RTX 3080

I am not purchasing anything until imports stabalise and the Rand recovers to a more desirable level. I think the 'pandemic' will still be around until March next year.

I have already seen some imports double in price.

Until then the 3080 is a SA wet dream. It will still sell though. We have many rich kids.
With one or many wives :unsure:
 
I see BZ is in the opinion that Igor may be right, but also that some AIB partners may have overestimated their boost clocks. He is also not certain that this may be the cause, but that Igor's discovery is plausible. All up in the air until Nvdia makes a statement I would say.

BZ also let out his pet peeve with Igor getting the cals wrong. I guess most people call these wrong in general. EVGA also got it wrong in their PR statement.

As it stands, the ASUS TUF is the AIB card to get to be reassured under what is known. BZ also mentioned that the Strix have another layout and cap implementation which questions why the TUF is kit close to best possible.

In BZ's argument, Nvidia is in the wrong say they didn't provide recommended spec and the partners is in the wrong when they have exceeded their own clock limitations.
 
I see BZ is in the opinion that Igor may be right, but also that some AIB partners may have overestimated their boost clocks. He is also not certain that this may be the cause, but that Igor's discovery is plausible. All up in the air until Nvdia makes a statement I would say.

BZ also let out his pet peeve with Igor getting the cals wrong. I guess most people call these wrong in general. EVGA also got it wrong in their PR statement.

As it stands, the ASUS TUF is the AIB card to get to be reassured under what is known. BZ also mentioned that the Strix have another layout and cap implementation which questions why the TUF is kit close to best possible.

In BZ's argument, Nvidia is in the wrong say they didn't provide recommended spec and the partners is in the wrong when they have exceeded their own clock limitations.
Not so sure it's as much to do with the clocks as it has to do with the capacitors they've chosen to use. If you watch the Jayztwocents vid he mentions that they eventually figured out that it was his microphone that was causing the cards to crash all the time - once he removed it, they were able to complete their testing. Apparently the frequency interfered with the capacitors. That's just piss-poor design. Note that this doesn't happen with reference cards, so fault is firmly with AIB partners.
 
Not so sure it's as much to do with the clocks as it has to do with the capacitors they've chosen to use. If you watch the Jayztwocents vid he mentions that they eventually figured out that it was his microphone that was causing the cards to crash all the time - once he removed it, they were able to complete their testing. Apparently the frequency interfered with the capacitors. That's just piss-poor design. Note that this doesn't happen with reference cards, so fault is firmly with AIB partners.

As Gamers Nexus tweeted,


I laughed how BZ pointed out it is SP-CAPS and not POSCAPS, a single article on the internet made everyone a 'expert' on capacitors. Strangely Igor said it right in his video, perhaps a translation issue with his article?

Nvidia still needs to provide the hardware spec the AIB partners have to base their design on. I am surprised this wasn't picked up like by EVGA which has a dedicated R&D department which does these studies like with Kingpin.

A sample issue tied to logistics may also be the culprit at play here. Nvidia don't sample every GPU in production.

I am sure throats are being slit.
 
While the whole cap thing is a bit of a stuff up, a lot of these cards cards are crashing past where even the manufacturers specced the boost clock, so not sure how much of a leg people have to stand on barring not following the tech spec provided by nvidia.
 
Not so sure it's as much to do with the clocks as it has to do with the capacitors they've chosen to use.

Frequency & noise go hand in hand. I've always hated RF circuits back in the day, your design might have been by the book but reality kicked you in the nuts. To this day I reckon RF design has elements of black magic...
 
As it stands, the ASUS TUF is the AIB card to get to be reassured under what is known. BZ also mentioned that the Strix have another layout and cap implementation which questions why the TUF is kit close to best possible.

Sorry to burst your bubble here, but there are TUF owners who reported crashing.

The Palit I have is stable as a rock. Not a single crash. Boosts way above 2000. I even upped the power to 109%, core 80+ and memory 500+ to reach 10ghz. Runs like a champ. A few guys over Carbonite also dont have issues with the palit.

I'm hoping for a custom bios to unlock the 350w power limit and see what this bad boy can really do.
 
  • Love
Reactions: Yuu
Sorry to burst your bubble here, but there are TUF owners who reported crashing.

The Palit I have is stable as a rock. Not a single crash. Boosts way above 2000. I even upped the power to 109%, core 80+ and memory 500+ to reach 10ghz. Runs like a champ. A few guys over Carbonite also dont have issues with the palit.

I'm hoping for a custom bios to unlock the 350w power limit and see what this bad boy can really do.

This is good news and I hope that you are right, there is a Reddit thread discussion the CAP counts,


and regard to the ASUS cards, according to the debaters some reviewers received a six count SP-CAP TUF cards, but I didn't see reviewers bring this up or experiencing any issues as the consumer base did. Hardware Unboxed received a 6 six count MLCC TUF to review, but as noted in the Reddit post,

Grapevine says that there are reports of instabilities on ASUS TUF and Strix cards as well. So 6x MLCC does not make you immune.

This being the internet, issues tend to be over exaggerated or trolled and the AMD crowd will surely add to the discussion, nonsense or otherwise, however this is a real concern.

I see the Palit Gaming Pro OC's are listed as:

3080: MLCC x 1 SP-CAPs x 5
3090: MLCC x 2 SP-CAPs x 4

and every ASUS card is listed as MLCC x 6, but Buildzoid said otherwise about the Strix. Like I said, this is the internet.

This does come back to my initial conclusion, this may very well be a sample issue (or inadequate sampling size) in which Nvidia takes part to approve their spec. Selective sampling may also be an issue, a rather big concern, as it would have been done to hide 'uncertainties'. Then it isn't a 'CAPs' issue per se.

Too much speculation going on at the moment, it is only Nvidia which can highlight the issue at hand. EVGA, by their own PR, admitted this being a POSCAP issue and their own post have been severely scrutinised since.

I am happy that you are in a happy place with your Palit.
 
This is why you never buy launch releases or pre-order... You're the first commercial beta testers paying a premium to do so.
Kind of mental, because all the scalpers weren't even selling the physical cards, they were just reselling their preorders. And people still paid through the nose for them.

No matter what tech you buy, there's generally always usually going to be issues with the first generation. AIBs learnt a hard lesson with this, because previous cards were built on nVidia boards that had already been thoroughly tested. This was just a shot in the dark for most of them I reckon.
 
I believe its way too early to be speculating why the cards are crashing, there are so many factors that could contribute. There are reports that these crashes could be linked to the power cable splitters that are being used to power the cards.

BTW i see wootware has some Palit 3080 pre orders available now.
 
Ew. Been sticking to Gigabyte for a few years now, let's see what they come up with.

Bah. I had the 1080ti GTX Aurous Extreme Edition...

Ran perfectly fine boosted on most games except PUBG.

Had to reduce the stock boost to get it to not crash...
 
Believe they just down tuned the cards to run below 2000mhz.
 
And now we just wait from some cards to actually arrive in SA
 
Top
Sign up to the MyBroadband newsletter
X