Tuesday, February 24th 2015

"It Won't Happen Again:" NVIDIA CEO Breaks Silence on GTX 970 Controversy

In the wake of bad PR, and a potentially expensive class-action lawsuit over the GeForce GTX 970 memory controversy, NVIDIA CEO Jen-Hsun Huang wrote a candid letter addressed to everyone concerned, explaining in the simplest possible language what went wrong with designing and marketing the chip, how it doesn't affect the design-goals of the product, its quality or stability, and how it could be misconstrued in a whole different ways.

Huang's explanation of the issue isn't much different from the one we already have, but bears the final stamp of authority from the company, especially with the spate of discrepancies between what NVIDIA representatives post on GeForce forums, and what ends up being the company's position on certain things. Huang's letter signs off with "we won't let this happen again. We'll do a better job next time."

The transcript of Huang's letter follows.

Hey everyone,

Some of you are disappointed that we didn't clearly describe the segmented memory of GeForce GTX 970 when we launched it. I can see why, so let me address it.

We invented a new memory architecture in Maxwell. This new capability was created so that reduced-configurations of Maxwell can have a larger framebuffer - i.e., so that GTX 970 is not limited to 3GB, and can have an additional 1GB.

GTX 970 is a 4GB card. However, the upper 512MB of the additional 1GB is segmented and has reduced bandwidth. This is a good design because we were able to add an additional 1GB for GTX 970 and our software engineers can keep less frequently used data in the 512MB segment.

Unfortunately, we failed to communicate this internally to our marketing team, and externally to reviewers at launch.

Since then, Jonah Alben, our senior vice president of hardware engineering, provided a technical description of the design, which was captured well by several editors. Here's one example from The Tech Report.

Instead of being excited that we invented a way to increase memory of the GTX 970 from 3GB to 4GB, some were disappointed that we didn't better describe the segmented nature of the architecture for that last 1GB of memory.

This is understandable. But, let me be clear: Our only intention was to create the best GPU for you. We wanted GTX 970 to have 4GB of memory, as games are using more memory than ever.

The 4GB of memory on GTX 970 is used and useful to achieve the performance you are enjoying. And as ever, our engineers will continue to enhance game performance that you can regularly download using GeForce Experience.

This new feature of Maxwell should have been clearly detailed from the beginning.

We won't let this happen again. We'll do a better job next time.

Jen-Hsun
Source: NVIDIA
Add your own comment

140 Comments on "It Won't Happen Again:" NVIDIA CEO Breaks Silence on GTX 970 Controversy

#26
GhostRyder
Not sure how to count this, the wording hints at multiple things instead of sounding like an apology.
the54thvoidlol,

It's a very clever apology, not sincere in the slightest but very clever.
Yea I am with you on that, it really has no sincerity in it but more than anything it almost sounds to me at least like he is saying we are ungrateful to them for this. Maybe I am reading into it to much...
ToothlessWhat the hell guys. You all have miscommunications, and don't make a big deal. We're all human, we make mistakes, get over it.
NV had a miscommunication, so what, no one is perfect. Quit bashing them when they at least own up to their mistake.

Next time I see one of you mess up, I'll just link this thread into your post and remind you of how much of an ass you are.

NO HUMAN IS PERFECT.
We are all human, and we all do make mistakes. However your telling me a multi-million dollar company did not catch these types of things before sending it out to the public along with specs sheets for months? That's a different story and I am pretty sure NVidia has some sort of way to check these things before speaking or showing to the public since most businesses want to cover themselves as much as possible.
heydan83So basically they tried to make the 970 a 4gb card and they saw that they would not make it and maybe they saw that advertising the card has a 3gb card wouldn´t be that excitement from a 780 3gb card and decided to advertise it as a 4gb card.
That sounds about right honestly if we are going off of the intentional theory.

