-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
NVIDIA RTX 50 SERIES - Technical/General Discussion
- Thread starter scubes
- Start date
Soldato
- Joined
- 21 Jul 2005
- Posts
- 20,800
- Location
- Officially least sunny location -Ronskistats
People still bought them and many paid over the msrp just like now. Just highlighting.
Much easier to make up gaps in the data you have then to predict the future
Jensen wants exactly that. He wants to be like Apple and exclusively manufacture all their own hardware.
Not at all my friend. Not sure if you noticed a page or so back, but things went off on a tangent which had absolutely nothing to do with GPUs or PC hardware really.
Any discussion related to 5000 series is fine in my book, especially if it involves the experience of owning/buying one.
Though with how things appear, very few have the experience of buying a 5000 series GPU
Anyhow, we have folks hating on framegen, my point was that framegen was the next evolutionary step after DLSS and upscaling. Fake frames, fake resolution, is anything real at this point?
And these 12VHPWR cables... would we reckon the fire/melting risk is less on lower power cards like 5070ti and 5080? Is there any implementation of these where they can be safely used? I ignored the 12VHPWR cable that came with my PSU, but if a future GPU upgrade forces me to use it... I should probably take a closer look.
I have a ‘really into the weeds’ technical question regarding the PCIE 5.1 spec - hoping that someone could clarify this for me.
There’s quite a lot of info available regarding the increased options available for Sense 0 and Sense 1 pins in the new 12V-6x2 spec:
hwbusters.com
I’m struggling to understand the significance of the new options being available for 0w and 150w power states.
From the second link above:
^^^ the formatting there could be a LOT better but it looks like the first table is for the new standard, whereas the second table is for the old standard.
From the first intel link:
I’m currently sat here thinking… “but what does it all mean, Basil?!”
I understand that the new pin lengths help ensure a more robust connection (hence H++ being better) but I’m ultimately trying to understand if the info I’ve set out in this post (re: SENSE0 and SENSE1) is, or isn’t, an additional reason to be weary of using a H+ socket with a 50 series card. It might be a total nothing-burger, but I’m not sure.
Any thoughts, oh wise ones?
There’s quite a lot of info available regarding the increased options available for Sense 0 and Sense 1 pins in the new 12V-6x2 spec:
ATX v3.1 & PCIe CEM 5.1 are official! - Hardware Busters
Hardware Busters - ATX v3.1 & PCIe CEM 5.1 are official! - PSUs
hwbusters.com
I’m struggling to understand the significance of the new options being available for 0w and 150w power states.
From the second link above:
^^^ the formatting there could be a LOT better but it looks like the first table is for the new standard, whereas the second table is for the old standard.
From the first intel link:
I’m currently sat here thinking… “but what does it all mean, Basil?!”
I understand that the new pin lengths help ensure a more robust connection (hence H++ being better) but I’m ultimately trying to understand if the info I’ve set out in this post (re: SENSE0 and SENSE1) is, or isn’t, an additional reason to be weary of using a H+ socket with a 50 series card. It might be a total nothing-burger, but I’m not sure.
Any thoughts, oh wise ones?
Last edited:
Associate
- Joined
- 17 Jan 2025
- Posts
- 70
- Location
- Somerset
From what I can gather from reading up is anything over 360w is a *danger* zone and anything below should be fine.
600w should be fine through the cable, but only aslong as the amps stay even, its only when more current is drawn through 1 or 2 cables it becomes too much and over the margin.
I'm no specialist so still do your own research, I have to use a 12v-2x6 with my 5090 when it arrives as that's the only option I have, otherwise I need to buy a new psu again.
A 5060 maybe.
The 5070 is rumoured to be a 250W card so maybe the 5060 will shave 25W off that.
Soldato
- Joined
- 21 Jul 2005
- Posts
- 20,800
- Location
- Officially least sunny location -Ronskistats
MicroCenter confirms GeForce RTX 5070 Ti prices: one MSRP model and $930 average price.
Deserves that top gun song!
Deserves that top gun song!
Last edited:
Associate
- Joined
- 11 Sep 2020
- Posts
- 918
- Location
- London
I power limit my 4090 to 70% (315w) and have done for some time (was 80-90% PL before that). Having said that, I do not want to pull the power connector out to check the end
Last edited:
Associate
- Joined
- 11 Sep 2020
- Posts
- 918
- Location
- London
Caseking prices are absolutely shocking, so expect the same from OCUK....
