Discussion Can graphics cards with external PCIE power connectors run off of just the power from the PCIE slot?
- Thread starter Order 66
- Start date
- Status
jordanbuilds1
Prominent
- Jan 16, 2024
- 275
- 51
- 760
you will get output, but the preformance will prob be on par with a lp version of a card.
- Jun 19, 2011
- 33,688
- 1,700
- 159,590
Nope, likely to either just not post or give a warning screen that says to plug in the pci-e power connectors.
jordanbuilds1
Prominent
- Jan 16, 2024
- 275
- 51
- 760
- May 27, 2010
- 28,394
- 5,771
- 144,340
Cards check to see if the PCIe grounds are hooked up. If they aren't they won't POST or they will prompt the BIOS to ask you to plug in the cables.
I forgot to plug my GTX980 back in once after some routine maintenance. It told me to plug it in. Might have been m GTX1080, or both.
I forgot to plug my GTX980 back in once after some routine maintenance. It told me to plug it in. Might have been m GTX1080, or both.
- Apr 13, 2023
- 2,170
- 919
- 3,570
But why not make the cards power limited to 75W without being plugged in, rather than not posting?
hotaru.hino
Glorious
- Sep 1, 2020
- 9,822
- 3,567
- 46,090
If you want people to make sure they use the product as intended, then prevent them from using the product not as intended.
Imagine a case where someone sort of half plugs in the PEG cables. Or at least enough where it looks like a good connection is made. They boot the computer up, but get crippled performance. Then they go through the rigamaroll of software fixes, maybe report the card doesn't seem to use more than 75W, and people could conclude the person has a bad card.
When in reality, they just needed to push the cables in a bit further.
- Apr 13, 2023
- 2,170
- 919
- 3,570
that’s a good point. I still feel like it would be nice for the (admittedly niche) use case of putting a relatively powerful GPU in an office pc with a proprietary power supply, and just using the slots power for the GPU. Pointless, sure, but I would have a use for it.
hotaru.hino
Glorious
- Sep 1, 2020
- 9,822
- 3,567
- 46,090
Proprietary PSUs tend to be proprietary only on the motherboard side, because the system builder designed/made the board so they control it. They can't do proprietary plugs for other things like video cards or storage drives, because they don't make them (video cards from Dell or whatnot are rebadged). So I don't really see a problem there.
We could have the standards make a slot-based power thing like what Apple did and what MSI and ASUS are trying to do, but I don't think everyone will be on board with that. If it becomes standard, it's most likely going to optional, and that's just going to create a headache for people using cheaper boards that likely won't include it to save cost.
- May 27, 2010
- 28,394
- 5,771
- 144,340
Yeah, the current implementation from ASUS makes the expansion slot longer which certainly won't work for all form factors.
Has its own problems anyway. Not going to see 675W through the motherboard without some drastic changes. So it would have to become some in-between standard. Not to mention the need for people to then plug in the PCIe or additional EPS cable to the motherboard.
I think there is too great an expense in board layers on the cheaper side of things for it to get adopted. Why make all boards with a huge power plane if the scenario of an APU is most common for off the shelf desktops?
Dell video cards are sort of re-badged. Usually made by Foxconn or PNY, but they are Dell approved designs, usually with their own 'custom' PCB/Cooler, likely from Cooler Master anyway.
It would have been confusing as well, but they should have just switched to dual EPS instead of triple PCIe 8-pin and called it a day. I like the 12VHPWR connector in theory, just the implementation in practice is bad, and still confusing since the same connector has multiple power delivery options.
Not sure what will happen if OEMs adopt it, 400W PSU with a 12VHPWR cable that only does 75W or something? Going to be people buying big cards that fit in the chassis, and wondering why they don't perform as well. Assuming they won't get mad at being under powered. I know some of the cards will already work with 3/4 8-pin plugged into the 12VHPWR adapter, they just disable the ability to adjust the card power.
Or worse, they will create a whole new class of computer that comes with a 12VHPWR connector which means a huge power supply, costing hundreds of extra dollars and then ship it with an RTX4060.
Has its own problems anyway. Not going to see 675W through the motherboard without some drastic changes. So it would have to become some in-between standard. Not to mention the need for people to then plug in the PCIe or additional EPS cable to the motherboard.
