Radeon released RX 480. (Card is Released)

[Tmdz]

No the driver or bios can not balance power distribution, they can only limit TDP and in that way get the PCI-e slot keep in line with 75w

For 6 or 8 pin to draw more power they must redesign VRM on PCB, but most customs cards from MSI, ASUS, Powercolor will do it and prolly use 8pin connector, even if they use 6pin their VRM design might use more power from the 6/8pin connector

If that is the case, there will not be overclocking of the 480 any more. They might limit the memory frequency to be below 2200MHz.

And that would mean lower and slower hashrates, they really needed 512bit memory instead of the 256bit...

[1715052430]

1715052430

[Auponef]

No the driver or bios can not balance power distribution, they can only limit TDP and in that way get the PCI-e slot keep in line with 75w

For 6 or 8 pin to draw more power they must redesign VRM on PCB, but most customs cards from MSI, ASUS, Powercolor will do it and prolly use 8pin connector, even if they use 6pin their VRM design might use more power from the 6/8pin connector

If that is the case, there will not be overclocking of the 480 any more. They might limit the memory frequency to be below 2200MHz.

And that would mean lower and slower hashrates, they really needed 512bit memory instead of the 256bit...

Compared to R9 390, which has 512 bit, RX 480 has only 256 bit memory bus, but the memory speed is faster.

[blg42598]

I see the power issues with the rx 480. Does using a powered riser solve this problem?

[reb0rn21]

Powered riser will solve 480x problem only if its TOP quality with connector which can provide 80W, and I can not answer which one will provide that!

Chines ones are not so good most you need to resolder +12V line better and ground to have better contact that can hold 80W

[giagge]

AMD comunicate to resolve the problem for more watt with upgrade .

[Sanguintan]

AMD comunicate to resolve the problem for more watt with upgrade .

The solution is additional power connections, either 8 pin or dual 6.

That is too late for the people who already bought the RX 480 cards. If you wait until the AIB out, there is not much to mine.

[spyshagg]

The 6 pin grounds are all grounded (one should be sensing, not grounded) on the reference rx480.  So in practice it can deliver the same power as the 8 pin.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

As such, when both the GPU and the full 8GB of ram are being pushed, it will overload the pci-e


This being a physical  circuitry problem, all the drivers can do is to throttle the card (decreasing performance). Or, if there is a margin, to throttle the gpu Voltage.

[Bemerand]

The 6 pin grounds are all grounded (one should be sensing, not grounded) on the reference rx480.  So in practice it can deliver the same power as the 8 pin.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

As such, when both the GPU and the full 8GB of ram are being pushed, it will overload the pci-e


This being a physical  circuitry problem, all the drivers can do is to throttle the card (decreasing performance). Or, if there is a margin, to throttle the gpu Voltage.

If the GPU  power is mainly supplied by the PCIE slots, then it is good news as we can under volt that.

[Charloz24]

Wow by reading all this, I feel lucky that I miss the boat buying rx 480. They were OOS really fast.

[spyshagg]

The 6 pin grounds are all grounded (one should be sensing, not grounded) on the reference rx480.  So in practice it can deliver the same power as the 8 pin.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

As such, when both the GPU and the full 8GB of ram are being pushed, it will overload the pci-e


This being a physical  circuitry problem, all the drivers can do is to throttle the card (decreasing performance). Or, if there is a margin, to throttle the gpu Voltage.

If the GPU  power is mainly supplied by the PCIE slots, then it is good news as we can under volt that.

The GPU is partially PCI-E. Its mostly 6pin

[Subw]

If the GPU  power is mainly supplied by the PCIE slots, then it is good news as we can under volt that.

no you can't.

[Subw]

Wow by reading all this, I feel lucky that I miss the boat buying rx 480. They were OOS really fast.

reference 480 is bad card for mining. i can't believe AMD failed it so much on so many points

[asdalani]

The 6 pin grounds are all grounded (one should be sensing, not grounded) on the reference rx480.  So in practice it can deliver the same power as the 8 pin.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

As such, when both the GPU and the full 8GB of ram are being pushed, it will overload the pci-e


This being a physical  circuitry problem, all the drivers can do is to throttle the card (decreasing performance). Or, if there is a margin, to throttle the gpu Voltage.