Last edited:
It may help when you know the difference between a short circuit and open circuit ; a short circuit essentially means two things are connected whereas an open circuit is the opposite.
So in the bottom chart you could have no sense pins connected (current can not flow on any sense pins) but that would still tell the device that it could draw 100/150W ; In the top chart that changed to not drawing power if no sense pins are connected (if we can't send a current down the sense pins) and only drawing 100/150W if both sense pins are connected (if we can send a current down the sense pins and that current is returned to us on another sense pin).
Short circuit = current flows from one point to another
Open circuit = current can not flow.
Ground = current is flowing but it's not reaching the other point.
Last edited:
Thanks @Murphy - all noted. The bit I’m struggling with is that I presume what determines the power being ‘drawn’ is… surely, the GPU. That’s the part which needs more or less power, depending on what your PC is doing at a given time.
In which case, are these pins movable or fixed in the PSU? If movable, how does the PSU know when to change them. And what scenario are these changes in specs looking to ‘solve’? Asking with full humility and acknowledgement of my current ignorance.
Sorry for the ‘additional questions’ - I didn’t want to massively overload my initial post.
In which case, are these pins movable or fixed in the PSU? If movable, how does the PSU know when to change them. And what scenario are these changes in specs looking to ‘solve’? Asking with full humility and acknowledgement of my current ignorance.
Sorry for the ‘additional questions’ - I didn’t want to massively overload my initial post.
Last edited:
Real flames at least?
Still sat here waiting for any kind of update from OCUK on my order, is my order real? lol
What you're struggling with is why IMO calling them sense pins is a bit misleading as they're a way to tell the GPU if a cable/PSU is connected and if so what's the maximum wattage that can be drawn from the PSU or though the cable.
The sense pins, what is or isn't connected to what, would be set when the PSU or cabled is manufactured. It's kind of like how the old 6 & 8 pin PCI-E power connectors had the same number of 12v sources (3) but the 8 pin added a couple of extra pins to indicate to devices that 'hey I'm a better PSU/cable so you can draw more power from/through me'
Ahhhh - I see. Thank you for helping me with this.
That explains what the new ‘0w status’ is achieving on the PSU side then with a H++ then.
Not only is it the case that the various pins have revised (more appropriate) lengths, but the ability for the GPU to draw power has significantly changed: the GPU is effectively now being told “you can currently only pull a limited amount, which because of the new spec may now may be nooooothing at alll [/flanders wiggle] if there is an improper connection.”
By contrast, if there is an improper connection with an old H+ socket the GPU will always think it can draw power if there is a connection - because there was never a 0w state.
So this is an additional layer of redundancy ON TOP of the various pins having revised lengths in the H++ socket.
This feels really significant to me as it’s only really the ‘pin length changes’ that are typically discussed on this topic.
Again, thanks for helping - and do please nudge me in the right direction if I’m still misreading the situation.
Edit: … and if all of that is right, it seems ridiculous that there was never a 0w state in the initial spec?!
Last edited:
It does a bit and it's only something that occurred to me after you brought this up so thanks....
I'm beating Jufes with my current work in progress undervolt/ overclock on my 5080. Got a graphics score of 35,140 in TIMESPY with a core voltage of only 0.97. Max power draw about 350w....
I scored 29 339 in Time Spy
AMD Ryzen 7 9700X, NVIDIA GeForce RTX 5080 x 1, 32768 MB, 64-bit Windows 11}
www.3dmark.com
With my less aggressive core volts of only 0.94 and max power draw of about 315w I'm pretty much matching....
I scored 28 945 in Time Spy
AMD Ryzen 7 9700X, NVIDIA GeForce RTX 5080 x 1, 32768 MB, 64-bit Windows 11}
www.3dmark.com
And, with only 0.91v(!) and max power draw of ~295w, getting 34,109....
I scored 28 576 in Time Spy
AMD Ryzen 7 9700X, NVIDIA GeForce RTX 5080 x 1, 32768 MB, 64-bit Windows 11}
www.3dmark.com
For ref in his vid he was doing a quick and dirty sliders only overclock in Afterburner but he was maxing out power to 375w and presumably volts were at 1.02 at least. Really quite impressed with these 5080s you know, especially the undervolting.
Last edited: