Discussion To hell with nVidia's definition of SFF ready. What should be the actual SFF standard?

[GuilleAcoustic]

Similar to R9 Nano and older ITX cards!

It's actually 4mm shorter than the R9 Nano ... not that it matters a lot as it is still shorter than an ITX mobo, but it proves that it's still possible and that almost 300mm triple fans 115W GPU like the Gigabyte AERO 4060 are just lazy non-sense.

 

[SFFMunkee]

It's actually 4mm shorter than the R9 Nano ... not that it matters a lot as it is still shorter than an ITX mobo, but it proves that it's still possible and that almost 300mm triple fans 115W GPU like the Gigabyte AERO 4060 are just lazy non-sense.

Except they can markup big chonky cards and sell more. So $$ wins again even if it’s wasteful from every other metric.

Capitalism: Line must go up.

 

[robbee]

Lenovo have a nice ITX sized RTX 4060, never seen it before. It has some interesting dimensions (150mm length).

View attachment 3707

Would be a great fit with a Mini STX motherboard if the power connector wasn't aimed to the rear

 

[SaperPL]

Would be a great fit with a Mini STX motherboard if the power connector wasn't aimed to the rear

It's shorter than ITX board, the power cable can be bent right around the board's end so it's not a problem. In most cases you will have to have additional space in front of the board to even install the case, so you'll have that room for it. I'd prefer if card would go this way instead of having connectors on the side.

 

[robbee]

It's shorter than ITX board, the power cable can be bent right around the board's end so it's not a problem. In most cases you will have to have additional space in front of the board to even install the case, so you'll have that room for it. I'd prefer if card would go this way instead of having connectors on the side.

For ITX it’s indeed perfect, but I was referring to STX, which has about the same dimension length-wise

 

[SaperPL]

For ITX it’s indeed perfect, but I was referring to STX, which has about the same dimension length-wise

My bad, seems like temporary blindness. But are there mini STX boards not requiring shenanigans with converting M.2 to pci-e to use graphic cards?

 

[yeshyyyk]

And just a PITA when fiddling around behind a PC case IMO

In my personal view, if you want to call something SFF-Ready AT ALL it should be based on the original PCI/PCIe spec

View attachment 3706

So I would suggest absolutely max of 111mm tall, 40mm thick (i.e. DUAL SLOT not 2.1, 2.2, etc) with considerations for power connectors and cable routing etc. for full sized cards.

Badge / Category	MAX Length (mm)	MAX Height (mm)	MAX Thickness (mm)	Example Card
SFF Long	312	111	40	IDK if there are any true dual-slot cards that could fit here???
SFF Standard / Reference	267	111	40	AMD Reference RX 6700 XT AMD Radeon RX 6700 XT Specs AMD Navi 22, 2581 MHz, 2560 Cores, 160 TMUs, 64 ROPs, 12288 MB GDDR6, 2000 MHz, 192 bit www.techpowerup.com nVidia RTX 3070 FE NVIDIA GeForce RTX 3070 Specs NVIDIA GA104, 1725 MHz, 5888 Cores, 184 TMUs, 96 ROPs, 8192 MB GDDR6, 1750 MHz, 256 bit www.techpowerup.com
SFF Short	225	111	40	PowerColor Fighter RX 6600XT PowerColor Fighter RX 6600 XT Specs AMD Navi 23, 2589 MHz, 2048 Cores, 128 TMUs, 64 ROPs, 8192 MB GDDR6, 2000 MHz, 128 bit www.techpowerup.com
SFF ITX	150-180	111	40	HP RTX 3060 Single Fan HP RTX 3060 OEM Specs NVIDIA GA106, 1777 MHz, 3584 Cores, 112 TMUs, 48 ROPs, 12288 MB GDDR6, 1875 MHz, 192 bit www.techpowerup.com PowerColor Radeon RX 5600 XT ITX PowerColor RX 5600 XT ITX Specs AMD Navi 10, 1560 MHz, 2304 Cores, 144 TMUs, 64 ROPs, 6144 MB GDDR6, 1750 MHz, 192 bit www.techpowerup.com
SFF LP	190	69	40	Gigabyte RTX 4060 LP GIGABYTE RTX 4060 Low Profile OC Specs NVIDIA AD107, 2475 MHz, 3072 Cores, 96 TMUs, 48 ROPs, 8192 MB GDDR6, 2125 MHz, 128 bit www.techpowerup.com
SFF LP ITX	170	69	40	nVidia RTX 4000 SFF Ada NVIDIA RTX 4000 SFF Ada Generation Specs NVIDIA AD104, 1560 MHz, 6144 Cores, 192 TMUs, 80 ROPs, 20480 MB GDDR6, 1750 MHz, 160 bit www.techpowerup.com

Anyways, just my 2c

Most recently, I think Asrock 7900XT/X Creator works for SFF Long, Nvidia has nothing / closest is Gigabyte 4070Ti Super AI TOP

Powercolor Reaper 9070/XT is sooo close, it's the only dual slot option for those cards

(but on the non XT you lose 30% TDP / 80W for no space savings)

Similar to R9 Nano and older ITX cards!

