For those trying to build a budget miner or gaming pc, NEVER EVER buy the 7950 from Sapphire. I've had 6 cards out of 10 blow up. And no, no overclocking at all :') I've atleast been able to trace the fault... memory. The ones with Elpida are 10x more likely to blow up. At first the PC would just shutdown, and would not start. Removed the card et voila! boot! After testing them in another rig I almost instantly blew them up, turned out the memory fried itself. Sadly no pictures anymore, SD card died 3 days ago I'm not saying all Sapphires are bad, not at all. It's just the 7950 ones (later batches) that should be avoided. Can anyone confirm this with the 280's from Sapphire and/or others? Thanks in advance!
They blew up did they? Funny, as myself and several others have been running Sapphire built 7950's for well over a year without problems. Honestly, this post stinks of troll. A few more details on your 'problems' would probably alleviate that.
I have 2 sapphire hd 7950 in my rig running rock solid. I've also owned quite a few sapphire cards the past, never had a single problem. I would suggest you look into what made 6 out of 10 cards go pop, that is not normal and there's got to be a common denominator.
Mine is 7970 and only problem I have are coil whine (first time ever to hear it from a GPU so it bugs the hell out of me) and various AMD driver related crashes that have been fixed over time.
Sorry if I made it look like I was trolling. I wasn't. And for those wanting the whole story: I build 3 miners for some friends, using 7950's from Sapphire. I COULD provide some batch numbers (if I knew where to look...). First rig was made from a 990FXA-UD5, 8GB DDR3 1100T and a 1200w be quiet! Dark Power Pro 10. And finally: 3 7950's from sapphire. I ordered them from our local shop where I work. After just two weeks he came back and said that the PC wouldn't start. I decided to visit him that same evening (he is a very old friend, we've been friends since the age of 4). Removed the cards, put them in until it wouldn't start, tested the faulty card in my rig, same thing. After coming to the conclusion that I would simply RMA it my friend called saying the SECOND card died. He tested it with the very same method, and without the faulty card it booted right away. So I put in the second card in my rig and yipes! Nasty smell. One of the memory chips on the right "blew up". Sadly I haven't got the pictures anymore. And to be very clear, I nor my friend overclocked it. Not even for fun PSU was checked, motherboard was checked everything was working just fine. Second rig was for my colleague who found the idea of mining very exciting, so be it. I build him a nice rig, using a ASRock FM2A88X Extreme4+, 5400K and 2 7950's and a XFX 850W Core Edition. He would later upgrade to 3 cards. Same thing, same methods and same ending. 3 cards in total DEAD. Last one was this morning. All those cards have been sent back and we got new ones instead. And those too blew up (one of the new cards was among the 2 for the second PC). The third one was for our CEO, tech savvy guy :') For him it was just an experiment, and trying various things. Build him a nice PC using 990FXA-UD5, 1200w be quiet! Dark Power Pro 10, 3 7950's, 16GB and an FX6300. This wasn't a dedicated miner. Same thing, parts were good, cards not. 2 cards were replaced. I tested EACH faulty card in my own rig, and most of them blew up one or more of their memory chips. Elpida memory, 50F chips. I might be able to post some pictures when the dead card from this morning arrives. Our cards come from Wave, located in the Netherlands which is the same as Alternate. Which is a top notch supplier and retailer with excelent service! Won awards almost 10 years in a row :') Which I can confirm. And for those who are still running their cards, could you provide any numbers? Like batch numbers? So if any info is available please help
If defect rate in someones hands is significantly higher than average defect rate for same product in general owners population, problem is very likely in that particular user. I by chance own Sapphire one too, things I did to that card are from other world even for most of users here. And that card lives, and lives, and lives. You may check PSUs for defect and VccSA and PCIe voltages in BIOS as possible cause for killing blow.
Well that was my initial thought too, but things don't simply add up. 3 different systems, all other parts working perfectly fine including all the voltages. I know these people, I trust them when they tell me they didn't OC. As a matter of fact, I helped them set up their systems :') As said before, my CEO/friend is tech savvy, not that much in the hardware department though but still... He too confirmed that everything was as it should be. This has been going on since the start of 2014, I've been searching for an answer for three whole months now... Which is why I'm posting
Maybe your friend should look for a better pc builder then. Looks like you're just bad with hardware. J/k, strange problem indeed.
