• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

I won an artifacting 2070 Super

Status
Not open for further replies.
Associate
Joined
3 Jun 2013
Posts
148
Location
Hull
So over christmas I entered a few competitions on Hexus.net and found out on Monday that I had won an MSI 2070 Super Ventus OC (Yay!). The competition was run by Hexus but the prize was provided by and shipped to me by a system integrator based in the U.K. (not sure I can mention the name because they're technically an OcUK competitor).

I received the prize yesterday and as I had suspected, I didn't get shipped a retail card. The card (and only the card itself) arrived sandwiched between two polystyrene pieces that I assume were taken from retail boxes. They were taped together and shipped in one of those DPD expresspak packages. The card doesn't appear to have been damaged at all in transit, but I did notice that the protective covers on both the pci-e connector and the IO connectors had been removed, so the card has been tested in their factory if not used and returned.

Before finding out that I'd won the card, I'd ordered a g12 bracket and h55 in order to quieten down my aging 980ti, but had intended on now using it on the 2070 super instead. Before doing so, I obviously wanted to make sure the card was functioning correctly with the stock cooler so I decided to run a few benchmarks.

Unfortunately, at stock settings, I have been seeing pretty extreme levels of artifacting on both Unigine Valley and Superposition, though Heaven appears to run OK (at 1440p). Firestrike also artifacts, but firestrike extreme mostly just crashes, with high amounts of artifacting, flickering and just a black screen at parts where it should be rendering something.

While most of the issues are exclusive to benchmarks, I have also noticed bright and flickering pixels in some games, though it's much more difficult to reproduce these problems reliably. My first sign that something wasn't quite right was when I tried to run the 'Campaign' benchmark on TW:W2 and what should be a quite bright scene was very very dark, as if the shaders weren't functioning correctly.

Clearly the card isn't stable at stock clocks. I can remove *most* of the artifacting issues if I apply -100Mhz to the core, though the flickering doesn't stop unless I apply -200Mhz to the core. Perhaps I'm a bit out of the loop, but this isn't particularly normal is it? Having to apply such an aggressive underclock in order to make a card stable at stock settings?

Had I bought this card myself at retail I would 100% send it back and ask for either a replacement or a refund. The problem I have is that this is essentially a freebie. The cynical side of me suspects that this is a card that has been tested and determined to be not suitable for a customers system and so it has been offered as a competition prize instead, though I have no way of knowing that for sure.

I have the contact details of the person who organised the delivery of the card and so I could see if they can do something to help me, but suspect this might end up being a bit headache to deal with. Do I just accept the poor quality of the card and consider myself lucky I won a GPU at all? Or is it worth pursuing this and trying to get a card that isn't effectively malfunctional on day 1? I just don't want to be in a position where I slap the G12 + H55 on it (and so, I believe, void any warranty) and have the card die completely in a month - though I have to say, trying to RMA a card where I have no receipt and no retail packaging seems like a bit of a nightmare too!

What do you think would be the best course of action here?
 
Soldato
Joined
24 Oct 2005
Posts
16,277
Location
North East
Just contact hexus and tell them about the problem and see what they can do, either they may assist in a rma as in u send them the card and they rma it if they have the receipt or can get it from whoever supplied the card. Failing that, downclock the ram see if that helps or then the clocks. Failing that, try a bios flash maybe to a one that might work with lower stock clocks? Maybe the cooler has bad contacts with the thermal paste etc and a strip down and clean and reapply might work?

How many drivers have you tried, have you DDU between em etc as well.

Maybe needs latest windows 1909? Maybe needs a bios update, im not sure but worth checking perhaps.

Even if you downclock that -200mhz its still gona be really good.

If the packaging wasnt great maybe it developed some moisture or something and that might have done something? Clutching at straws here lol.

Tho you said you have contact info from the one who sent it i think, perhaps if it was sent special delivery it is covered under damage in transit tho proving it might be difficult given the poor way it was sent i.e unboxed as in not how it should have been. Unless the sender could prove it worked before sending it.

I still reckon a rma is possible if you try the serial number on it with the correct manufacturers site, some rma places i dont think require a reciept just the serial tho i could be wrong. Maybe ocuk has a rep from said company here u can call on to this topic for advice? Or just contact the makers directly with a email and describe you won it and its not working etc tho i might leave out the way it was sent to be on the safe side lol.

Could maybe try a diff pci slot?
 
Associate
OP
Joined
3 Jun 2013
Posts
148
Location
Hull
Sorry, I should have mentioned the troubleshooting steps I've taken already in the previous post.

I've placed the card in a system that has been running my 980ti flawlessly for 5 years. Since noticing the problems I've double checked pci-e connectors, swapped display ports on the card (don't have another DP cable to test, but I swapped back to my 980ti and it works just fine), reinstalled Nvidia drivers with DDU. Thermals on the card (at least the core) seem completely fine - during full load seeing high 60s low 70s. The card boosts to ~2Ghz at stock clock during benchmarking, so that seems to be operating normally too. After doing some googling, I'd found that some of the early RTX cards has problems with failing micron memory (which is what this card has) causing issues similar to what I am experiencing, but lowering the memory clock to it's minimum amount in afterburner makes no difference. The only thing that fixes the problem is the large core under clock.

Because I'm out of the loop I'm not sure exactly how these factory overclocked cards are supposed to function - but does having to under clock a card that is factory overclocked make that card faulty?
 
Associate
Joined
8 Apr 2003
Posts
1,181
Location
NFA
Card was effectively DOA if it was artifacting straight away. So yeah competition people need to replace it. I've had a nightmare with a 2080 I won that died after 10mths light use. Neither the manufacturer or the competition running (competitor) willing to take responsibility for a card that was less than year old. Won't put my business Zotacs way again after the rudeness of their support and the way Eurotech rma people responded.

