Bookmark and Share


UPDATE: Nvidia Denies Plans to Recall GeForce GTX 600 Due to Performance Degradation.

Nvidia Corp. reportedly plans to recall all the consumer desktop graphics cards based on the code-named Kepler architecture, including GeForce GTX models 670, 680 and 690. Apparently, Nvidia GK104 graphics processors suffer from performance degradation after usage under heavy load for long periods of time. reports that "chips may be suffering from serious performance degradation over long periods of heavy load". The exact effects of performance degradation are not indicated: it is unclear whether certain stream processors within GPUs stop working, or the chips drop clock-speed. As a result, the company is now rumored to initiate recall process for the GK104-based products. The information is not confirmed by Nvidia or makers of graphics cards.

The creditability of the information is unclear. In fact, it can be doubted as before making it to the market, all the new GPUs are not only vigorously tested in various applications, but also pass Microsoft HCL tests, which not just stress graphics chips, but take hours to complete. Therefore, chances that chips may degrade in performance over time are pretty slight.

Nvidia did not comment on the news-story.

Tags: Nvidia, Geforce, Kepler, GK104, 28nm


Comments currently: 32
Discussion started: 05/21/12 11:26:24 AM
Latest comment: 11/30/15 11:22:25 PM
Expand all threads | Collapse all threads


this a fail of epic proportions ... if it is true that is
4 1 [Posted by: veli05  | Date: 05/21/12 11:26:24 AM]
- collapse thread

It would be a fail if true. At least NV could simply refund the $$ to consumers to save face. However, all the poor investors who bought into the Facebook IPO can't get their $ back today!

I guess no price wars this generation.
6 4 [Posted by: BestJinjo  | Date: 05/21/12 12:12:50 PM]

This is not first time. If you remember, near 4 years ago there was a problem caused by material problem. G84, G86 and G92 series notebook GPU's was affected from this problem.
3 0 [Posted by: veleciraptor  | Date: 05/21/12 12:07:38 PM]
- collapse thread

Doesn't sound like the same type of issue though. The bumpgate disaster related to the cracks between the bump and the substrate of the chip, with results in a loss of signal and the chip dies completely.

This sounds like the chip's performance degrades under loads. The bumpgate issue was easily fixed but if this issue is true it could be related to the 28nm TSMC process for these chips couldn't be fixed easily until TSMC does a re-spin.

NV should consider Global Foundries as a supplier!
6 3 [Posted by: BestJinjo  | Date: 05/21/12 12:17:22 PM]

show the post
4 15 [Posted by: JanGozi  | Date: 05/21/12 12:18:07 PM]

Good job, NVIDIA. Sounds like there aren't any end users reporting this issue. Its good that NVIDIA stands behind their product like this and does a recall before all the consumers yell foul.

Its like the Intel Southbridge SATA port issue.
2 1 [Posted by: MonkRX  | Date: 05/21/12 01:04:18 PM]

This will really require some solid evidence from testing to confirm. Without such proof in hand, this is simply an interesting accusation. It could be true, but extraordinary claims require extraordinary proof.
2 2 [Posted by: Divide Overflow  | Date: 05/21/12 01:24:12 PM]

show the post
1 5 [Posted by: ato  | Date: 05/21/12 01:47:01 PM]
- collapse thread

Xbitlabs simply reported on a rumor. They never presented it as a fact. That's what the News section tends to have from time to time. Nothing wrong with that since it was reported as a rumor to begin with.

This is no different than Gamespot or IGN reporting on next generation Xbox or PS4 hardware rumors, etc.
9 2 [Posted by: BestJinjo  | Date: 05/21/12 02:00:19 PM]

One GPU supplier is reported to be recalling certain model Nvidia Kepler cards because "they did not receive the proper performance validation"...

This may be a function of Nvidia's enginering issues with Kepler resulting in horribly low yields. I think in a week or two Nvidia will be forced to admit they have issues with Kepler.

Several years ago Nvidia refused to acknowledge massive GPU failure issues until Charlie D. @ SemiAccurate was able to independently confirm the defects from numerous Nvidia card makers.
9 9 [Posted by: beenthere  | Date: 05/21/12 02:11:19 PM]
- collapse thread

show the post
0 9 [Posted by: kailrusha  | Date: 05/21/12 11:17:23 PM]
3 4 [Posted by: MyK  | Date: 05/22/12 03:58:12 AM]
That article is sensationalized. EVGA never issued an official recall for GTX670/680 cards. They only noted that users with cards that have issues can get a replacement in the form of a FTW card. The issues specifically related to early batches of EVGA cards.

EVGA has isolated this problem to an early batch of GTX 670 Superclock cards (P/N: 02G-P3-2672-KR) that were not properly screened during QA/QC procedure. Sounds like EVGA didn't do proper validation, not NV.

