kevinc313
CLASSIFIED ULTRA Member
- Total Posts : 5004
- Reward points : 0
- Joined: 2/28/2019
- Status: offline
- Ribbons : 22
Thursday, August 26, 2021 2:27 PM
(permalink)
Many people have reported 3080 Ti FTW3's doing about 425w in software reported power draw at max PL, when the bios is supposedly 450w. Had the opportunity to test a 3080 Ti FTW3 against a 3080 Ti Strix in the same system on the same loads, also while observing power draw at the wall with a kill-a-watt meter. Per my observations and calculations the 3080 Ti FTW3 under reports it's power by about 7%, so at 100% power it is reporting 395w but actually drawing about 425w. When you are at max power the card is drawing the ~450w it is supposed to, but only reporting ~415w-425w. 425w is an awful lot of power to be trying to keep cool, for a card right out of the box. The Strix on the other hand is pretty accurate in it's power reporting, but has 10-15C worse in it's vram temps.
post edited by kevinc313 - Thursday, August 26, 2021 2:34 PM
|
Cool GTX
EVGA Forum Moderator
- Total Posts : 31353
- Reward points : 0
- Joined: 12/12/2010
- Location: Folding for the Greater Good
- Status: offline
- Ribbons : 123
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:11 PM
(permalink)
What were your observed kill-a-watt meter readings ?
The "under reporting" --> is it the same on X1 & GPU-Z ?
Learn your way around the EVGA Forums, Rules & limits on new accounts Ultimate Self-Starter Thread For New Members
I am a Volunteer Moderator - not an EVGA employee
Older RIG projects RTX Project Nibbler
When someone does not use reason to reach their conclusion in the first place; you can't use reason to convince them otherwise!
|
kevinc313
CLASSIFIED ULTRA Member
- Total Posts : 5004
- Reward points : 0
- Joined: 2/28/2019
- Status: offline
- Ribbons : 22
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:26 PM
(permalink)
Cool GTX
What were your observed kill-a-watt meter readings ?
The "under reporting" --> is it the same on X1 & GPU-Z ?
Sorry I don't have my notes handy at the moment. iirc it was doing 540w at the wall at 100% / gpu 395w software. 92% psu efficiency, 70w for system and other GPU . So about 425w actual. When you compared with the Asus, at a variety of power level settings, the ASUS was drawing very noticeably less power. So it's not just the Kill-a-watt, it's supported by the other GPU's readings. To be clear, the card is drawing 7% more board power than is reported by software. Reading board power in HWiNFO.
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:32 PM
(permalink)
Talon also did the same test as you. He used a clamp meter and a ups. Similar results. https://www.overclock.net....1791823/post-28835729He saw that GPU-Z was wrong. My PSU let's me measure power draw, I'll check later myself as well.
post edited by KingEngineRevUp - Thursday, August 26, 2021 3:37 PM
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:34 PM
(permalink)
|
kevinc313
CLASSIFIED ULTRA Member
- Total Posts : 5004
- Reward points : 0
- Joined: 2/28/2019
- Status: offline
- Ribbons : 22
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:37 PM
(permalink)
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Thanks, I saw that a while back and the overall underreport didn't sink in with me, was more focused on the #2 underreport. People should be aware that it's not just the max setting though, it's across the entire range. Even at 50% power it's underreporting by a lot. And especially concerning at stock settings. If you daisychained #1 and #2 like a noob and ran stock settings, you could easily exceed 288w on one cable.
post edited by kevinc313 - Thursday, August 26, 2021 5:12 PM
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:38 PM
(permalink)
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his mind again.
post edited by the_Scarlet_one - Thursday, August 26, 2021 5:15 PM
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 3:42 PM
(permalink)
|
kevinc313
CLASSIFIED ULTRA Member
- Total Posts : 5004
- Reward points : 0
- Joined: 2/28/2019
- Status: offline
- Ribbons : 22
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 4:01 PM
(permalink)
KingEngineRevUp
This would explain why the FTW3 hold more voltage than the Strix in TimeSpy GPU Test 2
https://docs.google.com/s...rtpof=true&sd=true
No back to back benches but I felt like the FTW3 had a bit more zip in games at stock settings.