What about air flow then? If those systems were for mining, not everyone want his system to be loud and additional fans eat energy too. And I know miners are crazy about every watt... Honestly isn't it possible they reduced rpm even on those cards by setting manually some fixed speed? I have modified bios in which I have custom fan curve (min50% as accelero is pretty silent), power target, clocks. So i do not have to care about Power management in CCC.
Haha, at first sight it nearly gave me a heart attack :') Yeah I know about the noise, fans, watts and how much they care about it. I took an old server I had lying around, which is a windtunnel, so they didn't get that hot. About 65 degrees celcius, and from experience I can say that these chips don't really like 75 or above. I did want to mod the BIOS, only warranty kept me from doing it Others were in well ventilated cases Power target was set to +20, stock clocks and fan speed varried among builds. The one from my CEO was in our server room so nice cool and clean. Second (colleague) was in his guest room, about 18 degrees re reckon. The one my friend had was in his room, about 20 degrees. I even measured the temps with an IR gun :S
sapphire and MSI is a no no in my list, however I may buy them for customers who want a standard non-overclocking pc. The price is cheap so do the component. I have experience sapphire graphic card cannot be OC'ed at all.
Well you don't want this batch, even if they are meant for stock :') They were all stock... Just asked my colleague if I could borrow his "dead" card, I'm going to post some pics... Not sure about putting another one in my test rig... And in my experience Sapphire is fine, and MSI too. I've had all of the so called "better" better manufacturers and they still went down. It doesn't matter which which company you buy from, there's always that one sh*tty card
Wtf?! That's some awful luck! Was it something like this? http://www.youtube.com/watch?v=Ggh2Mu4Qkgk I've had a 7970 Dual-x for 2 years and a 7970 Vapor-x for 10 months without problems...so far. c:
Strange i had 2 sapphire 7870ghz editions and i overclocked the crape out of them for a year and they kept running and are still running in a friend of mine his pc,i think sapphire is more like one of the best brands to get for amd cards,i had a 5870 and 4890toxic and they all stil run after lots of usage.
Never seen a harddrive go up like that before. My MSI 7950 has Elpida memory on it and it had no issues running F@H 24/7 for a few months at 1200/1400.....
Well as I mentioned before, it's probably only this specific 7950 batch. Yes, I've had other cards from Sapphire that needed an RMA (TORTURE! ) but that also goes for other brands. I was hoping anybody could share some info on the Sapphire 280, or 280X perhaps, something more specific like memory used and other components ^^ And no the cards just made a KRGHRGRHGGHHH sound and started smelling... like *(@&$() :3eyes: And I believe the cards runs fine Sykozis, might be the earlier batch using 60F chips 60F translates to 6000:4=1500MHz for those wondering. My paranoid "theory" is that Sapphire just loaded a truckload of 7950's from cheapest components possible so they could ship the last chips. Making as much "profit" as possible. It isn't weird to see that happen in the industry.
cryptocurrency mining is no more stressful than F@H is....and there's a few F@H team members running Sapphire cards.
I know, I even tried having them fold for us But in the end it's all about the coins :3eyes: I hope they don't experience the same, it's frustrating. Those were not my cards, okay. But I DID recommend them, based on earlier info. I do not doubt Sapphire, just this batch. Anyway anyhow I'm off to bed, got a CentOs server waiting for me at 7ish in the morning c:
I had a sapphire 7950, too and it worked just fine for 2 years.. I just bought a sapphire r9 280x now and i have random black screens. I sent to sapphire explaining, and they told me to go to the reseller for hardware check and RMA.. Why u say it is "torture". Did u have a bad experience?
Well I don't know about RMA procedures in Greece (guess the same) but that sounds right. Your supposed to contact your retailer first, after all you bought the card from them With torture I meant that previous cards went through hell :') Like my old TNT2 Ultra, running at 185-210.
