NVIDIA Kepler cards suffering from fatal flaw, could see recall

[cointoss]

---

It appears that TSMC and NVIDIA disagree whether or not NVIDIA's new Kepler architecture has a fatal flaw or not. On one had, we are hearing reports from TSMC that the new Kepler video card "chips may be suffering from serious performance degradation over long periods of heavy load" and could be the cause of a future recall by NVIDIA.**On the other hand, NVIDIA is saying all is fine. Bryan Del Rizzo, spokesman for NVIDIA stated, "There is no truth to this." NVIDIA denies that there will be any sort of recall over this report from TSMC. NVIDIA has, however, not provided any further details other than the previous statement denying that there is an issue.*EVGA has already had to recall all of their GTX 670 SC cards due to a hardware issue, so the idea of a recall could be correct. Is this the hardware fault that has caused the recall? Additionally, this could be the reason that the 670/680/690 are not in stock anywhere, even after TSMC promised NVIDIA a majority of its manufacturing resources.*A company can't keep a major hardware flaw secret for long as a mass amount of users would take to the internet with stories of the problems they are having. I imagine there is at least some truth to this report and it will come out soon enough. Let's not forget the bumpgate scandal where NVIDIA held off for as long as possible from admitting a design flaw to avoid compensating affected customers.*

Source: Tweaktown

Unsa kaha ni katinood

[lloyd_joy]

hhhmmm... basin "glassivation crack" phenomenon ni... weak ang die structure... ang mga metal layers mo collapse after the deposition process sa upper layer(s)... dili ni makita ug visual inspection lang kay ang crack, naa sa ubos na mga layers sa die... ayayayay... i'm just shooting the stars... hehehe... but this can happen though... as well as a handful of problem root cause contributors if ever it is a valid "Quality issue" on the wafer/die manufacturing process...

kani na case, ang wafer fab house (TSMC) naman ang ni divulge for a recall... kay kun dili nila ipahibaw ni sa Nvidia, balik man gihapon sa ilaha ang mga claims sa Customer's nila... dako kaayo sila (TSMC) mabayran ani kay process issue man nila ni kun naa gyud ha...

[jimmy128]

na hadlok man nuon ko dayon sa akong upgrade ><

[CrasHBURN]

I think this is not true. I'm using 2 GTX 680 in SLI. I have no problems. Only when using Adaptive VSync, mag stutter. But when using the regular Vsync, everything is smooth.

[mr.a]

Lantawa sa FreeBSD VuXML Documenting security issues in FreeBSD and the FreeBSD Ports Collection: CVE-2012-0946

[CrasHBURN]

About the Adaptive Vsync stuttering issue, nVidia is working on it and promised to release a new driver by next month. Though when using regular Vsync, I think its running like "adaptive vsync" already. Basically, no issues in my system.

mr.a, please explain further. I'm not sure if this is related to the thread.

[lloyd_joy]

if the performance of a semiconductor will decrease over time... its not a software issue (driver's issue) but, it's on the design or process when the semiconductor (GPU chip on this case) was made... if the claim by TSMC is valid, need gyud ni'g heads-up... "walking wounded" naman ang product (GPU) if ever the issue raised by the Wafer Fab (TSMC) is really true... kay mas naka hibaw sila sa process nila man gud and no driver update can solve the real "quality issue" (again, if this is really valid)... huwat lang gud ta'g updates from both Nvidia & TSMC...

pero, ako, lami gyud kaayo ang GTX680... its still my top GPU choice... agad lang gyud sa budget... hehehehe...

[lloyd_joy]

nangita ko'g write ups... sample ni... "Yield issue" man kuno... if yield issue, ang lots kay low yield during final testing, gamay ra good units na ma ship out. ang fall-out (failed units) kay i-FA (failure analysis) man gyud ni para masabtan unsa ang cause sa high failures... dri nakita sa TSMC ang problema... in-progress ang reliability test man nila pud so, nakita sa batteries of reliability test (accelerated testing man ni ang mga reliability tests) kun asa kutob ang life sa kepler chip. this is the basis sa TSMC on their claim kay kun dili sila mo sulti sa Nvidia, balikan man sila ani... pwerteng dako-a sa bayad gyud ani na case (Quality Issue)... labi na ang failure (Customer complaint), kay gikan sa end user like us.


TSMC 28nm Capacity Issues Causing Nvidia Kepler GPU Launch Delays

Nvidia Unsatisfied with TSMC, TSMC Gives Nvidia 28nm Priority | Custom PC Review

Nvidia Blames TSMC's 28nm Process Technology for Slow Sales - X-bit labs

[lloyd_joy]

aha, here's the article... kumbate Kepler... you're still no. 1!...

Nvidia Denies Plans to Recall GeForce GTX 600 Due to Performance Degradation - X-bit labs

[siopao1984]

history will repeat itself. mao na sa una nahitabo about those NVIDIA mobile GPU failure katong 2006-2008 na laptop models

poor soldering sa die ang root cause

read article here: Nvidia GPU Settlement Covers Apple, Dell, and HP

mao ni ang mga affected models

Apple MacBook Pro (15-inch) May 2007 – September 2008
Apple MacBook Pro (17-inch) May 2007 – September 2008
Dell Inspiron 1420 May 2007 – September 30, 2008
Dell Inspiron 9400/E1705 March 2006 – February 28, 2010
Dell Latitude D620 December 2005 – November 30, 2008
Dell Latitude D630 February 2007 – September 30, 2008
Dell Latitude D630c February 2007 – September 30, 2008
Dell Latitude D820 November 2005 – March 31, 2008
Dell Latitude D830 March 2007 – September 30, 2008
Dell Precision M2300 August 2007 – September 30, 2008
Dell Precision M4300 March 2007 – September 30, 2008
Dell Precision M6300 December 2007 – March 31, 2009
Dell Precision M65 November 2005 – January 31, 2008
Dell Vostro 1700 March 2007 – September 30, 2008
Dell Vostro Notebook 1310 December 2007 – September 30, 2008
Dell Vostro Notebook 1400 May 2007 – September 30, 2008
Dell Vostro Notebook 1510 December 2007 – September 30, 2008
Dell Vostro Notebook 1710 January 2008 – July 31, 2008
Dell XPS M1210 March 2006 – November 30, 2008
Dell XPS M1330 March 2007 – September 30, 2008
Dell XPS M1530 September 2007 – January 31, 2009
Dell XPS M1710 March 2006 – February 28, 2010
Dell XPS M1730 November 2007 – January 31, 2009
Dell XPS1710 March 2006 – April 30, 2009
HP Compaq Presario v30xx March 2006 – October 31, 2008
HP Compaq Presario v31xx March 2006 – October 31, 2008
HP Compaq Presario v32xx March 2006 – October 31, 2008
HP Compaq Presario v33xx March 2006 – October 31, 2008
HP Compaq Presario v34xx March 2006 – October 31, 2008
HP Compaq Presario v60xx May 2006 – October 31, 2008
HP Compaq Presario v61xx May 2006 – October 31, 2008
HP Compaq Presario v62xx May 2006 – October 31, 2008
HP Compaq Presario v63xx May 2006 – October 31, 2008
HP Compaq Presario v64xx May 2006 – October 31, 2008
HP Compaq Presario F500 October 2006 – September 30, 2008
HP Pavilion dv20xx March 2006 – December 31, 2008
HP Pavilion dv21xx March 2006 – December 31, 2008
HP Pavilion dv22xx March 2006 – December 31, 2008
HP Pavilion dv23xx March 2006 – December 31, 2008
HP Pavilion dv24xx March 2006 – December 31, 2008
HP Pavilion dv60xx May 2006 – April 30, 2009
HP Pavilion dv61xx May 2006 – April 30, 2009
HP Pavilion dv62xx May 2006 – April 30, 2009
HP Pavilion dv63xx May 2006 – April 30, 2009
HP Pavilion dv64xx May 2006 – April 30, 2009
HP Pavilion dv90xx May 2006 – November 30, 2008
HP Pavilion dv92xx May 2006 – November 30, 2008
HP Pavilion dv93xx May 2006 – November 30, 2008
HP Pavilion dv94xx May 2006 – November 30, 2008
HP Pavilion tx1xxx March 2007 – May 31, 2008

let's keep our fingers crossed.