If the GPU  power is mainly supplied by the PCIE slots, then it is good news as we can under volt that.

The GPU is partially PCI-E. Its mostly 6pin

Even the partially PCIE, it has already killed so many mother boards. Most of the motherboards are quite old ones.

[dumpida]

The 6 pin grounds are all grounded (one should be sensing, not grounded) on the reference rx480.  So in practice it can deliver the same power as the 8 pin.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

As such, when both the GPU and the full 8GB of ram are being pushed, it will overload the pci-e


This being a physical  circuitry problem, all the drivers can do is to throttle the card (decreasing performance). Or, if there is a margin, to throttle the gpu Voltage.

If the GPU  power is mainly supplied by the PCIE slots, then it is good news as we can under volt that.

The GPU is partially PCI-E. Its mostly 6pin

Even the partially PCIE, it has already killed so many mother boards. Most of the motherboards are quite old ones.

These motherboards will not have warranty left. So that is a big loss to the miners and lost time in mining.

[Tmdz]

It probably was a cost saving measure in manufacturing where this design would cut costs by some tiny amount and thats why you end up with a poor design.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

I think you could also add the memory they use is clocked at a high speed and most likely running inefficient, pulling more power.  So instead using a wider bus on the interface they speed up the traffic.  Fine but then it seems they looked for an additional corner to cut to save costs, its like they said oh lets use 6 pin to save that fraction of a penny in manufacturing and then pull the rest from the mobo.

The electrical engineering guys want to build things right to the proper specs but then you have these assholes that look for corners to cut to increase profit.

[spyshagg]

the 6 PIN was a marketing play, really. Hence the "disguised" extra ground pin which makes it effectively an 8 Pin. They knew they needed the power.


Not sharing the gddr5 load between the pci-e and 6 pin was one mistake. The other was having two gpu phases pull power from the pci-e.


They screwed up, honestly.

[AlphaSun]

It probably was a cost saving measure in manufacturing where this design would cut costs by some tiny amount and thats why you end up with a poor design.

The problem is two fold:

1- AMD soldered the RAM vrm's to pull its power exclusively from the pci-e, without touching the 6 pin.
2- AMD hard soldered two pci-e power lanes to the GPU.  So instead of the gpu pulling all its power from the 6 pin, its also pulling from pci-e.

I think you could also add the memory they use is clocked at a high speed and most likely running inefficient, pulling more power.  So instead using a wider bus on the interface they speed up the traffic.  Fine but then it seems they looked for an additional corner to cut to save costs, its like they said oh lets use 6 pin to save that fraction of a penny in manufacturing and then pull the rest from the mobo.

The electrical engineering guys want to build things right to the proper specs but then you have these assholes that look for corners to cut to increase profit.

The memory are running too fast. That indicates they have high voltages. So the power consumption is high.

[generalt]

Wow by reading all this, I feel lucky that I miss the boat buying rx 480. They were OOS really fast.

reference 480 is bad card for mining. i can't believe AMD failed it so much on so many points

To be fair AMD's main focus isn't just mining.  They aren't Bitmain and I believe that their main concern is fps rather than hash rate.  Just like people are not going to invest in creating an asic for an algo that may be gone in six month AMD is not going to invest heavily into optimizing their video cards to mine coins.

[Zitdadast]

Wow by reading all this, I feel lucky that I miss the boat buying rx 480. They were OOS really fast.

reference 480 is bad card for mining. i can't believe AMD failed it so much on so many points

To be fair AMD's main focus isn't just mining.  They aren't Bitmain and I believe that their main concern is fps rather than hash rate.  Just like people are not going to invest in creating an asic for an algo that may be gone in six month AMD is not going to invest heavily into optimizing their video cards to mine coins.

I believe is also not for everyday gaming. The Ethereum mining does not stress the GPU like 3D marks.

[Ambros]

During stress test that measured 90W from pci-e, it's not difficult to believe in a motherboard literally burning..

That being said, AMD has announced a new driver that should fix the problem.