Win something, sell it straight away to avoid toxicity.
 
Associate
OP
Joined
3 Jun 2013
Posts
148
Location
Hull
I'm sorry to hear that dude. It's a shame to have something lucky to happen only for it to become a massive headache.

I'm going to contact the company that sent me the card and see if I can get this rectified - if they don't play ball I'll send an email to Hexus themselves and see if they can sort the problem out on my behalf. I just wasn't sure if I was making a mountain out of a molehill, and whether sorting out a working card for me would be a bit too much to ask for giving that I'm not actually a customer of theirs.

Hopefully this will have a happy ending!
 
Soldato
Joined
24 Oct 2005
Posts
16,277
Location
North East
Might not fix it but did you try upping the power limit?

Only other thing i can think of is check if the gpu had a newer bios available that was made to fix the problem, could be worth a look? Btw do them cards have dual bios with a switch on the pcb?

As its a 20xx card msi afterburners OC Scanner might be a good test as it can test things for stability and tell you if theres limits and what they are, i think. Might be worth a run i think its safe but maybe ask others to be sure if its safe or helpful.
 
Last edited:
Soldato
Joined
2 Aug 2016
Posts
4,039
Location
Third Earth
I entered that competition! Congratulations! (I think)

Ultimately you entered a competition to win a working GPU, not a broken one. Therefore it’s entirely reasonable for you to expect a working GPU.

Expect nothing less!
 
Soldato
Joined
21 Dec 2019
Posts
6,439
Location
Planet Thanet
I entered that
Didn't win :(
Send it to me if don't want it :D
Seriously though hexus isn't some tiny little website
Contact them about it

Edit
And sure hexus and the not to be mentioned competitor wouldn't want bad social media coverage over something like this
 
Soldato
Joined
26 Oct 2010
Posts
2,981
Location
Leatherhead
Winning something that then turns out not to work really sucks. I’d have a look through the terms and conditions of the competition to see if there’s anything relevant you can quote at them. Surely they’re obliged to provide you with a working card?
I hope you get it sorted.
 
Associate
OP
Joined
3 Jun 2013
Posts
148
Location
Hull
Just been compiling some video evidence of the problems to link to in emails and thought I could post it here as well. I've noticed that if I run the Unigine benchmarks at 1080p the problems go away (though that isn't true of Firestrike), and if I lower the AA at 1440p to 2x the problems go away too. It seems to be only at very high GPU load that the core clock fails. Obviously that still means a defective GPU, but I'm concerned if I send it back they will test it at lower GPU loads, not encounter any artifacting and so assume that it's something else on my end. I can pretty consistently get artifacting to appear in TW:W2, though it's not especially severe (at the moment!) and it's difficult to replicate the exact same artifacting problem if I quit and reload. Most other games seem to be O.K. at the moment, but I don't really have many highly demanding games to test out. I'll link to some examples of the issues I've been having below, for those who are curious:

Unigine Valley:

https://www.youtube.com/watch?v=WaC1fFztEA8&feature=youtu.be

Firestrike (Extreme crashes and so messes with the recording, but it like this but worse):

https://www.youtube.com/watch?v=NGIVXD-hbNM&feature=youtu.be

TW:W2:

https://www.youtube.com/watch?v=PfFPJfEWugQ&feature=youtu.be

Shader issue that is difficult to reproduce in TW:W2:

https://imgur.com/a/F33ANTa
 
Associate
OP
Joined
3 Jun 2013
Posts
148
Location
Hull
Also, the Ventus appears to be MSI's "Budget" card, and so, for whatever reason, in MSI afterburner I can't increase the power limit at all. I did try upping the core voltage last night by +5mV at a time to see if that would stabilize the stock core clock, but got up to about half way on the slider in afterburner with no improvement and didn't want to push it any further.
 
Soldato
Joined
24 Oct 2005
Posts
16,277
Location
North East
Is there no way you can test with a different cable dp or hdmi, ive just found a few topics on google about coloured sparkles artifacts and it might be a cable.

Also maybe try the clean nvidia drivers without gf experience etc in it. Not sure the link to em as dunno site, but i think there mentioned where to get em in nvidia drivers topic last page or two.
 
Last edited:
Associate
OP
Joined
3 Jun 2013
Posts
148
Location
Hull
I have the ROG swift (the original TN version) and it only has a display port connection, so no possibility to check HDMI I'm afraid. I also don't have a spare DP cable either, but I've just swapped out the 2070 for my 980ti (using the same cable) and rerun the benchmarks and have no artifacting at all. Is it possible for a display port cable to function perfectly fine with one card but malfunction with another?
 
Associate
OP
Joined
3 Jun 2013
Posts
148
Location
Hull
I also feel that because I don't appear to suffer any artifacting on lower demand games, I doubt it's a problem with the cable? I ran Warframe for about an hour last night maxed out at 1440p and didn't seem to have any artifacting. I would have thought that if the cable was the issue if would also be present in lower demand titles? Now that I think of it, if it was a cable issue it wouldn't appear on any recordings I make would it? When I record with shadowplay it's recording the signal produced by the GPU itself right? And not what actually appears on my monitor?
 
Soldato
Joined
24 Oct 2005
Posts
16,277
Location
North East
I dunno on the specs but maybe the 980ti isnt pushing the cable to the max like the new card is cos its well its a lot better and pushing more bandwidth etc? Could be wrong like.

Do you have a tv with hdmi or something u can move the pc near to and try that? Like a living room tv?

Your probably right that its not the cable, just thought was worth a shot is all.
 
Status
Not open for further replies.
Back
Top Bottom