28 replies to this topic

[cSand]

Class action suit against nVidia!

http://arstechnica.c...ard-ram-issues/

Edited by cSand, 23 February 2015 - 11:56 AM.

[Hawk819]

No surprise there. It was bound to happen. I mean, there were a lot of complaints on the 700 series cards as being underpowered and unable to do what Nvidia said they could do. This is no surprise to me.

I've got a GTX 760 4GB DDR5 card, and it's not up to par the way Nvidia says its suppose to be.

Edited by Hawk819, 23 February 2015 - 12:06 PM.

[Basskicker]

Pitiful always hearing how companies are cutting corners like this expecting people to not find out. It's a business, and the bottom line is to make money... but at what point do you sacrifice your integrety as a company to make a couple extra bucks?

[cSand]

AMD fan or no, this tweet was hilarious after the news came out:
https://twitter.com/...511204951855104

[Bill Lumbar]

I guess this is one good reason I have not put a Nvidia card in any of my rigs since my 8800GTX. Loved the card, but something about Nvidia just seems so dirty.

[xWiredx]

Honestly, this is bad but not nearly as bad as AMD's false advertising way back on it's original Phenom chips or its 2900XT GPUs yet there was no official class action suit for those that I can find. Us oldies are aware of the shortfalls in their performance after AMD touted major advantages over their competitors, though, and how AMD pulled the curtains in the sneakiest ways possible on those. There were unfixable hardware bugs that were not disclosed by AMD. There was a missing feature or two that was advertised on the 2900XT. No suits. Claiming one company or the other is dirty in comparison is pretty biased.

Also, the only real substantial part of the claim is the difference in ROP count. The card was not advertised ever to have complete access speed parity across all 4GB of VRAM. Usually people assume that is the case, but it was not advertised as such. The card can access all 4GB, it simply accesses the last 500MB slower than the rest (which technically savvy people know is not an issue while at 1080p). The difference in ROP count is somewhat disconcerting, though, and had I bought a 970 I probably would be a little angry about that.

Personally, I've not had a problem with either one. I've had a 9600XT, 3870x2, 2x 5770, 2x 6870 from ATI/AMD, Geforce 2 MX400, 9800GTX+, GTX 275, 2x GTX 660 Ti, and GTX 980 from Nvidia.

[cSand]

Neither company is exactly Mr Clean.


But hey, if you're an nVidia guy and you want to switch sides, AMD is giving discounts for 970 owners

[Flapdrol]

I went with nvidia because they generally score better in cpu limited benchmarks, and my pentium cpu doesn't have a lot of cpu time to spare. Turns out that's because of better multithreading, so I might as well have gone amd. 100 euro gtx 670 wasn't that bad of a deal though.

[Bill Lumbar]

I didn't say that Nvidia was "dirty", I simply said "seems" dirty. I have bought cards from both camps.... just no Nvidia cards lately. Not biased here at all...just calling a spade a spade. Amd's original Phenoms had false specs? There is a difference in a company claiming/bragging on their new line of upcoming processors being the ***** vs a company that mislabels the specs of the product, IMO. As long as the spec are correct.... one action just gets you laughed at after the release of your product if it doesn't perform as claimed, the other action can get you sued or so it would seem.

Please don't get me wrong, most companies are willing to throw all ethics out the window for "the almighty dollar" as their bottom line, some go even further then others do.

IF you are bringing up the B3 bug in the original Phenoms Amd released, yes it was a bug. There was a "fix" for it, just many didn't like the fix if I remember right, because it "lessened" performance even more on the chips if I remember right. This is one reason I didn't jump on board with upgrading my Amd 6400X2 BE when they first came out with the new quads. This is why I rarely jump on bleeding edge tech from any company, unless I can see the results first hand, before I drop my cash on it. This is why I got out of the hardware/have the best and baddest chasing game several years ago. I get some like to have the newest and greatest thing, best of the best, and have the money to spend on it. More power to them, to each their own. However, I have found it to be a never ending cash cow for the companies they spend their money with. After all, more is always better right?

Edited by Bill Lumbar, 23 February 2015 - 01:27 PM.

[xWiredx]

Yeah, off-topic, but AMD got nailed for showing performance numbers for Phenom chips that didn't exist on in-house hardware that didn't exist and did a whole black ops thing to take the info down just in the name of having a marketing edge over current Intel chips.

At least Nvidia said "yes, we definitely screwed up" after getting caught. Neither of them is particularly respectable, having manipulated benchmarks to favor them in the past as well as these current shenanigans.

[Bill Lumbar]

My only point was that showing performance numbers from benchmarks that don't add up, is very different then claiming specs on hardware that do not match the end product, IMO. One is kinda shady yes.... but the other is just... wrong. How is one to shop for hardware if the specs don't match the product?

I am one that takes ALL benchmarks with a grain of salt, and do not just buy the hype. I have no bias for any of the major companies that produce CPU's/GPU's but I do tend to root for the little guys as long as they produce something that works decent, and is priced decent as well.

Edited by Bill Lumbar, 23 February 2015 - 01:45 PM.

[Oderint dum Metuant]

The only part that has leverage is the ROP count.