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 4:31 PM
(permalink)
KingEngineRevUp
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his mind again.
Yea I'm resisting posting this thread over there. Lol
post edited by the_Scarlet_one - Thursday, August 26, 2021 5:15 PM
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 4:39 PM
(permalink)
talon951
KingEngineRevUp
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his mind again.
Yea I'm resisting posting this thread over there. Lol
My PSU is a Corsair RM1000i, it lets me read several readings. Power draw form wall and power delivered to devices. Let me know if you think I should do any test on my end, but it sounds like the both of you have confirmed what we have all believed. The FTW3 needs a firmware update to fix its sensors.
post edited by the_Scarlet_one - Thursday, August 26, 2021 5:15 PM
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 4:54 PM
(permalink)
KingEngineRevUp
talon951
KingEngineRevUp
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his mind again.
Yea I'm resisting posting this thread over there. Lol
My PSU is a Corsair RM1000i, it lets me read several readings. Power draw form wall and power delivered to devices. Let me know if you think I should do any test on my end, but it sounds like the both of you have confirmed what we have all believed. The FTW3 needs a firmware update to fix its sensors.
Seems fine like it is. There's a reason why I haven't been concerned about the indicated low power limit.
post edited by the_Scarlet_one - Thursday, August 26, 2021 5:16 PM
|
kevinc313
CLASSIFIED ULTRA Member
- Total Posts : 5004
- Reward points : 0
- Joined: 2/28/2019
- Status: offline
- Ribbons : 22
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 4:56 PM
(permalink)
KingEngineRevUp
talon951
KingEngineRevUp
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his **** again.
Yea I'm resisting posting this thread over there. Lol
My PSU is a Corsair RM1000i, it lets me read several readings. Power draw form wall and power delivered to devices. Let me know if you think I should do any test on my end, but it sounds like the both of you have confirmed what we have all believed. The FTW3 needs a firmware update to fix its sensors.
I wouldn't believe a power meter alone (KaW, UPS based, PSU based, other), but seeing it consistently pull more power than another card when set the same, convinced me.
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 4:58 PM
(permalink)
talon951
KingEngineRevUp
talon951
KingEngineRevUp
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his mind again.
Yea I'm resisting posting this thread over there. Lol
My PSU is a Corsair RM1000i, it lets me read several readings. Power draw form wall and power delivered to devices. Let me know if you think I should do any test on my end, but it sounds like the both of you have confirmed what we have all believed. The FTW3 needs a firmware update to fix its sensors.
Seems fine like it is. There's a reason why I haven't been concerned about the indicated low power limit.
Yeah but it would be nice if the sensors read things correctly. Then we could avoid another Phoenixyz lol. I'm going to submit a ticket with EVGA, if you guys have time, it would be nice if more people did so.
post edited by the_Scarlet_one - Thursday, August 26, 2021 5:16 PM
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 5:06 PM
(permalink)
Guys like Phoenix will just scream EVGA is cheating by just changing the indicated power and ignore any facts/data people present.
|
the_Scarlet_one
formerly Scarlet-tech
- Total Posts : 24080
- Reward points : 0
- Joined: 11/13/2013
- Location: East Coast
- Status: offline
- Ribbons : 79
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 5:17 PM
(permalink)
So in the end, people are getting the exact power they are supposed to, but software is not telling them it is exactly what it should be? Looking at the thread on overclock, the thread starts by mentioning power balancing on Pin 2, was that with the older cards or with newer cards? I am trying to get a handle on all of the different I formation out there.
post edited by the_Scarlet_one - Thursday, August 26, 2021 5:23 PM
|
kevinc313
CLASSIFIED ULTRA Member
- Total Posts : 5004
- Reward points : 0
- Joined: 2/28/2019
- Status: offline
- Ribbons : 22
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 5:19 PM
(permalink)
the_Scarlet_one
So in the end, people are getting the exact power they are supposed to, but software is not telling them it is exactly what it should be?
Seems like a minor issue to have.
They are getting ~25w extra at stock settings.