I honestly see something more here overall than just an oversight, but that is just my opinion and Jen-Hsun's comment did not exactly help my opinion on that.
Posted on Reply
#27
BUCK NASTY
4P Enthusiust
I accept your apology Jen-Hsun Huang. Now give me my settlement money for the 6x GTX 970's that I own.
Posted on Reply
#28
Outback Bronze
They should have marketed a 3GB card so everyone would have been talking about the BONUS 512 of memory
instead of marketing a 4GB and everybody getting pissed about only having 3.5GB.
No excuses and extremely unprofessional from a multi million/billion dollar company.
Posted on Reply
#29
EarthDog
BUCK NASTYI accept your apology Jen-Hsun Huang. Now give me my settlement money for the 6x GTX 970's that I own.
SMH...LOL
Posted on Reply
#30
cokker
"This is a good design"

No, Jen-Hsun... No it is not...

This letter is about as apologetic as putting lemon juice in my eyes.
Posted on Reply
#31
nunyabuisness
if this is an apology letter. they its the worst!
no offer to refund, no offer to upgrade, nothing! frigging CHEAT
Posted on Reply
#32
metalslaw
"Instead of being excited that we invented a way to increase memory of the GTX 970 from 3GB to 4GB, some were disappointed that we didn't better describe the segmented nature of the architecture for that last 1GB of memory."

4 months ago...

Me - "Did you hear about that new feature with the gtx970... I'm so excited by it!"

Friend - "No, what is it?"

Me - "I don't know. They haven't told us about it. So i'm excited by it!"

Friend - "That's great logic dude! I'm excited now too."
Posted on Reply
#33
heydan83
metalslaw"Instead of being excited that we invented a way to increase memory of the GTX 970 from 3GB to 4GB, some were disappointed that we didn't better describe the segmented nature of the architecture for that last 1GB of memory."

4 months ago...

Me - "Did you hear about that new feature with the gtx970... I'm so excited by it!"

Friend - "No, what is it?"

Me - "I don't know. They haven't told us about it. So i'm excited by it!"

Friend - "That's great logic dude! I'm excited now too."
Completly agree with this, if they want us to be exicited about "adding aditional weird 1gb" at least they should told us on release date...
Posted on Reply
#34
Xzibit
Unfortunately, we failed to communicate this internally to our marketing team, and externally to reviewers at launch.
That's the BS part. The driver team knew since they were allocating low access resources into the 512mb sector. It doesn't fly that there was a communication problem if 2 separate departments hardware and software knew with software developing drivers continuing to optimize for it. Reeks of deception or cover your ass because the lawsuits are coming.
Posted on Reply
#35
RejZoR
Isn't Heng
heydan83I dont think this is a Nvidia vs AMD discuss, this is about consumer vs enterprises, and also all that you said is not "hidden the truth" to the consumer but maybe change the strategy, so even if AMD or any other company that make this kind of thing should be treated the same....
Exactly. I'm not against NVIDIA as such. I'm against them when they pull shit like this with GTX 970. I don't hate GTX 980, because that one is a good card. But at the same time it sucks to be their cusotmer, because you're placed in a shitty position because of it. Have a GTX 970 with potentially questionable performance or fork out a lot more for GTX 980 in which case you'll just be fueling their lies. After all, the controversy kinda forced more people to buy more expensive GTX 980 if they wanted NVIDIA card for whatever reason...
Posted on Reply
#36
HumanSmoke
the54thvoidlol,
It's a very clever apology, not sincere in the slightest but very clever.
More like an explanation worded to sound like an apology
BUCK NASTYI accept your apology Jen-Hsun Huang. Now give me my settlement money for the 6x GTX 970's that I own.
Don't count on too big a windfall
Posted on Reply
#37
esrever
Buy a table and and 4 chairs,
Some one sits on the fourth chair,
It collapses.
"Well we were originally going to sell it with only three chairs, but we found an innovative new way to add a fourth. The last one is made of cardboard and bubble gum. We thought you'd be excited!"
Posted on Reply
#38
renz496
jordan199Nvidia it's time for a new CEO, I hope will replace with someone who RESPECT the company and consumers...
if anything i think this is the hardest thing to happen. AFAIK they have much worse mess with Fermi and JHH still lead the company after that. plus JHH won't give up the CEO position even if other company to buy Nvidia lol.
Posted on Reply
#39
heydan83
esreverBuy a table and and 4 chairs,
Some one sits on the fourth chair,
It collapses.
"Well we were originally going to sell it with only three chairs, but we found an innovative new way to add a fourth. The last one is made of cardboard and bubble gum. We thought you'd be excited!"
I dont know why is everyone complaining about it, my table works great and I dont have any intentions to use the fourth chair either way, and if I need it, Ill change my hole set for a new better set..... of course of the same brand... just in case I want finish this by saying I'm being sarcastic...
Posted on Reply
#40
HumanSmoke
renz496
jordan199Nvidia it's time for a new CEO, I hope will replace with someone who RESPECT the company and consumers...
if anything i think this is the hardest thing to happen. AFAIK they have much worse mess with Fermi and JHH still lead the company after that. plus JHH won't give up the CEO position even if other company to buy Nvidia lol.
Companies as a general rule don't oust successful CEO's - especially when they are founders. Balance the GTX 970 issue againstboth a record last quarter and record year, and outselling their principle competitorwith discrete GPUs versus the competitions GPU + APU portfolio. If the role was reversed, how likely would it be that AMD or Intel would fire their CEO after a record year in revenue because of an issue with a single SKU?
heydan83I dont know why is everyone complaining about it, my table works great and I dont have any intentions to use the fourth chair either way
Seating for three, or up-end the chairs and you have seating for 12 of the lawyers involved in the civil suit.
Posted on Reply
#41
iO
But hey it only took them five weeks and an incoming class-action suit to officially respond on the topic...
Posted on Reply
#42
Iceni
I would love to defend Nvidia in this.

My own 970's are fantastic. They would have still been fantastic if they had not attempted to hide a "feature".

If they had not bothered to add in that last 1GB then the problem is where does the 780 end and the 970 start. So they tweaked the configuration and put-in a semi working feature with the hopes of controlling it with software designed to not allocate to that segment of memory.

What that statement is actually saying is we were hoping to force these cards to opperate as a 3.5GB card via drivers. So they had no intention of ever having the final user have access to the whole 4GB of memory. So in effect the statement he has given is more of a rod for his own back in this instance.

Bring me the popcorn this is better than Samsung V/s Apple.
Posted on Reply
#43
harry90
"it wont happen again" "next time well do better" lol there is no next time nvidia, no way im gonna buy nvidia ever.
Posted on Reply
#44
newtekie1
Semi-Retired Folder
harry90"it wont happen again" "next time well do better" lol there is no next time nvidia, no way im gonna buy nvidia ever.
So who will be buy? AMD? You remember when they released a card marketed at 1000MHz and then after a reasonable gaming session was found to drop down as low as 550MHz? It could last long enough to get through benchmarks, so the reviews were good, but actual user experience wasn't as good. Remember that? No? How about the time AMD released a card, let it go through all the reviews(which gave it very good reviews) and then released a driver that purposely and drastically reduced performance to stop the cards from dying pre-maturely? No. Don't remember that? Remember how they didn't even bother to apologize for either? Yeah, good times!

All you people need to get over your hate on nVidia because this is nothing compared to what some other companies have done and never even bothered to apologize for. What nVidia did was a marketing mistake, it doesn't affect the performance of the card we bought compared to the reviews we all read.
Posted on Reply
#45
Xzibit
newtekie1So who will be buy? AMD? You remember when they released a card marketed at 1000MHz and then after a reasonable gaming session was found to drop down as low as 550MHz? It could last long enough to get through benchmarks, so the reviews were good, but actual user experience wasn't as good. Remember that? No? How about the time AMD released a card, let it go through all the reviews(which gave it very good reviews) and then released a driver that purposely and drastically reduced performance to stop the cards from dying pre-maturely? No. Don't remember that? Remember how they didn't even bother to apologize for either? Yeah, good times!

All you people need to get over your hate on nVidia because this is nothing compared to what some other companies have done and never even bothered to apologize for. What nVidia did was a marketing mistake, it doesn't affect the performance of the card we bought compared to the reviews we all read.
Obviously you don't remember that AMD got crap over it because it never specified a base clock and used the words "Up To". Many people here defending Nvidia complained about that or they just conveniently forgot.