"EVGA has not recalled its 670 product line," as Joe Darwin with EVGA put it
6 2 [Posted by: BestJinjo  | Date: 05/22/12 07:12:26 AM]

mine is great. no problems whatsoever. It is evga so I never worry.
3 2 [Posted by: beck2448  | Date: 05/21/12 09:34:14 PM]
- collapse thread

^^^ "It's an EVGA so I never worry". You should see the link above then.
7 8 [Posted by: beenthere  | Date: 05/22/12 05:50:32 AM]
because their warranty and customer service is awesome, duh.
0 0 [Posted by: beck2448  | Date: 05/22/12 03:42:41 PM]

show the post
2 5 [Posted by: TAViX  | Date: 05/22/12 04:30:15 AM]

Let's straightened this up!
It's a common fact that overclocking of a GPU shortens common life cycle spin. Nvidias yield manufacturing layout is bad ( mooch more complaints about manufacturing problems then anyone else ). Kepler parts don't overclock very well. This are the facts.

Naw let's reconsider something else. GK104 whose intended to be a mainstream product. Bat when Nvidia sawed opportunity to even beat high end Amds GPU they pronounced it for a high end meanwhile real high end vent to Tesla part.
It's entirely possible that the clock speed of GK104 is already to high so it cud beat Amds flagship in that case degradation of GPU shud be slightly faster. On the other hand bad yield process cud case electrical interference between components that wold explained small performance gains from over clocking.

Until the degradation process don't exceeds warranty period Nvidia don't have to wary about it. Never the les if they miss calculated this they are in a big trouble.
This is the bad news for all premium manufacturers like EVGA, Gainward and etc because they will have to change they warranty policy's!
0 0 [Posted by: Zola  | Date: 05/22/12 04:37:45 AM]

Never one to take responsibility for their defective products... give Nvidia a few weeks to come up with some excuse why it's TSMC's fault, then they can officially start the recall.
7 9 [Posted by: beenthere  | Date: 05/22/12 05:54:02 AM]
- collapse thread

First you admitted to be an Intel hater constantly spewing that Intel bribes on a daily basis and now you are an NV hater because they had 1 issue with bumpgate?

What about AMD's 7900 series that promised us H.264 HD video decoding that doesn't work? Drivers that often crash, poor cross-fire scaling (basically unworkable Tri-Fire 7970), coil whine lottery, HD4800 series only getting quarterly driver updates? AMD raising prices across the board (HD6870 $230 --> HD7870 replacement for $350). You never bring these points up against AMD. Interesting.

Looks like you have been in denial as you have never criticized all the other AMD failures over the years: Phenom I, II, Llano, Bulldozer, 2900XT and 3800 series.

While Intel had 2 underwhelming CPUs in the form of Pentium 4 and Pentium D and NV had the disastrous 5800/5900 series, since then AMD has had a lot more poor products and yet you still keep supporting them no matter what, coming up with all kinds of "falsified fact" statements regarding Intel and NV to justify your position.

Intel doesn't need bribes. Since Core 2 Duo generation, they have produced a better product and that's why it sells. Unlike AMD though, Intel publicly admitted to the failure of Netburst and went back to the drawing board. In AMD's case Rory Read now starts telling us we don't need faster CPUs......I never knew RR became Steve Jobs, telling us what we need or don't need.
10 2 [Posted by: BestJinjo  | Date: 05/22/12 07:19:17 AM]
show the post
0 6 [Posted by: MyK  | Date: 05/22/12 09:11:18 AM]

BestJingo is trying to blow smoke up people's arse... again.

Reality don't change just because YOU try to SPIN it Bubba.

For the record Intel has been paying bribe money for over six months to OEMs to try and sell Ultra-expensive Books that are under-powered and over-priced by hundreds compared to AMD's laptops of equal or better mainstream performance.

Even with bribe money to lower the retail price, AMD's Llano and Trinity are far better laptops for mainstream consumers and that is why AMD Llano and now Trinity chips are selling faster than GloFo can produce the chippies.

AMD continues to take laptop market share from Intel because consumers are voting with their wallet instead of talking smack like some folks on X-Bits.

BTW, Ivy Bridge for laptop is a disappointment and WAY over-priced so Intel shot themselves in the foot with their first tri-gate transistor effort, overheating and minimal OC'in ability. Better luck next time or not?
7 9 [Posted by: beenthere  | Date: 05/22/12 09:04:37 AM]
- collapse thread

show the post
2 5 [Posted by: kailrusha  | Date: 05/22/12 07:54:57 PM]

Ok, let's think about it. Did anyone has the absolute proof 100% proof that the GK104 have this problem ?

I think it's some rumor, and needs to be a rumor, in my opinion the GK104 is fine.
3 0 [Posted by: CalinTM  | Date: 05/22/12 09:14:17 AM]
- collapse thread

THAT is why we should wait a few weeks to see what excuse Nvidia has and what evidence comes forth to substantiate the "rumors".
7 6 [Posted by: beenthere  | Date: 05/22/12 12:10:23 PM]

Not only xbitlabs ,GeForce 600 Series V-Sync Stuttering Issue?
Tom's Hardware has received several requests to look into a v-sync issue that a number of owners are seeing with Nvidia's new GeForce GTX 600-series graphics cards. A number of readers have pointed us in the direction of a very long forum thread where multiple customers express concern over stuttering problems with their Kepler-based graphics cards.
1 1 [Posted by: Blackcode  | Date: 05/23/12 02:45:03 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture