mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware > GPU Computing

Reply
 
Thread Tools
Old 2013-06-17, 15:59   #100
firejuggler
 
firejuggler's Avatar
 
Apr 2010
Over the rainbow

23×32×5×7 Posts
Default

Well i'm sure you won't complain if you receive a second GPU.
It's their mistake, not yours. However,should they contact you to get the second back, I wouldn't do any difficulties to send it back .
firejuggler is online now   Reply With Quote
Old 2013-06-18, 04:00   #101
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

217218 Posts
Default

Quote:
Originally Posted by kracker View Post
But I already got my replacement... I guess they got confused or something, I'll have to send them a message <snip>
DON'T! Google for "unsolicited goods iphone", for example.
That is kinda usual mistake the people in RMA do. It is not your fault, and any unsolicited goods have to be treated as unsolicited goods.
I know people who received a second mobile phone or second computer parts after they RMA and GOT the replacement already. Some tried to be "honest" and return the goods, it was a lot of headache... One of my friends was not able to return an iphone even after one year. In other cases, people at Samsung or whatever, didn't want to recognize their mistake, or they said they can't take the device back, and if they do, they don't know what to do with it. This is not a joke, I will try to search the web for it. If you want to be clear with your conscience, then you can write them a mail, thanking them so much, but under no circumstances return back the card, unless they expressly require it. If they say "sorry, it was a mistake, please return it", then it is a different story, but still, according with the laws, they have to pay the shipment. Otherwise, if you do, it will stay on a shelf for few years until someone finds it and asks about, eventually throw it away, or sent it back to you, because you know, the serial numbers are already registered in their DB as "RMA-ed to this or that customer", and it is a bigger pain in the ass for them to change all the paperwork. Who knows how these things work, knows what I am talking about.

Last fiddled with by LaurV on 2013-06-18 at 04:12
LaurV is online now   Reply With Quote
Old 2013-10-03, 01:30   #102
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

2·5,077 Posts
Default

I started this story over in the original Unhappy Me thread, when my (3rd RMA, I think) GTX 570 started heating up on account of one of three fans getting a sticky bearing. Gigabyte has now offered to send me a replacement fan, which suits me just fine. They have the idea that the fans on the card are different from each other in the number of wires connecting them. This turns out not to be the case.

I pulled all three, and they are identical three-wire fans. I took pictures of the three together, with legible model numbers, plus other details like the connectors. In the email back, I detailed the model number and specs. I also asked if, while we are at it, I could buy two spare fans, and possibly get a few spare mounting screws (two sizes, for the fans themselves, and for the shroud.) I figure I should at least get free shipping, and they might even throw in the fans. I'm still waiting for a response.

However, in the expectation of at least one replacement, and not having to send the unit in, I felt safe in attacking the dying fan. With a sharp punch, and a very small drill bit, I went at the plastic frame on the underside of the fan. I held the bit in my fingers and rotated the fan to get the bit to cut into the plastic. I very cautiously dug down to visible metal, and put a tiny amount of oil in the pit. After some working and turning the fan, it freed up somewhat, though not as much as the others. I then soaked a cotton swab in graphited, penetrating lock lubricant and blotted that into my excavation. This made still more improvement, so I cleaned up the residue and reapplied the label I had removed. I reassembled everything, stuck the card back in, and verified that all the fans were working. It is now back at work cranking out TF, at normal temperature.

I do look forward to having backup parts. I am delighted that Gigabyte offered that option. It is much easier for all concerned.

Last fiddled with by kladner on 2013-10-03 at 01:31
kladner is offline   Reply With Quote
Old 2013-10-04, 23:33   #103
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

11·157 Posts
Default

Very nice of Gigabyte to do that for you. GSkill does something similar with their memory. They'll send you new stuff before you have to send in the old stuff in so that in the mean-time you still have a semi-functional computer. I've got three sticks of 2GB and 2 sticks of 1GB 1333 MHz memory lying around but I might be tossing it out soon on account of I don't really want to put up with that kind of crap.

I'd rather be able to say I can't be without my RAM for very long so I'll send in the broken stuff when your fixed stuff comes in.
TheMawn is offline   Reply With Quote
Old 2013-10-05, 02:00   #104
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

2·5,077 Posts
Default

Still no follow up from Gigabyte. Somehow, I doubt they will want the old fan back, though.

On a side note, I am happy to report that the Asus GTX 580 has turned in a good DC with the VRAM running at its rated 2004 MHz, and the core clock at almost 50 MHz over its mild factory OC: 830 instead of 781 MHz.

The Gigabyte 570 still has to have the VRAM clocked down from 1900 to succeed. It failed at 1800, but I've had good runs with it a 1600 MHz. I hate to throw away 24 hours of work, more or less, just to see if I could nudge it up to 1700. It's kind of a shame as that card is supposed to have all hand-picked components.

Good on Asus, though!
kladner is offline   Reply With Quote
Old 2013-10-05, 06:10   #105
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

6BF16 Posts
Default

Cards geared toward gaming have the luxury of much higher tolerances for error in the VRAM. Even if one pixel turns bright red instead of green for one frame you'll never see it. That would be a pretty big memory error (relative to our tolerance for zero errors, of course) but it affects the gaming experience by virtually nothing at all.

I'd have to agree: Good on ASUS for making a card that can survive CudaLucas at almost-stock settings.
TheMawn is offline   Reply With Quote
Old 2013-10-06, 02:29   #106
apocalypse
 
Feb 2003

AE16 Posts
Default

I just had my 6-month-old Zotac GeForce GTX 570 release its magic smoke.

It seems like the computer overheated and shut down yesterday. I restarted it and updated CUDA to v. 5.5 for linux. When I restarted mfaktc, the card overheated and shut down the computer in seconds and I couldn't start it unless I replaced the video card. I dusted all the heat sinks and replaced the PSU with a new Coolmax 600W one, and as soon as I turned the computer on with the GTX 570 in place, it died in a flash of smoke.

Sadly, I didn't register for the warranty. I do wonder if something in the new CUDA driver contributed to its demise. In any case, I'll be out of the GPU computing race for a while.
apocalypse is offline   Reply With Quote
Old 2013-10-06, 03:34   #107
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

2·5,077 Posts
Default

I am very sorry to hear of your mishap. Do not assume that you are ineligible for warranty. Look at the Zotac web page and see what is needed to launch an RMA. Even if it seems that you may not have qualified, pursue it anyway. Be polite but persistent. Of course, it could help if you have a receipt, but even that depends on where you bought it. Most large retailers would have records of anything that recent, especially if a credit card is involved, and/or it is an online vendor.

As an encouraging example, let me mention that I participate a good deal in the Western Digital user forum. I have seen quite a number of cases in which a customer had a unit fail a few weeks out of warranty, but the company bent its own rules to accommodate.

It is really to a company's advantage to keep customers satisfied, and the cost is fairly minimal by their terms as they are recycling refurbished units.

I urge you to go for it!
kladner is offline   Reply With Quote
Old 2013-10-06, 03:56   #108
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

2·5,077 Posts
Default

Quote:
Originally Posted by TheMawn View Post
Cards geared toward gaming have the luxury of much higher tolerances for error in the VRAM. Even if one pixel turns bright red instead of green for one frame you'll never see it. That would be a pretty big memory error (relative to our tolerance for zero errors, of course) but it affects the gaming experience by virtually nothing at all.

I'd have to agree: Good on ASUS for making a card that can survive CudaLucas at almost-stock settings.
I have to say that those Direct CU Asus cards are physically built like tanks. It's nice to find that the electronic components also meet such standards.

To Gigabyte's credit, I do have to say that the first generation GTX 460 I have of theirs is a tough little beasty. The core of it runs good DCs at 830 MHz, and stock for it is 675. Factory OC is 715 MHz. The VRAM does fine at the stock 1800 MHz for CuLu.
kladner is offline   Reply With Quote
Old 2013-10-06, 23:47   #109
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

17×191 Posts
Default

Quote:
Originally Posted by James Heinrich View Post
For what it's worth, EVGA has good warranty policies. I had to RMA my GTX 570 last October, and I didn't have any hassles.
Unfortunately, almost exactly 1 year after my first GTX 570 died, the RMA replacement has just died.

Kind of a weird death though: for the last week or so the computer would randomly reboot once a day or so, no particular time or workload or anything, just reboot. Then today it just powered off and wouldn't come back on. After tinkering for a bit I found the computer would boot if I unplugged the 2x 6-pin PCI-E power cables (even with the GPU in the motherboard), but as soon as I plugged them back in the power supply wouldn't even turn on. Some kind of short, I guess. But no visible artifacting before it died.

Time for another RMA I guess...
James Heinrich is offline   Reply With Quote
Old 2013-10-07, 00:50   #110
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

2×5,077 Posts
Default

That's a strange set of symptoms. What comes to mind is something in the VRM assemblies. At least, an overload which causes the power supply to shut down in self-defense kind of suggests something in that area. This seems especially possible if nothing implicated the chip or memory, as artifacts might.

May the RMA be as quick and as painless as possible.

Last fiddled with by kladner on 2013-10-07 at 00:52 Reason: add good wishes
kladner is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Stockfish game: "Move 8 poll", not "move 3.14159 discussion" MooMoo2 Other Chess Games 5 2016-10-22 01:55
"Master" and "helper" threads Madpoo Software 0 2016-09-08 01:27
Aouessare-El Haddouchi-Essaaidi "test": "if Mp has no factor, it is prime!" wildrabbitt Miscellaneous Math 11 2015-03-06 08:17
Would Minimizing "iterations between results file" may reveal "is not prime" earlier? nitai1999 Software 7 2004-08-26 18:12

All times are UTC. The time now is 09:03.

Thu Jan 28 09:03:04 UTC 2021 up 56 days, 5:14, 0 users, load averages: 2.53, 2.45, 2.36

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.