stevetb77
Thanks to the both of you... oops - how long is the registration window?   =D
 

atfrico

+1    I agree with Obscuretmpyre, in addition, don't let the registration window go by and expired. 

stevetb77
Gee, this has me not feeling so great about the 3090 XC3 Hydro Copper I have sitting in a box.  Maybe I will wait for FTW3 HC or KP.. maybe.
There's like zero reviews for this card on the interwebs.  Maybe I will make a thread to try to get some opinions from folks who own it.

90 days from the date of purchase, register the GPU ASAP or you will lose your RMA option

ty_ger07

atfrico

kongfra

ty_ger07

kongfra
Have you actually tried to RMA and see what EVGA says?  Maybe you just have bad card?

There seems to be many bad cards. Should everyone RMA their card at their own expense and hope the replacement performs as expected, or should EVGA be made aware of the issue and provide some other fix?

mine hits 348W in most demanding games like Red Dead Redemption 2 and Battlefield 5 without any overclocking or messing with any settings except the fan curve.  It also stays very cool 69-72 under full load and this is the XC3 base model

Good for you. You got a good card. You are one of a few who has reported such.

It's hit and miss and comes down to component manufacturing tolerances.

 
What I am saying is if the card is not within spec has anyone tried to RMA and see what EVGA says?

My understanding when a product is not performing within spec, owned for a short period of time, you are either entitled to a refund, or simply get another Brand New Item of the same product of purchase. Most companies do these, EVGA shouldn't be any different.

"New" is subject to availability. Shipping is partially at the owner's expense. And until EVGA acknowledges and determines the cause of the issue, there is no certainty that the replacement won't have the same problem.

That isn't a great option.

yes, "New" still available, EVGA still moving 3080Tis right?

https://forums.evga.com/Timespy-Score-3090-XC3-Ultra-Hybrid-m3186968.aspx
https://forums.evga.com/3090-XC3-Ultra-underperforming-in-timespy-m3101000.aspx
 
Looks like a very old problem, even 3090 XC3 having similar performance issues. Probably the 3090 XC3s were drawing less than advertised 350w and people weren't aware to it.  

atfrico

ty_ger07

atfrico

kongfra

ty_ger07

kongfra
Have you actually tried to RMA and see what EVGA says?  Maybe you just have bad card?

There seems to be many bad cards. Should everyone RMA their card at their own expense and hope the replacement performs as expected, or should EVGA be made aware of the issue and provide some other fix?

mine hits 348W in most demanding games like Red Dead Redemption 2 and Battlefield 5 without any overclocking or messing with any settings except the fan curve.  It also stays very cool 69-72 under full load and this is the XC3 base model

Good for you. You got a good card. You are one of a few who has reported such.

It's hit and miss and comes down to component manufacturing tolerances.

What I am saying is if the card is not within spec has anyone tried to RMA and see what EVGA says?

My understanding when a product is not performing within spec, owned for a short period of time, you are either entitled to a refund, or simply get another Brand New Item of the same product of purchase. Most companies do these, EVGA shouldn't be any different.

"New" is subject to availability. Shipping is partially at the owner's expense. And until EVGA acknowledges and determines the cause of the issue, there is no certainty that the replacement won't have the same problem.

That isn't a great option.

yes, "New" still available, EVGA still moving 3080Tis right?

No, I mean that it is EVGA's option depending on warehouse inventory. In the past, EVGA has sent out refurbs within the first 30 days, even though the card is still in production and is still "moving".

ObscureEmpyre

ty_ger07

atfrico

ty_ger07

atfrico

kongfra

ty_ger07

kongfra
Have you actually tried to RMA and see what EVGA says?  Maybe you just have bad card?

There seems to be many bad cards. Should everyone RMA their card at their own expense and hope the replacement performs as expected, or should EVGA be made aware of the issue and provide some other fix?

mine hits 348W in most demanding games like Red Dead Redemption 2 and Battlefield 5 without any overclocking or messing with any settings except the fan curve.  It also stays very cool 69-72 under full load and this is the XC3 base model

Good for you. You got a good card. You are one of a few who has reported such.