You seem to be implying its okay and should be accepted because others do it as its common practice. That's a sad state to be in as a consumer.
Posted on Reply
#46
Naito
I think Nvidia need to do less apologizing and more explaining of how they will correct the performance issues seen in some scenarios. It's obvious that having segmented memory seems to slow the entire memory system down when it has to deal with the smaller segment. Perhaps they need to add a 'switch' to their drivers in which the card completely ignores the smaller 512MB segment as to keep the 3.5GB segment at maximum speed. This may help remove micro-stutter and the other various issues reported.
Posted on Reply
#47
newtekie1
Semi-Retired Folder
XzibitObviously you don't remember that AMD got crap over it because it never specified a base clock and used the words "Up To". Many people here defending Nvidia complained about that or they just conveniently forgot.

You seem to be implying its okay and should be accepted because others do it as its common practice. That's a sad state to be in as a consumer.
I do remember they got crap over it, but they never apologized, they got crap over their drivers deliberately reducing performance too but didn't apologize either.

What I'm directly saying, not implying anything, is that people need to cut nVidia some slack. Their mistake was on paper, on paper that was private, it didn't affect the performance we were promised through reviews. The card has not changed just because the specs on paper did. People shouldn't be buying cards on specs anyway, or name, they should be buying them based on the performance they get in reviews. And nVidia did not change that performance, AMD on the other hand did. AMD did cheat in the reviews, nVidia didn't. Yet, nVidia apologizes and people still want to bash, AMD never even comments on the problems when they get caught.
NaitoI think Nvidia need to do less apologizing and more explaining of how they will correct the performance issues seen in some scenarios. It's obvious that having segmented memory seems to slow the entire memory system down when it has to deal with the smaller segment. Perhaps they need to add a 'switch' to their drivers in which the card completely ignores the smaller 512MB segment as to keep the 3.5GB segment at maximum speed. This may help remove micro-stutter and the other various issues reported.
That is the thing, having the 512MB only improves performance. Turning if off means that the card would start paging out to system RAM instead of accessing the 512MB, that is way worse.
Posted on Reply
#48
heydan83
newtekie1So who will be buy? AMD? You remember when they released a card marketed at 1000MHz and then after a reasonable gaming session was found to drop down as low as 550MHz? It could last long enough to get through benchmarks, so the reviews were good, but actual user experience wasn't as good. Remember that? No? How about the time AMD released a card, let it go through all the reviews(which gave it very good reviews) and then released a driver that purposely and drastically reduced performance to stop the cards from dying pre-maturely? No. Don't remember that? Remember how they didn't even bother to apologize for either? Yeah, good times!

All you people need to get over your hate on nVidia because this is nothing compared to what some other companies have done and never even bothered to apologize for. What nVidia did was a marketing mistake, it doesn't affect the performance of the card we bought compared to the reviews we all read.
I would say at least skip this generation donr support this kind of things
Posted on Reply
#49
damric
Dear Jen-Hsun,

Don't worry. As a consumer, my memory is shorter than that of an American voter. I'll still buy your stuff, even if you use lead paint and baby fetuses for thermal grease.

Your dearly beloved,

Fanboy.
Posted on Reply
#50
HumanSmoke
XzibitObviously you don't remember that AMD got crap over it because it never specified a base clock and used the words "Up To".
Probably comes down more to degree ( 700+ posts in four front page articles and a few forum threads - something AMD's throttling didn't even approach even though it affected more than one SKU, and many user scenarios). Everyone kind of gets it as this point, and while it is fun for some people to pile on because the company is - lets face it, a stereotype tech villain, most of these threads are populated by people who don't own the card, and wouldn't buy Nvidia products in general. Great for blowing off steam, but in the greater scheme of things means close to nothing.

When there are twice as many forum posts about (un)/locking overclocking of Nvidia mobile graphics cards than there is about the NSA secreting spyware in hard drive firmware across the whole planet, it's probably time to take stock of what kind of yardstick you measure wrongdoing by...IMO
Posted on Reply
Add your own comment
Jul 16th, 2024 11:28 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts