, the AMD news is not priced in yet. Still for the medium term investor Nvidia at that price is a good bet.

I think Jensen stepping down is fairly unlikely - it'd be a blow for Nvidia's brand long term.

You may have a point about the market being spooked - again though, I don't see that shareprice having been supported by Nvidia's fundamentals. It's purely speculative, and this drop does have a lot of the features of the end of a bubble (even down to a bull trap at the end of October). I guess we'll see from next week which direction things go.

btkm8unsl0w5r4

Forum posts are publishing and are subject to liable law not slander law I know this from 15 years moderating.

Reading about a topic from reliable independent sources is what research is. Do you have any to support you opinion? *** edit, I see you added to your post, but if you want that info just go back to my post on the matter a week ago, Nvidia are already a huge player in autonomous cars and have contracts with most of the big german manufacturers and tesla, and have most of the GPU deep learning business sown up, your opinion is however interesting is contrary to reality *** Tech bubbles tend to burst much more dramatically, Nvidia is also massively profitable, its just the profits were not what was flagged and dividends will suffer. I dont know about a bull trap, but I do know when somebody is bullsh1tting.

, nice try all the same. Of course nobody should risk any money they cannot afford to loose and I am in no way giving advice, it just seems to be the consensus that a buy on Nvidia may be a good idea.

To put this into perspective read this. Seems even you guys on here couldn't clear the pascal stock enough.

(CercleFinance.com) - Nvidia's forecasts fell short of Wall Street expectations as the chip-maker wrestles with high levels of inventories following the crypto-currency boom, which sent its shares down sharply on Friday.

The company reported late on Thursday non-GAAP earnings per diluted share of 1.84 dollar, up 38% from a year earlier, for the third quarter ended 28 October, on revenues that are up 21% at 3.2 billion dollars, although falling 2% short of expectations.

Nvidia said that results reflected 'excess channel inventory' post the crypto-currency boom and its 'correction.'

So-called 'miners' use hgih-performance computers and servers to process crypto-currencies, such as bitcoin.

Nvidia also gave a revenue forecast for the current quarter that fell short of Wall Street expectations.

'The January quarter revenue outlook was worse, coming in 20% below consensus and the company blamed excess inventories of older (Pascal) crypto GPUs in the channel,' analysts at Jefferies said.

The company expects fourth-quarter revenue of 2.7 billion dollars, plus or minus 2%.

Nvidia shares are currently in a spin - tumbling 18.7% after these results.

Copyright (c) 2018 CercleFinance.com. All rights reserved

And here are the company stats for an idea of the fundamentals in relation to those figures. And the historic Price earning ratio showing its around average for the last 7 years.

6034073

Genevieve Disgusting Self-improvement wrote: »

Forum posts are publishing and are subject to liable law not slander law I know this from 15 years moderating.

And when you see me libel someone, you can remind me of that. :P Until then, it remains a fierce strong way to describe a forum post.

Genevieve Disgusting Self-improvement wrote: »

Reading about a topic from reliable independent sources is what research is.

These aren't independent sources though Fitz, they've skin in the game themselves. You can't look at industry and academia as the same thing as we both well know.

Genevieve Disgusting Self-improvement wrote: »

Do you have any to support you opinion? Tech bubbles tend to burst much more dramatically, Nvidia is also massively profitable, its just the profits were not what was flagged and dividends will suffer.

Sure, I've a link to Jean-Paul Rodrigue on bubbles below.

Bubbles come in all sorts of shapes and sizes. The tech bubbles you're talking about, unless I'm very much mistaken, are companies with a single overhyped product and no past record. Nvidia doesn't fit that profile - it does however not have profits or revenue that support this current level of shareprice.

It's super profitable as a company but bad as an investment at the moment - unless you intend on using the shareprice itself to drive your returns (which, again, fits into the characterization of this as a speculative bubble).

Genevieve Disgusting Self-improvement wrote: »

Of course nobody should risk any money they cannot afford to loose and I am in no way giving advice, it just seems to be the consensus that a buy on Nvidia may be a good idea.

Definitely - never invest what you can't lose. I can also see what the consensus is. I just happen to disagree with it, but, you know, Dunning-Kruger so amateurs beware.

Genevieve Disgusting Self-improvement wrote: »

I dont know about a bull trap, but I do know when somebody is bullsh1tting. , nice try all the same.

You're getting oddly personal about it again Fitz. I'm not really sure why? Is it true what people say? Do you happen to be sitting on a good bit of Nvidia stock? :P

This is what a bull trap is btw.

Note the features of the Classic bubble stages.

Now look below. See any features that look oddly familiar? This scale looks a little funky for the bull trap, but if you look at just the last six months or so, it'll snap into focus.

Genevieve Disgusting Self-improvement wrote: »

And here are the company stats for an idea of the fundamentals in relation to those figures. And the historic Price earning ratio showing its around average for the last 7 years.

Now look at the earnings per share and compare that to the rise in share price over the same period. They don't line up - and when they don't line up, that means the share price is based on speculation. And when the price of something is primarily based on speculation, that means that there's a strong chance it's a ...

btkm8unsl0w5r4

Ok now that you have done the research and edited your previous posts with some accurate information, has you opinion on the shares changed? I know what a bull trap is thanks.

Earnings per share means nothing over time as there was a stock dilution and EPS does not take into account reinvestment, R&D and all that jazz. Its doesnt scale with stock price. Price earnings ratio is what matters if you want to see if a stock is overpriced in comparison of the PE ratio of the comparable shares in the industry, thats why I put in that graph. People call bubbles all the time...you can find correlation in pretty much any stock price over time with a classical bubble. Stocks go up and stocks go down. The up and down of Nvidia has had very good reasons for it, not just speculation.

Its not personal, its just your opinion does not seem that informed, but I am glad this discussion has given you the impetus to do the research, and its made me look more closely than I have in a while.

No I am not sitting on Nvidia stock,