The VRAM is a non issue, it has 4GB weather it has .5 that is accessed slower is irrelevant

[Thorqemada]

CACHE + ROP missing.
Memory Interface works at best at 224 Bit instead of 256 and at worst at 32 Bit instead of 256 Bit.

And they finally made People realize who this Jensun Guy is and how he behaves...a sect leader on top of a Computer Hardware Company imo.

[DarthPeanut]

For reference, if you bought a gtx970 from some retailers like newegg.. they are offering a full refund for these cards or a 10% gift code.

[-Natural Selection-]

I knew about it when I bought mine. The temps and performance of the card still makes it way worth the money. Once 4Kp is a thing then Ill worry about it, but until then Ill enjoy its blackmagic

[DarthPeanut]

Agreed. The gtx970 is still the best performing card for the money overall with it's only shortfall benchmarks being in the 4k benchmarks to the r9 290x. Running cooler, quieter, and with less power draw is nice as well.

[pbiggz]

*has an AMD card*



So i assume its true then that the GTX970 has trouble using all of its video memory?

Edited by pbiggz, 24 February 2015 - 12:49 PM.

[DarthPeanut]

Explains it in more in depth in the OP article. Basically it has 3.5 gb of high priority ddr5 ram and .5 gb of slower access speed low priority ram. So it does technically have 4 gb of ram but not all of it is as advertised ddr5. It can use it all 4 gb but after the 3.5 gb of ddr5 is used it begins using the slower .5 gb segment.

ETA: It still outguns everything short of 290x cards @ 4k.

Edited by DarthPeanut, 24 February 2015 - 01:47 PM.

[Bill Lumbar]

Am I mistaken, or are not some of the many complaints with the 970 gamers and pro developers that need the 4gbs of DDR5 as advertised running into stutter problems while using the card in some circumstances that have been shown to relate back to the false specs and the slower .5 ram? and or the lesser Rop?

[Sundervine]

Performance differences of the 980 and 970 first at 2688x1512 Very High which would use less than 3.5 gigs of memory. Second test was at3456x1944 which required more than 4 gigs of memory.

On GTX 980, Shadows of Mordor drops about 24% on GTX 980 and 25% on GTX 970, a 1% difference. On Battlefield 4, the drop is 47% on GTX 980 and 50% on GTX 970, a 3% difference. On CoD: AW, the drop is 41% on GTX 980 and 44% on GTX 970, a 3% difference. As you can see, there is very little change in the performance of the GTX 970 relative to GTX 980 on these games when it is using the 0.5GB segment.

Full article is at anandtech,http://anandtech.com...mory-allocation if you want to know more. It really is a non issue, the ROPS nor the slow ram make a real difference. This is really only about people that want to have something to complain about. For the money it really is the best card out there. False advertising is false advertising. I do not agree with it nor condone it, but this really might have been a mistake on the advertising team which really probably did not even know to look at the difference. I for one might have made this mistake so do not blame for more than not having the engineering department check the stats before the product was shipped.

Also I was until this last generation telling everyone to get AMD. Even my brother was told to get, and had AMD. AMD has had far more failures to communicate than Nvidia, including recently but not exclusive to their driver stuttering issue. http://anandtech.com...r-roadmap-fraps. Which was a much bigger problem than this and effected ALL AMD cards. Thus any AMD person claiming NVIDIA is bad has nothing to say at this time. There is a huge difference in things Nvidia has been doing for generations of cards, and a driver that damaged the performance all the cards.

P.s. It has a full 4 gigs of gddr5, all the same make and model chips. It however has 1 of 5 ROP controllers disabled. This has never been done before, by anyone that I am aware of. The disabling of part of a ROP controller instead of the whole thing. This is something new they are able to do now with their new design. Thus the reason since it has never been done before I can understand the problem in the advertising team. And the engineers have far more pressing things to do than to go and explain this to the marketing team that probably only half gets what is going on anyway. It is understandable with new technology, and yes this is completely new technology in its own right. Most do not understand the actual control this gives them, and the benefits we will reap from it.

I will try to explain. Before on almost every generation of both Nvidia and AMD chips, they had both a fast access ram cache and a slow access cache. This is nothing new at all, only the newly initiated was unaware of this. However before they had to fully disable a ROP that was bad. Yes when making a chip there are bad sections. Easiest way to explain it, AMD had three core processors remember. Wonder why? The chip had to many flaws to be able to thermally use all four cores. Disable one and it is now good to go. Thermals withing acceptable limits. Well a 970 and 980 are essentially the exact same card. The 970 just had more flaws, thus had to have some of it turned off. Before the whole ROP controller had to be disabled. Thus making the card have 48 ROPS instead of the current 56. A huge difference overall instead of a smaller difference. That means they will be able to use more of the more flawed chips than they used to. This allows for one of two things, either higher performance or lower cost. If they go the lower cost route they can use more of each platter, or they can go more performance and disable less of the chip than before. They sound the same but they are not. This one hickup for the process is only a good thing they will learn from it and their marketing team sure learned from it. Thus from now on they will be more honest on this front.

This news is all good no bad at all.

Edited by Sundervine, 24 February 2015 - 05:04 PM.