It's hit and miss and comes down to component manufacturing tolerances.

What I am saying is if the card is not within spec has anyone tried to RMA and see what EVGA says?

My understanding when a product is not performing within spec, owned for a short period of time, you are either entitled to a refund, or simply get another Brand New Item of the same product of purchase. Most companies do these, EVGA shouldn't be any different.

"New" is subject to availability. Shipping is partially at the owner's expense. And until EVGA acknowledges and determines the cause of the issue, there is no certainty that the replacement won't have the same problem.

That isn't a great option.

yes, "New" still available, EVGA still moving 3080Tis right?

No, I mean that it is EVGA's option depending on warehouse inventory. In the past, EVGA has sent out refurbs within the first 30 days, even though the card is still in production and is still "moving".

I’ve gotten both brand new and refurbished cards from EVGA through various RMAs over the years. I’d say there’s a strong likelihood that a new one would be sent considering the 3080 Ti isn’t quite two months old.

KingEngineRevUp
I posted this in the other thread with Kram (where Kram just kept ranting and not reading).

The problem might not be fixable with a firmware update. And I said might, not confirming anything.

1. Over at overclock.net we observed a lot of 3080 Ti we're not hitting their intended power limits

2. A user used 2 different external hardware methods to read the current being drawn and found GPU-Z power readings were quite different

3. It is possible the XC3 is drawing it's power limits but not what GPU-Z is reporting

4. Why this might not be fixable, if 3 is true, increasing more power might pop the fuses EVGA put to protect these cards

So it might not be possible to fix these cards with software, they might be stuck the way they are and that sucks, it would be pretty crap situation for the XC3 owners.

Wonder if there's someone that has means and method's of measuring their power draw from their hardware. Even if you have a PSU that can tell you PCI-E power draws. Just curious how they compare to GPU-Z.

The problem is the XC3 Hydro Copper is under water, and everyone who owed a Hydro Copper card would expect better performance across the board to be better than the GPUs running on Air.
I would just return the GPU, it is still functional, it is not broken. reason for returning it? simple, it doesn't perform on specs.
The other option: I would try sell it. There are many others out there who would like to get the GPU that Kram has at the moment

kram36
 
So you're telling everyone a different bios would not help the issue? May I present to you my results using a two 8 pin Zotac bios on the XC3 that allows for more power. Not great, but better then the EVGA bios.
 
Port Royal
 
https://www.3dmark.com/pr/1137081
 
Time Spy
 
https://www.3dmark.com/fs/25942832
 
Fire Strike Extreme
 
https://www.3dmark.com/spy/21652985
 
 

As I suspected, EVGA needs to release a revised bios then🤔.  How is the GPU performing Kram aside from the benchmarks

atfrico

kram36
 
So you're telling everyone a different bios would not help the issue? May I present to you my results using a two 8 pin Zotac bios on the XC3 that allows for more power. Not great, but better then the EVGA bios.
 
Port Royal
 
https://www.3dmark.com/pr/1137081
 
Time Spy
 
https://www.3dmark.com/fs/25942832
 
Fire Strike Extreme
 
https://www.3dmark.com/spy/21652985
 
 

As I suspected, EVGA needs to release a revised bios then🤔.  How is the GPU performing Kram aside from the benchmarks

I only put it on just to test those benchmarks.
 

KingEngineRevUp

kram36

KingEngineRevUp
I posted this in the other thread with Kram (where Kram just kept ranting and not reading).

The problem might not be fixable with a firmware update. And I said might, not confirming anything.

1. Over at overclock.net we observed a lot of 3080 Ti we're not hitting their intended power limits

2. A user used 2 different external hardware methods to read the current being drawn and found GPU-Z power readings were quite different

3. It is possible the XC3 is drawing it's power limits but not what GPU-Z is reporting

4. Why this might not be fixable, if 3 is true, increasing more power might pop the fuses EVGA put to protect these cards

So it might not be possible to fix these cards with software, they might be stuck the way they are and that sucks, it would be pretty crap situation for the XC3 owners.

Wonder if there's someone that has means and method's of measuring their power draw from their hardware. Even if you have a PSU that can tell you PCI-E power draws. Just curious how they compare to GPU-Z.

I read it and replied to it and then gsrcrxsi stated taking you to task on that post, so I did not interrupt your and his discussion from then on.
 
So you're telling everyone a different bios would not help the issue? May I present to you my results using a two 8 pin Zotac bios on the XC3 that allows for more power. Not great, but better then the EVGA bios.
 
Port Royal
 
https://www.3dmark.com/pr/1137081
 
Time Spy
 
https://www.3dmark.com/fs/25942832
 
Fire Strike Extreme
 
https://www.3dmark.com/spy/21652985
 
 

Again, all gsrcrxsi said was "I've used equipment too and I'm also an engineer." Never did he prove or do his own measurements to disprove the other engineers work.

And no, I said that there are fuses on the EVGA cards and we need to measure the power draw so if we do start finding alternative bios we don't blow those fuses and cause another mess.

The EVGA cards load balance is absolute crap. Pin 2 might be drawing more than Pin 1 on the XC3 like it is the FTW3.

XC3 owners might be in a worse situation than you already are and that's a bigger black eye on EVGA on these cards if true.

The reason some of this research came about was because the Galax bios gives the FTW3 card 1000W, so to safely use it, real power draw from cards should be measured to not risk blowing the fuses. Hardware measure power draw was higher than GPU-Z. Pin 2 was a lot higher than GPU-Z was reporting. So if anyone went off of GPU-Z, they would blow a fuse using the Galax bios.

How does this relate to the XC3, it's possible that EVGA screwed this up by setting a hardware limit on Pin 1 and 2 and they didn't load balance the pins correctly. When you're trying to draw more than 150W, Pin 2 might actually be drawing more than it should and it can be the reason you can't go above 300W because once Pin 2 might be hitting the stop button even though Pin 1 and PCI-E are still game to draw more power.

Sorry, I had to edit this post a few times because of mobile and browser keeps refreshing.

The max I have seen plug 1 is 132w and plug 2 was pulling 123.4w but that was only for a split second. Here is the info from that time.
 
PCIe Slot Power [W] , 8-Pin #1 Power [W] , 8-Pin #2 Power [W] , Power Consumption (%) [% TDP]

        67.2                         132.0                         123.4                                92.2
 
That was during a Time Spy run and GPU-Z said the PerfCap Reason PWR. Kind of odd since the max power I hit was only 92.2%

talon951

kram36

I read it and replied to it and then gsrcrxsi stated taking you to task on that post, so I did not interrupt your and his discussion from then on.
 
So you're telling everyone a different bios would not help the issue? May I present to you my results using a two 8 pin Zotac bios on the XC3 that allows for more power. Not great, but better then the EVGA bios.
 
Port Royal
 
https://www.3dmark.com/pr/1137081
 
Compared. Notice the higher average clock speed? https://www.3dmark.com/compare/pr/1137081/pr/1136708# 
 
Time Spy
 
https://www.3dmark.com/spy/21652985
 
Compared. Notice the higher average clock speed? https://www.3dmark.com/compare/spy/21652985/spy/21648694
 
Fire Strike Extreme
 
https://www.3dmark.com/fs/25942832
 
Compared. Notice the higher average clock speed? https://www.3dmark.com/compare/fs/25942832/fs/25941800
 
 

Interesting.

Did you get what appeared to be good power readings in GPUZ or were they obviously wrong and it was just apparent more power was being used based on clocks and scores?

I'm guessing messed up readings. I've assumed this is caused by the difference in controllers but have no real proof. The only bios out of a large number I tried on my FTW3 was one from a Palit card (same controller).

I didn't measure any readings during those test. I ran Time Spy first just to see what it would do and it still didn't beat the Stock FE card score from Jayz2Cents, I was bummed and just ran the other test to get the scores and clock speeds.

ty_ger07
Yes, sure.  But not guaranteed.  They strive to provide a user with a working replacement, but do not guarantee that it will be new.  That's my point.  I always provide this information to users so that they will not believe that it is a fact, and so that they will temper their expectations.  It is not a fact that they will receive a new card.  It has been proven to be not a fact by users experiences reported on this forum, and per the terms of the warranty.

Just imagine doing an RMA on one of these XC3 cards and receive back a XC3 card someone already did an RMA on because of the terrible performance. EVGA should guaranty anyone doing an RMA on one of these will get a new card, but EVGA hasn't even bothered to comment on the terrible performance, so I don't see EVGA doing an RMA on them right now.
 
Even a bios update to help out the terrible performance would be better than ignoring the issue.

Dante_Tel
So Kram, I took a look at your Motherboard as mentioned in your specs from these sites. Are you aware that your mobo only goes up to PCIe 3.0?

Do you have reason to believe that PCI-E 3.0 causes the power limit of the card to be reduced?  Otherwise, I fail to see how it could be related.

kram36

KingEngineRevUp
I posted this in the other thread with Kram (where Kram just kept ranting and not reading).

The problem might not be fixable with a firmware update. And I said might, not confirming anything.

1. Over at overclock.net we observed a lot of 3080 Ti we're not hitting their intended power limits

2. A user used 2 different external hardware methods to read the current being drawn and found GPU-Z power readings were quite different

3. It is possible the XC3 is drawing it's power limits but not what GPU-Z is reporting

4. Why this might not be fixable, if 3 is true, increasing more power might pop the fuses EVGA put to protect these cards

So it might not be possible to fix these cards with software, they might be stuck the way they are and that sucks, it would be pretty crap situation for the XC3 owners.

Wonder if there's someone that has means and method's of measuring their power draw from their hardware. Even if you have a PSU that can tell you PCI-E power draws. Just curious how they compare to GPU-Z.

I read it and replied to it and then gsrcrxsi stated taking you to task on that post, so I did not interrupt your and his discussion from then on.
 
So you're telling everyone a different bios would not help the issue? May I present to you my results using a two 8 pin Zotac bios on the XC3 that allows for more power. Not great, but better then the EVGA bios.
 
Port Royal
 
https://www.3dmark.com/pr/1137081
 
Compared. Notice the higher average clock speed? https://www.3dmark.com/compare/pr/1137081/pr/1136708# 
 
Time Spy
 
https://www.3dmark.com/spy/21652985
 
Compared. Notice the higher average clock speed? https://www.3dmark.com/compare/spy/21652985/spy/21648694
 
Fire Strike Extreme
 
https://www.3dmark.com/fs/25942832
 
Compared. Notice the higher average clock speed? https://www.3dmark.com/compare/fs/25942832/fs/25941800
 
 

Is your processor overclocked to ~5.0Ghz in some of those runs? There seems to be a difference between the two set of results and the cpu clock speed of ~500Mhz. Unless I'm looking at the clock speed incorrectly?

Dante_Tel

Aruzedragon

Dante_Tel
So Kram, I took a look at your Motherboard as mentioned in your specs from these sites. Are you aware that your mobo only goes up to PCIe 3.0? For max speed and performance these cards require no lower than the current PCIe 4.0 to run without hindrance. I mean, you can run a 4.0 card in a 3.0 slot but you're going to see varied levels of performance drop due to your slot not being up to spec. That's honestly my best guess after running all the other numbers and specs myself, the problem isn't the cards, it's your Motherboard not being up to spec.

I am having the same issue, and I have pcie 4.0 also. Not to mention, pcie 3.0 has more than enough bandwidth for gaming application. This isn't about pcie 3 vs 4 at all.

I mean I've checked his numbers and from what I've seen in the past (i.e. trying to socket a GPU in a system that isn't up to spec) it definitely matches with what I've seen before. This problem can be caused by any number of mismatches in the system though, not just the pcie slot itself.

I disagree.  In this instance, the rest of the system is inconsequential.  There is no configuration mismatch which is going to cause the card to power limit at 290 watts instead of 350 watts.  When the card says that it is power limited at a power consumption way lower than expected, that is the problem we are focusing on which isn't related to the rest of the system.

ty_ger07

Dante_Tel

Aruzedragon

Dante_Tel
So Kram, I took a look at your Motherboard as mentioned in your specs from these sites. Are you aware that your mobo only goes up to PCIe 3.0? For max speed and performance these cards require no lower than the current PCIe 4.0 to run without hindrance. I mean, you can run a 4.0 card in a 3.0 slot but you're going to see varied levels of performance drop due to your slot not being up to spec. That's honestly my best guess after running all the other numbers and specs myself, the problem isn't the cards, it's your Motherboard not being up to spec.

I am having the same issue, and I have pcie 4.0 also. Not to mention, pcie 3.0 has more than enough bandwidth for gaming application. This isn't about pcie 3 vs 4 at all.

I mean I've checked his numbers and from what I've seen in the past (i.e. trying to socket a GPU in a system that isn't up to spec) it definitely matches with what I've seen before. This problem can be caused by any number of mismatches in the system though, not just the pcie slot itself.

I disagree.  In this instance, the rest of the system is inconsequential.  There is no configuration mismatch which is going to cause the card to power limit at 290 watts instead of 350 watts.  When the card says that it is power limited at a power consumption way lower than expected, that is the problem we are focusing on which isn't related to the rest of the system.

I mean I'm not here to argue, I'm just going by what I've experienced and seen in the past. Worked back then for me. Had a rig a little while back that I ran an older Ryzen CPU in and it capped the mobo's specs below PCIe 3.0 (at that time), would not allow certain ram configs, wouldn't run the 1080ti up to par (wattage and overall performance was down). Upgraded the CPU to a Ryzen series 2 and maxed the mobo at 64gb ram proper and the very same GPU shot up to proper performance and stayed there. Like I said I'm not here to argue but still that is an example of what I experienced in the past, so from then on I never willfully ignored even the smallest detail on a build and I've been far happier for it to be fair.

kongfra

ty_ger07

kongfra
Have you actually tried to RMA and see what EVGA says?  Maybe you just have bad card?

There seems to be many bad cards. Should everyone RMA their card at their own expense and hope the replacement performs as expected, or should EVGA be made aware of the issue and provide some other fix?

mine hits 348W in most demanding games like Red Dead Redemption 2 and Battlefield 5 without any overclocking or messing with any settings except the fan curve.  It also stays very cool 69-72 under full load and this is the XC3 base model

Good for you. You got a good card. You are one of a few who has reported such.

It's hit and miss and comes down to component manufacturing tolerances.

 
What I am saying is if the card is not within spec has anyone tried to RMA and see what EVGA says?

I RMA’d my first 3080 Ti XC3 HC. But not for the low performance issue. The official RMA reason was the RGB stopped working completely. So it was a bit of an open and shut case. I paid (2x MSRP) collateral for a cross-ship RMA, and got a brand new replacement.

My first card DID have the low power and low clock issues. And it was quite severe. FurMark would only pull 270W and clocks under 1000MHz. Most other tests had low power draw and low clocks, very often less than the 1725MHz advertised clocks. Before the RGB failed, I did attempt to troubleshoot the issue with EVGA, but suffice to say, EVGA’s tech support team seems like a bunch of regulars just reading a script. They know as much about their own product as any random person on Reddit. They didn’t have access to their own BIOS files to send me to try to reflash it, and they couldn’t offer any explanations about why it was acting the way it is. Just the boiler plate stuff, what is the system specs, what are the voltages, etc. they had no interest in escalating it to actual technicians.

Being able to have a valid reason for RMA was a bit of a godsend. The replacement card acts much better (at least pulls 300W in FurMark, and higher power draw and clocks across the board) but it’s still not acting the way it should, and exhibits strange behavior that honestly it shouldn’t be, lower enforced power limit with increased mem controller load.

But I’m not going to RMA it a second time just for this reason. I can only hope they admit the issue and try to fix it via BIOS.