• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

RX 7900 XTX reference at possible vapor chamber design problem + very high hot spot (110c)

Status
Not open for further replies.
The amount of acrimony that has been created by the perceived focus on the negative aspects of a product that is 3 weeks old is worrying.

I don't think it's worrying. Worrying is the steep price for something which shouldn't be launched, in the first place.
This is a product that had to wait 2-3 months before its teething issues are ironed, and before then, there are more important cards to be launched - where the majority of sales is anyways - Radeon RX 7400 through Radeon RX 7800 XT.
We don't even know as of now if these will be launched or when.
 
Christ alive, what happened to turning over a new leaf in 2023? Y'all have had two friendly warnings from mods already in this thread.

If you still just can't resist the temptation to drag the thread off topic or attack others for their opinions against or supporting AMD, then I guess this thread can end here.
 
It can't just be because the "vapor chamber" can't handle the heat load. When the 6900 XT uses the same type of heatsink & has about the same thermal heat load is watts when overclocked.
 
It can't just be because the "vapor chamber" can't handle the heat load. When the 6900 XT uses the same type of heatsink & has about the same thermal heat load is watts when overclocked.
There is talk of other underlining issues too, but I digress.

The AIB cards obviously won't be an issue too, but they have the pesky problem of costing the same as the 4080.
 
It can't just be because the "vapor chamber" can't handle the heat load. When the 6900 XT uses the same type of heatsink & has about the same thermal heat load is watts when overclocked.
I think it's a design and engineering error. Just like Nvidia's lit 16 pin input.
 
I must have a good card because im able to play MW2 and my 7900 stays boosted >= 2500MHz the whole time. Hotspot is at about 110.c with the fan at 100%.
I'm also coming from a Radeon 7 which never got to it's target boost (target was 1800MHz but was always around 1750MHz) so i'm happy to see it stay boosted. I'm happy with my purschase.
 
The issue with the 4090's was user error from what I've read or am I wrong.
The full truth of the problem can only be available to us if we are very skilled engineers and have tested it ourselves with all the necessary measuring instruments. As long as we rely on information read somewhere, there is always the possibility of being misled.
 
I don't think it's worrying. Worrying is the steep price for something which shouldn't be launched, in the first place.
This is a product that had to wait 2-3 months before its teething issues are ironed, and before then, there are more important cards to be launched - where the majority of sales is anyways - Radeon RX 7400 through Radeon RX 7800 XT.
We don't even know as of now if these will be launched or when.
It is January 1st. Theorizing negatively about something that would belie historical information is also not good. When have you seen AMD launch the cards you are referencing at launch of a new Node? When you talk about price is it AMD or the Distributor that determines Street Price? Then there is the price so the retailer can pay it's employees too. I see people talking about the 4080 being the same price but in Canada there is no 4080 for $1399 or even $1599. It brings me back to the fact the cards are 3 weeks old. Was Vega smooth sailing from the begininng? The only GPU that did not have Gremlins at launch was Polaris because RDNA1 had a few problems that people tried to make into don't buy reasons. Outside of that if you look at raw Gaming performance in any review (biased or not) and you will see that there are not many cards faster than the reference card on the 7900XTX you won't be disappointed. I also have noticed that when people like to compare them the the 7900XTX has 24GB of VRAM and the 4080 16 so.

The full truth of the problem can only be available to us if we are very skilled engineers and have tested it ourselves with all the necessary measuring instruments. As long as we rely on information read somewhere, there is always the possibility of being misled.
The recommendation that everyone came to was to leave your side panel off.
 
I think it's a design and engineering error. Just like Nvidia's lit 16 pin input.

How so it's already been used on an older card, Like I mentioned the Heatsink is slightly larger than the ones on the RX 6900 XT. It works prefectly fine on those card ,even with overclocks. They would've already had issuses with those cards as they released the RX 6950XT with higher wattage & bascially the same cooler.
IMO, it's most likely because the Z-hight on the chip with the chiplet's is causing a a gap. He did not check to see if they where Flat or even on them. Chiplets are not all the exact same height, We've seen this with ryzen cpu's where they got the chiplet idea. Sometmes the IHS makes really poor contact with the one or two of the chiplets.

Moving the card from vertical mount to horizontal mount makes the thermal paste seal open up. Then moving it back it wouldn't change back, since the gap let in air to the space after the paste was heated the thinned out, &/or the thermal paste drops off the die/chiplets
 
It is January 1st.

RTX 4090 was launched in September, how many months ago?

Theorizing negatively about something that would belie historical information is also not good. When have you seen AMD launch the cards you are referencing at launch of a new Node?

Well, this is a good manufacturing practice which was used previously - it's called "a pipe cleaner". The logic is that the larger dies are more complex and do need an indeed mature process, while the so called pipe cleaners normally use much smaller dies and are easier to make - higher yields, etc.
Radeon HD 4770 was such a pipe cleaner.

The Payoff: How RV740 Saved Cypress - The RV870 Story: AMD Showing up to the Fight (anandtech.com)

For its first 40nm GPU, ATI chose the biggest die that made sense in its roadmap. That was the RV740 (Radeon HD 4770):

NVIDIA however picked a smaller die. While the RV740 was a 137mm2 GPU, NVIDIA’s first 40nm parts were the G210 and GT220 which measured 57mm2 and 100mm2. The G210 and GT220 were OEM-only for the first months of their life, and I’m guessing the G210 made up a good percentage of those orders. Note that it wasn’t until the release of the GeForce GT 240 that NVIDIA made a 40nm die equal in size to the RV740. The GT 240 came out in November 2009, while the Radeon HD 4770 (RV740) debuted in April 2009 - 7 months earlier.

When you talk about price is it AMD or the Distributor that determines Street Price? Then there is the price so the retailer can pay it's employees too. I see people talking about the 4080 being the same price but in Canada there is no 4080 for $1399 or even $1599. It brings me back to the fact the cards are 3 weeks old. Was Vega smooth sailing from the begininng? The only GPU that did not have Gremlins at launch was Polaris because RDNA1 had a few problems that people tried to make into don't buy reasons. Outside of that if you look at raw Gaming performance in any review (biased or not) and you will see that there are not many cards faster than the reference card on the 7900XTX you won't be disappointed. I also have noticed that when people like to compare them the the 7900XTX has 24GB of VRAM and the 4080 16 so.

Canada is only market, here in Europe RTX 4080 is cheaper than the RX 7900 XTX.

My point is very simple - when you pay 1400 euros for an RX 7900 XTX, you demand perfection and nothing else.
Next time AMD will have even less sales.
 
I don't see anything new, I have tried 4 6800XTs from different manufacturers and 2 6900XTs, all of them were trash. The hotspots were always reaching 100C on them, having my case side panel off/on, no change.

Winter here and having 20C+ in the room, I don't know the exact number. So during summer, when 35C is nothing unusual here, I would be reaching 110C+ on all of them and would throttle.

AMD GPUS are trash. My 3090RTX is colder when consuming 470W. Pathetic. This has been the continuous problem since 5700XT.
 
How so it's already been used on an older card, Like I mentioned the Heatsink is slightly larger than the ones on the RX 6900 XT. It works prefectly fine on those card ,even with overclocks. They would've already had issuses with those cards as they released the RX 6950XT with higher wattage & bascially the same cooler.
IMO, it's most likely because the Z-hight on the chip with the chiplet's is causing a a gap. He did not check to see if they where Flat or even on them. Chiplets are not all the exact same height, We've seen this with ryzen cpu's where they got the chiplet idea. Sometmes the IHS makes really poor contact with the one or two of the chiplets.

Moving the card from vertical mount to horizontal mount makes the thermal paste seal open up. Then moving it back it wouldn't change back, since the gap let in air to the space after the paste was heated the thinned out, &/or the thermal paste drops off the die/chiplets
Yes, I did not say that there is a problem with the cooler, as you said, there is a problem in the contact part of the cooler, as you said, the chiplet design may cause this.
But such a question had to be taken into account. Give 1000 dollars and such a problem. Companies want a lot of money. In this case, they need to offer a problem-free product to their customers.
 
RTX 4090 was launched in September, how many months ago?



Well, this is a good manufacturing practice which was used previously - it's called "a pipe cleaner". The logic is that the larger dies are more complex and do need an indeed mature process, while the so called pipe cleaners normally use much smaller dies and are easier to make - higher yields, etc.
Radeon HD 4770 was such a pipe cleaner.

The Payoff: How RV740 Saved Cypress - The RV870 Story: AMD Showing up to the Fight (anandtech.com)





Canada is only market, here in Europe RTX 4080 is cheaper than the RX 7900 XTX.

My point is very simple - when you pay 1400 euros for an RX 7900 XTX, you demand perfection and nothing else.
Next time AMD will have even less sales.
None of what we think matters at the end of the day but I know for at least 2 Generations from AMD and Nvidia that they have been giving the consumer the high end first and giving you the cut down nodes later in the manufacturing process. Specifically AMD usually releases an updated part that is refined at the hardware level with all of the previous Software Updates applied. Perfection is unachievable for humans otherwise we would have 100% success in everything we do. It doesn't matter what the price is as I would take 24GB over 16GB any day. If you have an older card and don't like the prices buy a 3070 or 6800XT and be happy. They will only get cheaper. If you are brave enough to buy a brand new GPU on brand new architecture you should also be sensitive to the fact that often there are Gremlins that Drivers often solve. Just look at the experience that is 6000 series or the 2080TI.
 
Guys,

This is just all part of the master plan to clear 30 series inventory :rolleyes:
 
I never had any issues with my reference 6900XT, but I'm definitely glad I went with a Red Devil for my 7900XTX.
 
I wont mind taking one over for cheap, was going to watercool it anyway.
 
I wont mind taking one over for cheap, was going to watercool it anyway.
that's what im hoping for as well. Massive price drop on these and Ill have to just get a vortical bracket :)
 
Any 7900XT users affected? Mine seems fine.
 
Any 7900XT users affected? Mine seems fine.

You got an MBA 7900XT? Cooler looks a bit different, one screw hole less around GPU.
 
You got an MBA 7900XT? Cooler looks a bit different, one screw hole less around GPU.

Yeah the shape of the vapor chamber looks different too, so it's probably not impacted (though, if it is, maybe not enough folks have bought one to notice?
 
Status
Not open for further replies.
Back
Top