But if it takes 60W (or even 15W less for 4060Ti models) less after ~9years, is it really similar? That's what I was trying to get at in my post anyways

 

[SaperPL]

Powercolor Reaper 9070/XT is sooo close, it's the only dual slot option for those cards

What is wrong with it? I think it is actually the only reference sized card released this year? What do you mean by "sooo close"?

Do you mean that it's 41mm and not 40?

 

[yeshyyyk]

What is wrong with it? I think it is actually the only reference sized card released this year? What do you mean by "sooo close"?

Do you mean that it's 41mm and not 40?

Yeah, just per his opinion / PCI/e spec

It's a fine card size-wise, I do wish 9070 non XT was smaller since it's 30% less TDP so 30% less size

And then to go further, Gigabyte 4070Ti Super AI TOP and Inno3D X2 4070Ti Super do better, so Reaper could be dual fan / non XT could maybe be 200mm length? (15-20W difference between 4070TiS and 9070XT, surely it's not much /it can be forgone)

Honestly I don't know how crazy AMD / Nvidia are on adhering to their TDP spec, but I wouldn't mind losing some TDP, anyways you can reclaim performance with undervolt (perhaps Intel is the biggest offender, but they certainly don't care about TDP spec on laptops)

And then maybe I am rambling on, but with copper: maybe it is inaccurate to compare CPU and GPU coolers, but maybe something like a TR AXP90-47 full copper, but x2 on a GPU, with shorter copper block (so it's < 40mm tall / 2 slot).

And then add 3rd one if needed lol, or instead of 90mm wide copper block you can go to 110 (as per spec) to compensate for losing the 7mm on height

I think that should be able to easily cool 200W in the first case and 300W in the 2nd? If you want to get a little more optimal you could maybe "trade" a few mm off the fan (15mm to 10-12) for more copper

(would be cool to see the trade-off for fan size vs copper/aluminum, but I'm pretty sure it's in favor of copper lol)

I don't know how everything works, but I strongly feel if Galax was able to do >300 "W/L" all the way back in 2017, then we should be able to do the same now (and not just on 5090FE / 575W card)

Galax GTX 1070 Katana: Single-slot Graphics

Galax's GTX 1070 Katana caught our attention for counting itself among the few single-slot gaming graphics cards available today. In fact, to the best of our knowledge,...

www.techspot.com

Besides thinking that it would run warmer and louder than most GTX 1070s, I wasn't sure what to expect from the Katana's performance in this category so I was pleasantly surprised by the card's 82 degree result, which was a degree under the Founders Edition card despite providing slightly better performance. Of note, the card wasn't too noisy out of the box but maintaining that 82 degrees after our overclock required us to increase the fan speed to a pretty loud setting.

It doesn't mean anything to me to put 100W card in long / thick / wide card, even if it meets spec. But let's say a higher length low profile, or a thicker shorter card would be interesting

 

[SaperPL]

Discussion about specs aside, do we have any way of making things happen?

I can only envision this working in a way that we come with some document that describes required specs to get some kind of badge or marking of "SFF.NET Ready" and we get it to be "signed" by as much people in the community as possible, so it works as an open letter to vendors explaining that we want this and explanation of what requirements are there to freely put the badge/marking on your product (open self certification in contrast to how SFF-Ready works with nvidia). And when this happens we need to reach out to media with this somehow and ask mods here to put it on the front page somehow.

The question is, what is the process of getting there to have that document made.

 

[SFFMunkee]

Discussion about specs aside, do we have any way of making things happen?

I can only envision this working in a way that we come with some document that describes required specs to get some kind of badge or markin eg of "SFF.NET Ready" and we get it to be "signed" by as much people in the community as possible, so it works as an open letter to vendors explaining that we want this and explanation of what requirements are there to freely put the badge/marking on your product (open self certification in contrast to how SFF-Ready works with nvidia). And when this happens we need to reach out to media with this somehow and ask mods here to put it on the front page somehow.

The question is, what is the process of getting there to have that document made.

We need to get community polls, then use that to get (commercial) case manufacturers onboard, then use THAT to get the card manufacturers on board
Otherwise this is all just tooting our own horns

What are some examples of similar action that we've seen in the PC enthusiast community we can use as a template of sorts?

 

[SaperPL]

then use that to get (commercial) case manufacturers onboard

I think going with that to graphics card vendors first makes more sense. Case vendors won't get onboard first if we don't convince card vendors because what does "convincing" actually mean here - it means asking to manufacture a product within spec. Risking manufacturing cases in the spec before the cards are there is a no-go.

For graphic card vendors it's a different story - there will be tons of cases supporting smaller cards and convincing them to go for this kind of badge/rating/award for few models they would have makes more sense and this is what we should be actually aiming for.

There won't be new cases restricted to such spec if there are no card vendors onboard. Unless your point is to get in touch with case makers to sign and back up the initiative without manufacturing cases to such restrictions, but then it can backfire if we end up with just few of them backing us while most of the case vendors will ignore us and this will actually show the card vendors that big case makers don't care.



I would use community polls to decide what should be in the spec and what not, but once that set and we decide and make a document with specification and explanation I would rather go for number of signatures than a poll. With poll you will get a lot of people that are not invested in the topic enough to understand that making such standard doesn't mean we're going to take away their big and beefy cards altogether and they will just vote against it because of that.

Also a poll with just numbers doesn't prove the actual numbers of people taking part in it. And requiring email verification isn't something that can stop you from generating a temporary email accounts just to vote.

What I think could make a lot of sense would be to make a signature page collecting nicknames + country + list of significant SFF cases those signers owned. We could have a threads where people on the forums/subreddit would write a comment and we would combine that into a page/section of signatures in the document, or we could just link to those threads and threads could have limit of one comment per user and there could be a requirement of the account being at least a year old with at least amount of posts, and the document would just show the numbers at the time of revising/updating it.

There are people that tend to get almost every new good and/or interesting case and they end up having multiple of them as a hobby and there are people buying and selling those cases once a new one comes - those are the people with actual experience with multiple SFF cases and trying to match the hardware to them should be the people that matter in signing such spec, not the anonymous random naysayers that will flood the polls if we get some youtubers on board to spread the word.

 

[SFFMunkee]

I think going with that to graphics card vendors first makes more sense. Case vendors won't get onboard first if we don't convince card vendors because what does "convincing" actually mean here - it means asking to manufacture a product within spec. Risking manufacturing cases in the spec before the cards are there is a no-go.

For graphic card vendors it's a different story - there will be tons of cases supporting smaller cards and convincing them to go for this kind of badge/rating/award for few models they would have makes more sense and this is what we should be actually aiming for.

There won't be new cases restricted to such spec if there are no card vendors onboard. Unless your point is to get in touch with case makers to sign and back up the initiative without manufacturing cases to such restrictions, but then it can backfire if we end up with just few of them backing us while most of the case vendors will ignore us and this will actually show the card vendors that big case makers don't care.



I would use community polls to decide what should be in the spec and what not, but once that set and we decide and make a document with specification and explanation I would rather go for number of signatures than a poll. With poll you will get a lot of people that are not invested in the topic enough to understand that making such standard doesn't mean we're going to take away their big and beefy cards altogether and they will just vote against it because of that.

Also a poll with just numbers doesn't prove the actual numbers of people taking part in it. And requiring email verification isn't something that can stop you from generating a temporary email accounts just to vote.

What I think could make a lot of sense would be to make a signature page collecting nicknames + country + list of significant SFF cases those signers owned. We could have a threads where people on the forums/subreddit would write a comment and we would combine that into a page/section of signatures in the document, or we could just link to those threads and threads could have limit of one comment per user and there could be a requirement of the account being at least a year old with at least amount of posts, and the document would just show the numbers at the time of revising/updating it.

There are people that tend to get almost every new good and/or interesting case and they end up having multiple of them as a hobby and there are people buying and selling those cases once a new one comes - those are the people with actual experience with multiple SFF cases and trying to match the hardware to them should be the people that matter in signing such spec, not the anonymous random naysayers that will flood the polls if we get some youtubers on board to spread the word.

Great response

 

[SaperPL]

I drafted out a doc showing how I envision such document with a standard, although note this is a rough draft without actual drawings and images yet, but with list of things that should end up in there.

Noteworthy is that I think we could get more people and case makers on board if we also tackle problems of other components at the same - there's a higher chance that someone that is still running a system from few years back has encountered a problem with choosing the right cpu cooler, optimising PSU cables etc, rather than just the GPU fitting, and despite the GPU fitting problem being the biggest one, I feel like broadening the subject may convince more people it's worth reading if they start from what problems they have experienced and go over the rest from that.

Not all components need a complex requirements though, but it can be as simple as expecting modular cables from TFX form factor to be consider SFF-Ready because for example Silverstone's 700W TFX has so many cables that you can't really handle in SFF case. Also enhance 7660b has shown that FlexATX can be worth using in SFF builds. Finally low profile CPU coolers should have specs showing how much of space they are expecting to have from the perforation before they start whining or causing turbulence as well as have testing done in actual SFF conditions and not in a full tower chassis with ton of fans in ambient temperature controlled chamber...

 

[SFFMunkee]

I drafted out a doc showing how I envision such document with a standard, although note this is a rough draft without actual drawings and images yet, but with list of things that should end up in there.

Noteworthy is that I think we could get more people and case makers on board if we also tackle problems of other components at the same - there's a higher chance that someone that is still running a system from few years back has encountered a problem with choosing the right cpu cooler, optimising PSU cables etc, rather than just the GPU fitting, and despite the GPU fitting problem being the biggest one, I feel like broadening the subject may convince more people it's worth reading if they start from what problems they have experienced and go over the rest from that.

Not all components need a complex requirements though, but it can be as simple as expecting modular cables from TFX form factor to be consider SFF-Ready because for example Silverstone's 700W TFX has so many cables that you can't really handle in SFF case. Also enhance 7660b has shown that FlexATX can be worth using in SFF builds. Finally low profile CPU coolers should have specs showing how much of space they are expecting to have from the perforation before they start whining or causing turbulence as well as have testing done in actual SFF conditions and not in a full tower chassis with ton of fans in ambient temperature controlled chamber...

Yesssssss to all of this

 

[SaperPL]

Just so we are on the same page - am I thinking about the same thing as rest of you guys? Keeping the possible fitting from pci reference design like this?

This is corresponding with how we assumed 112 mm reference size + 18 mm of room required for PEG cable bend in Sentry.

Note how stupid the recess for 12VHPWR is because of that requirement of 35 mm to the bend - it might make more sense to make it parallel on the top of the card instead of cutting into the board... Maybe the standard should show how stupid this is and require the 12VHWPR to be parallel like in the bottom drawing?

I also thought about potentially drawing the PEG 8-pin as flushed in within the 111,2 mm line, but it could be too greedy to ask for more than what we have had with reference cards. Still the question stands if perhaps when drawing up the standard, maybe we should try to make a step forward?

 

[SFFMunkee]

Just so we are on the same page - am I thinking about the same thing as rest of you guys? Keeping the possible fitting from pci reference design like this?

This is corresponding with how we assumed 112 mm reference size + 18 mm of room required for PEG cable bend in Sentry.

Note how stupid the recess for 12VHPWR is because of that requirement of 35 mm to the bend - it might make more sense to make it parallel on the top of the card instead of cutting into the board... Maybe the standard should show how stupid this is and require the 12VHWPR to be parallel like in the bottom drawing?

I also thought about potentially drawing the PEG 8-pin as flushed in within the 111,2 mm line, but it could be too greedy to ask for more than what we have had with reference cards. Still the question stands if perhaps when drawing up the standard, maybe we should try to make a step forward?

Yes! This is exactly what I’d hope for, and something that’s actually helpful for designing and building small cases / custom builds!

The more I think about it, the more I wonder if the excessively large coolers and low VRAM are in a large part (pun intended) down to market segmentation. That is, they don’t want people buying the (relatively) cheaper gaming cards for AI / professional uses, so make them impossible to fit in multi-GPU setups and limit the VRAM so they’re useless for AI (even if it hurts gaming performance too)

 

[SaperPL]

I put in some hours into outlining the height requirement in my doc, will have to try and make some proper drawings for thickness (2-slot requirements / explanation) and length, and then showcase some case topology variants which will probably burn most of the time here.

I'm not sure how should we treat a situation where we have 2-slot card that has 8-pin cable on top and is 305mm long - should this be generally acceptable that the cable will go in front of the fans or over the backplate because the case is exactly 305mm long? It feels like we should have cards that are 295mm long instead to be able to route the cable around, but then this cable would go exactly in front of the exhaust from radiator.

