bavor

Notchy44

bavor
I wonder if cards after a certain production date have a better chance of hitting close to 500 watts?

I have the 500W XOC BIOS on this card and I'm averaging 470+ average watts in Furmark and getting 460+ watts average in Quake II RTX

Well I can see your problem.... this is a oc bios and you have 0 overclock. Turn the core up to 150+ and mem up 850+ and I’m sure you will see a much higher average.

Actually, when I increase the voltage, GPU clocks and memory speed, the card draws less power in furmark.

willm1224

rgarodnick

willm1224
rgarodnick : Some other guy said it, Get the stock OC Bios on there and show them that you draw more than 75 W from pcie power during load, they’ll take it back then on a different premise. Although it’s the same premise, but worded differently.

Ah gotcha. So show > 75w on pcie slot on oc and <75w on xoc?

do the appeal thing with tons of frightening >80 pcie power I think that’s why they’re taking mine back that was probably in the notes from yesterday about my >75 watt pcie

Notchy44

bret0213
Those with 2014 that are not getting RMA approval. Put the original BIOS on and stress test the cards. See if PCIE goes above 75w, then it should get approved, I would think... Mine sits around 73-74 but still spikes above 80w. I know most are concerned with the OC power draw of 500w but the real problem is the power balance with the board pulling power from PCIE slot.

Im sorry but even at 85 watts that is not a issue for PCI e slot to handle. The issue is not this tbh, I know frame chasers WANTS to be right but I do not think it is the PCI E power load balancing.

bret0213

Notchy44

bret0213
Those with 2014 that are not getting RMA approval. Put the original BIOS on and stress test the cards. See if PCIE goes above 75w, then it should get approved, I would think... Mine sits around 73-74 but still spikes above 80w. I know most are concerned with the OC power draw of 500w but the real problem is the power balance with the board pulling power from PCIE slot.

Im sorry but even at 85 watts that is not a issue for PCI e slot to handle. The issue is not this tbh, I know frame chasers WANTS to be right but I do not think it is the PCI E power load balancing.

I understand your logic and most boards can safely allow more then 75w but that is out of spec. I for one use Riser cable and it is out of spec for those risers too. Speaking from experience, I had a faulty evga 2080ti black edition that would draw more 85-90w out of the PCIE slot. I ignored it for about 7-8 months because it was amazing silicon and I applied the same logic. In the end, I had to rma the card after it blew but lost my mobo and pcie riser too.

bret0213
I understand your logic and most boards can safely allow more then 75w but that is out of spec. I for one use Riser cable and it is out of spec for those risers too. Speaking from experience, I had a faulty evga 2080ti black edition that would draw more 85-90w out of the PCIE slot. I ignored it for about 7-8 months because it was amazing silicon and I applied the same logic. In the end, I had to rma the card after it blew but lost my mobo and pcie riser too.

arestavo

bret0213
I understand your logic and most boards can safely allow more then 75w but that is out of spec. I for one use Riser cable and it is out of spec for those risers too. Speaking from experience, I had a faulty evga 2080ti black edition that would draw more 85-90w out of the PCIE slot. I ignored it for about 7-8 months because it was amazing silicon and I applied the same logic. In the end, I had to rma the card after it blew but lost my mobo and pcie riser too.

 
Well, that tears it. I'm going to check out my stock OC VBIOS again tonight and see if it draws more than 75W. If it does, I will RMA - I don't want my motherboard (or anything else) to blow and take out the other expensive stuff.

Kylearan

slovak_killer

Kylearan
So, it's looking like not only are the "25" cards able to pull >500W, the 8 pin power connectors also have their own power limit of >161W!
But we still do not know what the PCIE slot limit is.  If it's 85W-90W, that's limiting the 8 pin.  If the 8 pin limit is 161W, then that's limiting the PCIE because the 8 pin is triggering the power limit.
This is also assuming the "TDP Limit" isn't triggering the power limit first.  Each individual rail has its own power limit.
 
So that means the card is technically capable hardware wise of pulling
 
161W * 3 + 85W=568W!  This assumes that we know what the 8 pin triggers its own power limit at.
I believe the FE cards trigger power limit at 150W at 100% and 170W at 112%.  I "believe" the PCIE slot limit is 68W.

I dont think youve seen this :D

How in Heaven's name did you manage this?  You're using the 1000W XOC Bios?

Kylearan

bavor
I checked my card.  It says made in Taiwan
PCV revision 0.1
With the factory BIOS it pulls 440-450 watts
With the 500W XOC BIOS, it pulls 460-470+ watts if I raise the power limits and leave the voltage and GPU clocks stock.
If I raise the voltage and GPU clocks it doesn't increase power draw and often the total power draw drops.
 
However something seems to be working right because my benchmark scores went up after installing the XOC BIOS.  Also, I have the #11 score on the Unigine Superposition 8K benchmark, #11 score on the Unigine Superposition 4K benchmark, and #13 score on the Unigine Superposition 1080p Extreme benchmark.  That's on the stock air cooler.
 
I have a feeling there is some issue other than a really simple hardware issue.  

This is simple, actually.
Remember that each rail has its own "power limit".
If a rail hits the power limit hard, it will prevent the other rails from going up, and the clocks will throttle harder.  If it's the PCIE rail triggering the power limit, the PCIE rail is the "least influential rail", so that being pushed back will have a much larger effect on everything if there is balancing going on, than an 8 pin rail hitting 180W and being pushed back.  
 
That is my guess.