I think there is too great an expense in board layers on the cheaper side of things for it to get adopted. Why make all boards with a huge power plane if the scenario of an APU is most common for off the shelf desktops?
Dell video cards are sort of re-badged. Usually made by Foxconn or PNY, but they are Dell approved designs, usually with their own 'custom' PCB/Cooler, likely from Cooler Master anyway.
It would have been confusing as well, but they should have just switched to dual EPS instead of triple PCIe 8-pin and called it a day. I like the 12VHPWR connector in theory, just the implementation in practice is bad, and still confusing since the same connector has multiple power delivery options.
Not sure what will happen if OEMs adopt it, 400W PSU with a 12VHPWR cable that only does 75W or something? Going to be people buying big cards that fit in the chassis, and wondering why they don't perform as well. Assuming they won't get mad at being under powered. I know some of the cards will already work with 3/4 8-pin plugged into the 12VHPWR adapter, they just disable the ability to adjust the card power.
Or worse, they will create a whole new class of computer that comes with a 12VHPWR connector which means a huge power supply, costing hundreds of extra dollars and then ship it with an RTX4060.
- May 27, 2010
- 28,394
- 5,771
- 144,340
RTX 3050 6GB is your best bet off the shelf for now.
https://pcpartpicker.com/product/TG...-6gb-6-gb-video-card-rtx-3050-ventus-2x-6g-oc
There are a few Quadro, but probably not what you want to pay for.
https://pcpartpicker.com/product/TG...-6gb-6-gb-video-card-rtx-3050-ventus-2x-6g-oc
There are a few Quadro, but probably not what you want to pay for.
- Apr 13, 2023
- 2,170
- 919
- 3,570
- May 27, 2010
- 28,394
- 5,771
- 144,340
That would be the Quadro. They don't call them that anymore, but it is what they essentially are.
Just a different market segment. The workstation drivers enable features that are denied to the geforce versions. These are often installed in desktops for work purposes such as CAD/Simulation etc where the drivers allow for far higher performance even if the silicon is the same. Usually very specific tasks though, general CUDA core related functions don't change much.
Also the case of taking the best GPUs that can be very efficient in the case for the Ada 4000 SFF card. 70W and that powerful is amazing.
The main thing is you can call up Nvidia when there is a problem, and they will fix it. Geforce, they'll get around to it eventually, but Quadro, it can be as little as a few days.
Once with Autocad 2010 I believe, we had some shiny new Lenovo workstations with Quadro 4000 (Fermi). With dual monitors, constantly having the cursor disappear. We talked with Nvidia and had a new Beta driver the next day. About one week later there was a new Quadro driver released.
Just a different market segment. The workstation drivers enable features that are denied to the geforce versions. These are often installed in desktops for work purposes such as CAD/Simulation etc where the drivers allow for far higher performance even if the silicon is the same. Usually very specific tasks though, general CUDA core related functions don't change much.
Also the case of taking the best GPUs that can be very efficient in the case for the Ada 4000 SFF card. 70W and that powerful is amazing.
The main thing is you can call up Nvidia when there is a problem, and they will fix it. Geforce, they'll get around to it eventually, but Quadro, it can be as little as a few days.
Once with Autocad 2010 I believe, we had some shiny new Lenovo workstations with Quadro 4000 (Fermi). With dual monitors, constantly having the cursor disappear. We talked with Nvidia and had a new Beta driver the next day. About one week later there was a new Quadro driver released.
Angel Fernandez
Distinguished
- Jul 25, 2015
- 46
- 10
- 18,535
Joseph_138
Distinguished
- Nov 22, 2016
- 747
- 184
- 19,190
Agree with this. This is the only use case where I could recommend using a 6gb 3050. It's the least expensive, and most up to date card that you'll find that doesn't require external power.
- Status
TRENDING THREADS
-
Question upgraded to 32gb ddr4 and now it wont boot above 2133mhz - help appreciated- Started by SophieTheMeh
- Replies: 4
-
-
-
-
Question Difficulty over Nord (and other) VPN configuration
- Started by kurtepearl22
- Replies: 0
Facebook
Twitter
Instagram