thebc2

howiejc

slovak_killerAgain, why this violent attitude towards EVGA ? As mentioned in this thread before, u got what u paid for, xoc bios is bonus. No its not reference design in fact its custom pcb, different IC used : uP9511R instead of reference MP2888A. Please just be patient and wait for official release of xoc bios rather than judging ftw3 card or even whole brand EVGA based on B E T A bios that is already in process of fixing.. Guys just be mature and dont direct your anger to this thread, its pointless, lets rather exchange statistics, infos etc and not hatred.

I have high standards. You expect a premium product when paying a premium price, especially if competitors offer something better for the same price. Also, that whole lack of MLCCs debacle which is clearly inferior to the reference design didn't help either.
Again, my fault for not thoroughly researching.


Ultimately, I really do hope EVGA proves me wrong and resolves this with future BIOS updates so that people can actually hit the 500W limit in reality.
 
Resorting to a shunt mod to avoid power limit throttling is pretty stupid if it's unnecessary and can be avoided.

100% this. At the end of the day, the FTW3 is more on par with a cheaper 3x8pin card like the gamingx trio, and Evga is charging you “Strix money” for a card a with inferior components and just pocketing the difference.

The Strix has smd capacitors for the vrm, MP2888As, and nothing but Mlcc capacitors behind the socket for the same price as the FTW3. Guess what other card has all of that? The new Kingpin.

bump
 

howiejc
I have high standards. You expect a premium product when paying a premium price, especially if competitors offer something better for the same price. Also, that whole lack of MLCCs debacle which is clearly inferior to the reference design didn't help either.
Again, my fault for not thoroughly researching.

 
The whole MLCC thing was blown out of proportion.  Pre release cards without MLCCs are still near the top of some of the overclocking leaderboards.  Its been shown that switching to MLCCs hasn't increased the stability of overclocks on cards that didn't have MLCCs to begin with.
 

howiejc
Ultimately, I really do hope EVGA proves me wrong and resolves this with future BIOS updates so that people can actually hit the 500W limit in reality.

Both the EVGA FTW3 Ultra 3090's I've had sustained 460-470 watts in Quake II RTX while playing the game.  At one point, I reset the meters in GPU-Z and played through a level in the game.  When I check after I finished the level the card averaged a little over 360 watts power draw with the 500 watt XOC BIOS.

I just talked to the administrator of this forum on phone support, they don’t have a second one....

willm1224
Can you help me figure out why this unattainable for my ftw3 ultra 3090? Nothing will get those numbers for me, I’ve flashed using nvflash as well as the update.exe my temps are low 60’s. I’m kind of at my wits end here

It's likely a power balance issue on the EVGA 3090 FTW Ultra. Look at the wattage being pulled from your PCIE slot. Mobo spec doesn't like more then 75w, you can pull more 80-90w safely but might destroy your mobo if it goes much higher. My card is around 80-90 and total board power peaks at 470ish with the BIOS. It's sad that this is a the case with 3x8 pin when zotac 2x8 pins are able to pull more, we got ripped off.

Ok, here is an update. I had to bypass all five fuses and remove all modded shunts (leaving the stock ones in place). At first I still was getting a red light on one of the PSU connectors. Took a closer look and could see that one of the fuses I had bypassed had part of the solder touching the fins of the cooler causing a short obviously. Promptly fixed the issue and card booted up and is alive. Adulation and jubilee turned to misery and pain as it was quickly realized that the card was in full gimp mode and refused to clock over 210 mhz.
 
Any ideas?
 

changboy
I want a waterblock for my card, WHERE ARE THEY ? Everyone have a WB but just us with the ftw3 dont have it lol arggggg.
 

You have to wait for Dec. (maybe Jan. 2021) from Watercool.de
Google-search for „WATERCOOL HEATKILLER GPU Compatibility pdf“
(sorry, i cant post links)
 
 
 

I get peaks of 460-480 in gameplay, not a constant 500w draw and avg 450-60 peaks 470-80 (quick) in timespy extreme. Pcie is at 80 w during said peaks (evga rep who administrates this forum sounded shocked, I’m like well it is what it is) 1 gpu pin peaked at 141 w averaged 135 w another 1 averaged 135w the third stays at 125 w

willm1224
I get peaks of 460-480 in gameplay, not a constant 500w draw and avg 450-60 peaks 470-80 (quick) in timespy extreme. Pcie is at 80 w during said peaks (evga rep who administrates this forum sounded shocked, I’m like well it is what it is) 1 gpu pin peaked at 141 w averaged 135 w another 1 averaged 135w the third stays at 125 w

GPU-Z doesn't report a max of anything higher than 440W for me on the XOC bios, even as an instantaneous peak. Not sure what I'm doing wrong. The third PCIe pin is always the lowest.

I have RMAd the card already. Didn't get any change in results.

bret0213

willm1224
I get peaks of 460-480 in gameplay, not a constant 500w draw and avg 450-60 peaks 470-80 (quick) in timespy extreme. Pcie is at 80 w during said peaks (evga rep who administrates this forum sounded shocked, I’m like well it is what it is) 1 gpu pin peaked at 141 w averaged 135 w another 1 averaged 135w the third stays at 125 w

We have been waiting for a response for 4-5 weeks now on this issue. I was told to RMA the card when I talked to a tech.

Ok, I guess I’ll just wait and see. As of now I just have to take what I have and hope it can be solved.if it requires an eventual rma (when it’s solved) then I’ll go that route.

Notchy44

howiejc

slovak_killerAgain, why this violent attitude towards EVGA ? As mentioned in this thread before, u got what u paid for, xoc bios is bonus. No its not reference design in fact its custom pcb, different IC used : uP9511R instead of reference MP2888A. Please just be patient and wait for official release of xoc bios rather than judging ftw3 card or even whole brand EVGA based on B E T A bios that is already in process of fixing.. Guys just be mature and dont direct your anger to this thread, its pointless, lets rather exchange statistics, infos etc and not hatred.

I have high standards. You expect a premium product when paying a premium price, especially if competitors offer something better for the same price. Also, that whole lack of MLCCs debacle which is clearly inferior to the reference design didn't help either.
Again, my fault for not thoroughly researching.
Ultimately, I really do hope EVGA proves me wrong and resolves this with future BIOS updates so that people can actually hit the 500W limit in reality.
Resorting to a shunt mod to avoid power limit throttling is pretty stupid if it's unnecessary and can be avoided.

Please don't listen to frame chasers with this whole you have to shunt mod your card to get the bios to work. My card is working perfectly fine with the bios as is a few other people on this forum. I bet a LOT more people are having success but they are not willing to take the time to post. This is a BETA bios so yes it still has time to mature. This card was not advertised with 500 watts and EVGA did not have to release this bios flash as the card performs above the ratings on the box STOCK.

I truly hope it's fixable because I highly doubt 100% of those who claim they are having problems are just trolls. It'd be stupid for Frame Chasers to post fake or incorrect data and ruin his reputation/brand, so it's likely his experience is a valid data point.
 
 
Now stock vs OC is a matter of perspective. Many serious gamers are well aware that they can safely push their premium product further, and they expect to be able to do so especially when the product is capable of it (aside from limits/locks from the AIB), otherwise they'd just buy a lower-priced/segmented item. The FTW3 Ultra 3090 is the highest tier air product aside from the Kingpin!!
 
If you are not one of those "more serious" gamers and always stick to stock, then you're simply not one of them. They consider a product's OC headroom as part of the product's appeal simply because they intend to OC, they like lower temps, like undervolting, etc.
 
Just because some people don't overclock, doesn't mean OC capabilities are completely irrelevant when there are those who do, and who may also eventually slap a water block on top of it later.
 
And let's not be so naive that EVGA didn't "have to" release a 500W BIOS. Of course they didn't have to if they did not want to compete with Asus. But clearly, they do.

cerealkeller
Shunt modding hides the power draw from the power monitoring software. So if you’re still power throttling, it’s going to show the same power draw in GPUz as it did prior to the mod

yes im aware my gpuz shows 350w pull for the gpu but im pulling over 900w on my outlet...

Kylearan
I scored this on a FE and PWR limit was not flagged at all, at +150/+600 and 100% voltage slider (unlocks one more "step" if boost wants to use it, which also may cause clocks to jump +15 mhz according to the curve).  Max TDP was 107% when MSI AB was set to 114% TDP.  Limited purely by clocks and MSVDD (no ability to adjust it on FE).
 
 

This is sad. My FTW3 Ultra on "500W" Bios with +200/+1200 gets around 14350. The PL is always flagged.

tt0ne
So apparently the card is not in gimp mode because in PrecisionX it says I have up to 255% power target and even though my card is running at 210 mhz the slider works and I can set my clock speed up to 1000 (the limit for PrecisionX - but I assume this is a GUI limitation that's all. My voltage slider works. Something is obviously wrong but it's not in gimp mode so I don't know what to call it. At this point I'm open to ideas and any instruction from those more experienced than me.
 

The card is in gimp mode because it's not boosting to its base clocks.  210 mhz is 2D clocks.  Can you run GPU-Z with a load on the GPU at 1000 mhz and post the "Max" values in GPU-Z wattage sensors, under a stress test like Heaven benchmark?  For all of the wattages.

Kylearan
The card is in gimp mode because it's not boosting to its base clocks.  210 mhz is 2D clocks.  Can you run GPU-Z with a load on the GPU at 1000 mhz and post the "Max" values in GPU-Z wattage sensors, under a stress test like Heaven benchmark?  For all of the wattages.

         Date        , GPU Clock [MHz] , Memory Clock [MHz] , GPU Temperature [°C] , Fan 1 Speed (%) [%] , Fan 1 Speed (RPM) [RPM] , Fan 2 Speed (%) [%] , Fan 2 Speed (RPM) [RPM] , Memory Used [MB] , GPU Load [%] , Memory Controller Load [%] , Video Engine Load [%] , Bus Interface Load [%] , Board Power Draw [W] , GPU Chip Power Draw [W] , MVDDC Power Draw [W] , PWR_SRC Power Draw [W] , PWR_SRC Voltage [V] , PCIe Slot Power [W] , PCIe Slot Voltage [V] , 8-Pin #1 Power [W] , 8-Pin #1 Voltage [V] , 8-Pin #2 Power [W] , 8-Pin #2 Voltage [V] , 8-Pin #3 Power [W] , 8-Pin #3 Voltage [V] , Power Consumption (%) [% TDP] , PerfCap Reason [] , GPU Voltage [V] , CPU Temperature [°C] , System Memory Used [MB] ,
2020-11-20 20:42:36 ,        1200.0   ,           1219.0   ,               31.0   ,               100   ,                  2988   ,               100   ,                  3008   ,           4306   ,          0   ,                        0   ,                   0   ,                    0   ,              463.8   ,                  24.4   ,              756.1   ,                428.6   ,              12.3   ,              36.0   ,                12.2   ,             29.4   ,               12.2   ,            382.4   ,               12.3   ,             16.0   ,               12.2   ,                       110.4   ,               1   ,        0.7370   ,               36.0   ,                  6862   ,

 
This was during a run on 4k Superposition.