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 5:20 PM
(permalink)
kevinc313
KingEngineRevUp
talon951
KingEngineRevUp
talon951
Yea beat me to it. This lines up with my findings back a while ago that people on this forum didn't believe.
https://www.overclock.net.../page-12#post-28835729
Phoenixyz gonna lose his **** again.
Yea I'm resisting posting this thread over there. Lol
My PSU is a Corsair RM1000i, it lets me read several readings. Power draw form wall and power delivered to devices. Let me know if you think I should do any test on my end, but it sounds like the both of you have confirmed what we have all believed. The FTW3 needs a firmware update to fix its sensors.
I wouldn't believe a power meter alone (KaW, UPS based, PSU based, other), but seeing it consistently pull more power than another card when set the same, convinced me.
I didn't believe it myself initially. I have several other cards. I checked my 3090 and 1080ti with the same meter which matched very closely. It was only the 2nd 8 pin on my 3080ti that didn't match. And also used my UPS as another verification. All I wanted from that originally was to make sure I wasn't going to blow a fuse while running the Galax XOC (which has borked power readings on the FTW3).
|
neteng101
Superclocked Member
- Total Posts : 153
- Reward points : 0
- Joined: 2/3/2018
- Status: offline
- Ribbons : 1
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 6:14 PM
(permalink)
talon951
I didn't believe it myself initially. I have several other cards. I checked my 3090 and 1080ti with the same meter which matched very closely. It was only the 2nd 8 pin on my 3080ti that didn't match. And also used my UPS as another verification.
All I wanted from that originally was to make sure I wasn't going to blow a fuse while running the Galax XOC (which has borked power readings on the FTW3).
Power meters don't lie - we don't know how much waste/inefficiency there might be though in the power delivery ie. power that turns into heat vs. actually doing something. Why is this phoenix person losing their minds? Software monitoring numbers has always been suspect!
|
homestyle
Superclocked Member
- Total Posts : 228
- Reward points : 0
- Joined: 11/21/2020
- Status: offline
- Ribbons : 0
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 6:25 PM
(permalink)
If true, this is even worse.
|
SoldierRBT
Superclocked Member
- Total Posts : 109
- Reward points : 0
- Joined: 1/6/2011
- Status: offline
- Ribbons : 0
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 6:43 PM
(permalink)
Does anyone know if the 3080 Ti FTW3 Hydro Copper has the same power issues?
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 6:50 PM
(permalink)
neteng101
talon951
I didn't believe it myself initially. I have several other cards. I checked my 3090 and 1080ti with the same meter which matched very closely. It was only the 2nd 8 pin on my 3080ti that didn't match. And also used my UPS as another verification.
All I wanted from that originally was to make sure I wasn't going to blow a fuse while running the Galax XOC (which has borked power readings on the FTW3).
Power meters don't lie - we don't know how much waste/inefficiency there might be though in the power delivery ie. power that turns into heat vs. actually doing something.
Why is this phoenix person losing their minds? Software monitoring numbers has always been suspect!
Well the GPU reads the 8-pin power at the shunts which are near the connector. So I would think the clamp meter would be fairly accurate. And I felt that was proven out to some extent by matching several connectors readings in GPUZ. He was running that Endwalker benchmark and couldn't match his friends Strix with his FTW3. That benchmark is fairly system dependent as it runs high framerates so most likely some other difference between the systems.
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 7:00 PM
(permalink)
neteng101
talon951
I didn't believe it myself initially. I have several other cards. I checked my 3090 and 1080ti with the same meter which matched very closely. It was only the 2nd 8 pin on my 3080ti that didn't match. And also used my UPS as another verification.
All I wanted from that originally was to make sure I wasn't going to blow a fuse while running the Galax XOC (which has borked power readings on the FTW3).
Power meters don't lie - we don't know how much waste/inefficiency there might be though in the power delivery ie. power that turns into heat vs. actually doing something.
Why is this phoenix person losing their minds? Software monitoring numbers has always been suspect!
I'll have to save you some trouble. This user asked me to troubleshoot his card, and I tried helping him for a few hours only to discover his card was fine, IIRC, it was getting 21500 on timespy graphic test. But on overclock.net, this user started a crusade and kept posting several times that the FTW3 is a scam and was telling everyone not to buy it because his FTW3 sucked at OC and couldn't get its rated power draw. I showed him evidence from Talon that it was, just not reporting correctly. I showed him a comparison of a timespy run from a Strix and FTW3 and their voltages. He even challenged me to beat his friends Strix benchmark scores, which I did so by 5% (today I'm 6% over his friend thanks to my WB). Nothing changed this users mind. Not sure what he's been up to since I blocked him. Warning you, it's probably a waste of time reading all this unless you want to laugh at someone's delusion and hysteria. Starts at post #665 https://www.overclock.net/threads/official-nvidia-rtx-3080-ti-owners-club.1791823/page-34
|
neteng101
Superclocked Member
- Total Posts : 153
- Reward points : 0
- Joined: 2/3/2018
- Status: offline
- Ribbons : 1
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 7:24 PM
(permalink)
talon951
He was running that Endwalker benchmark and couldn't match his friends Strix with his FTW3. That benchmark is fairly system dependent as it runs high framerates so most likely some other difference between the systems.
TimeSpy doesn't lie either - the graphics score almost totally hits the GPU, only GPU related bottlenecks can affect the score some (like reduced number of PCIe lanes). Can't tell you how many times I've heard people complaining about poor performance in some game like Cold War with high-end 30-series cards stuck into systems with older gen CPUs, or other situations that aren't the GPUs fault. Guess some people are just dense and will never learn.
|
hallmark888
New Member
- Total Posts : 76
- Reward points : 0
- Joined: 12/16/2010
- Status: offline
- Ribbons : 1
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 7:31 PM
(permalink)
Is the Strix throttling clocks more or less than the FTW3?
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 8:03 PM
(permalink)
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 9:50 PM
(permalink)
Acoustic and my results aren't very comparable. I'm 14C hotter and Acoustic made that run locking a relatively low voltage with a large offset on the core. I ran a lower offset but more voltage (and more power). Acoustic's run did not use the full PL. Which makes that run particularly impressive.
|
KingEngineRevUp
FTW Member
- Total Posts : 1030
- Reward points : 0
- Joined: 3/28/2019
- Status: offline
- Ribbons : 9
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 9:57 PM
(permalink)
talon951
Acoustic and my results aren't very comparable. I'm 14C hotter and Acoustic made that run locking a relatively low voltage with a large offset on the core. I ran a lower offset but more voltage (and more power). Acoustic's run did not use the full PL. Which makes that run particularly impressive.
Ah, I see. Well at least we have his stock run in that excel sheet. The voltage behavior can be compared to one another.
|
talon951
FTW Member
- Total Posts : 1026
- Reward points : 0
- Joined: 10/6/2020
- Status: offline
- Ribbons : 3
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 10:22 PM
(permalink)
KingEngineRevUp
talon951
Acoustic and my results aren't very comparable. I'm 14C hotter and Acoustic made that run locking a relatively low voltage with a large offset on the core. I ran a lower offset but more voltage (and more power). Acoustic's run did not use the full PL. Which makes that run particularly impressive.
Ah, I see. Well at least we have his stock run in that excel sheet. The voltage behavior can be compared to one another.
Yes it looks consistent with what Kevin found in directly comparing the cards.
|
kraade
CLASSIFIED Member
- Total Posts : 2433
- Reward points : 0
- Joined: 10/18/2020
- Location: Georgia
- Status: offline
- Ribbons : 7
Re: Found the missing ~25w of 3080 Ti FTW3 power.
Thursday, August 26, 2021 10:55 PM
(permalink)
More Watts doesn't equal more performance My 3080 on the 450w OC bios ran its fastest at 418w, and my KPE ran its fastest (closed case) at 468w on the LN2 520w bios = https://www.3dmark.com/3dm/64891803I have opened this up with the CL tool and made a at PR pulling 570w and scored a whapping 13988 IF you have the glass doors on and are not using extra cooling, I suggest that thinking your going to run at max power and its going to stay cool is completely unrealistic. Just because you push the sliders doesn't mean it going to pull your expectations, and what does that mean if anything/ It certainly doesn't mean you got a dud of a card or it has power issues. my2c...
|