Also the oversize only up front of the card may seem weird. I'm considering if sticking to purely reference size OR stealth design makes sense over allowing room for shenanigans with oversize when the power connector is on top of the card.

Finally - any ideas on how to split up the work with this? Anyone interested and willing to do some graphics for markings/badges that would look like a kind of award, or help with the text and formatting of the doc? Few people who commented here have quite a lot of posts on the forum, so I'd assume they are experienced...

 

[SFFMunkee]

I put in some hours into outlining the height requirement in my doc, will have to try and make some proper drawings for thickness (2-slot requirements / explanation) and length, and then showcase some case topology variants which will probably burn most of the time here.

I'm not sure how should we treat a situation where we have 2-slot card that has 8-pin cable on top and is 305mm long - should this be generally acceptable that the cable will go in front of the fans or over the backplate because the case is exactly 305mm long? It feels like we should have cards that are 295mm long instead to be able to route the cable around, but then this cable would go exactly in front of the exhaust from radiator.

Also the oversize only up front of the card may seem weird. I'm considering if sticking to purely reference size OR stealth design makes sense over allowing room for shenanigans with oversize when the power connector is on top of the card.

Finally - any ideas on how to split up the work with this? Anyone interested and willing to do some graphics for markings/badges that would look like a kind of award, or help with the text and formatting of the doc? Few people who commented here have quite a lot of posts on the forum, so I'd assume they are experienced...

I'm happy to have a crack at some documentation and planning for it, but I don't have contacts in the industry, that'll be our biggest hurdle

 

[SaperPL]

I'm happy to have a crack at some documentation and planning for it, but I don't have contacts in the industry, that'll be our biggest hurdle

That's not necessarily the point that everyone involved should have contacts. Also contacts in the industry don't necessarily make things happen on their own.

There are two points for not doing such things alone - first one is that one person will always miss something that can be later an issue and ammunition for those who oppose the idea. Even if explanation of why something isn't considered is not going to end up in a doc, it's good to be ready to respond to something when it comes up.

Second point is that the perception of one person doing something vs a group of people doing something is different. One person is a crazy guy thinking he knows better than industry professionals, doing this as a group in a way that they are willing to spend their time on an idea means that one crazy guy has managed to convince those people already that he has something going right.

If I were to compile whole thing on my own and show it off, and even if it gets pinned like this thread - it's still that one crazy guy trying to push through and most of the people will just ignore it and don't even read the thing. Put 10 experienced people on this and write this on top of the page that this was compiled, iterated and reviewed by them and they come up with it to this state over some time, and there's completely different perception of this.

If there's no group/community effort to push it, it'll die quickly the same as the threads under editorials that I wrote on the topic.


About contacting the industry - once we have the doc/the spec for SFF standard that we want to have, and we're agreed as a group that it makes sense, we can talk to staff here to somehow organise the page in a way that when you enter the website, the standard is in-your-face visible on top of the page, it can have slightly bigger area than pinned now just for some infographic banner to be visible, or perhaps if before the main page was the news site, maybe the main page could be the standard itself if we have enough people onboard.

Once we have a place for the standard on the web, we start both getting signatures under the spec and research gathering the list of GPUs (and other components if they are part of the standard) that are within spec. Once we have reasonable amount of signatures, we go to GPU vendors and tell them "look, you've made this amount cards that are matching the standard so far, you can use the our award/standard badge on those products. We will promote those cards as part of our standard on our website as go-to cards for SFF builds"

Once the place for the standard has a list of cards that are an easy pick for SFF builds that is being kept up to date with each release and people start using it to look for their purchase, this is the moment when this can start rolling and making sense to convince the vendors to stop ignoring it.

I do have contacts in nvidia/amd/intel's GPU software divisions through colleagues that I worked with, and we as DR ZĄBER do have contacts with youtube PC hardware media through the Sentry project, but nobody will bat an eye on this if we don't have something substantial going on.


Finally, this cannot be only paper design. We need physical prototypes and proof that they exist and they make sense. With the amount of stupid stuff that people print on their 3D printers just for fun, including case designs that will be used for just one video on youtube, I would expect that bringing onboard 3D printing SFF-oriented youtubers to this is another angle to promote the idea. Also modding cards, older cards, to have the connectors where we want them, cut down the coolers and shrouds to size and so on - showing this thing is physically possible. Anything that can be a news that a group of people is doing things like this is a chance to bring a lot more people to read and sign.