Fractal Design Torrent - Bad GPU temperatures

[Milos Sladic]

• What is up with o11 XL being so much better than torrent even tho they have the same gpu cooling system (3 fans at the bottom)?
• And also many of these cases with better temps dont even have fans under the gpu?
• What is going on? This case should obliterate everything in the GPU department by the looks of it.

[Bombastinator]

Might be that the o11 is simply a much much larger case.  More surface area, made of aluminum, larger interior takes longer to heat up. Just making guesses.

[Milos Sladic]

4 minutes ago, Bombastinator said:

Might be that the o11 is simply a much much larger case.  More surface area, made of aluminum, larger interior takes longer to heat up. Just making guesses.

Well, what I know is that torrent is 1 liter larger case and also interior made of metal. So cant be that. 

[Bombastinator]

18 minutes ago, MS1993 said:

Well, what I know is that torrent is 1 liter larger case and also interior made of metal. So cant be that. 

If the interior is metal but the exterior is plastic it would act as an insulator. There wouldn’t be cooling via that route.  Interesting that one has a higher interior volume when it looks like a much smaller interior volume.  Perhaps it’s the PSU basement that does that.

[thrasher_565]

On 9/11/2021 at 11:11 AM, MS1993 said:

• What is up with o11 XL being so much better than torrent even tho they have the same gpu cooling system (3 fans at the bottom)?
• And also many of these cases with better temps dont even have fans under the gpu?
• What is going on? This case should obliterate everything in the GPU department by the looks of it.

gn only dose test with stock fans or a set of 3 nutura fans. 2 in front and 1 in back. that's it.

 

hmm but that dose look odd.

 

maybe there so much air from the front it pushes the bottom air out?

[Blue4130]

On 9/12/2021 at 3:30 AM, Bombastinator said:

If the interior is metal but the exterior is plastic it would act as an insulator. There wouldn’t be cooling via that route.  Interesting that one has a higher interior volume when it looks like a much smaller interior volume.  Perhaps it’s the PSU basement that does that.

The case itself has next to no cooling via the panels unless they are connected to the cooler via heatpipes or the like. 

[Bombastinator]

7 minutes ago, Blue4130 said:

The case itself has next to no cooling via the panels unless they are connected to the cooler via heatpipes or the like. 

Possible.  I don’t know what the percentages are.  An aluminum case will be a better heat conductor than a plastic coated metal one.  How much I don’t know though. Probably varies according to the temp differential between the inside of the case and ambient.  I do know that if such a thing were a building an aluminum panel vs plastic covered steel would be considered immense.  If the sizes are near equal and the radiant heating of the case is near equal it could be the basement, or possibly design of the case fan positions. The torrent does look a lot more squashed. The fan holes seem to run into each other. Physics remains physics.  Something has to give.  GN seems very concerned with testing methodology though so I doubt that is the issue. 

[TVwazhere]

On 9/11/2021 at 2:11 PM, MS1993 said:

• What is up with o11 XL being so much better than torrent even tho they have the same gpu cooling system (3 fans at the bottom)?
• And also many of these cases with better temps dont even have fans under the gpu?
• What is going on? This case should obliterate everything in the GPU department by the looks of it.

Firstly, this is just one chart of many tests, so this isnt the "be all end all" for the case being "Bad GPU temps". My guess has nothing to do with volume of the case or it's materials (Seriously?), but more to do with the airflow principle. air must come in and flow out. 

 

The Torrent has 5 Pc fans; all of them are MASSIVE intakes. What the case lacks (that nearly every other PC above it on the list has) is a exhaust fan. Ejecting hot air out of the case is just as important as pulling in fresh air. Sometimes you can brute force exhaust by just putting in more fans or spinning them faster forcing air out the back, but this method is inefficient at best, especially when you have 5 fans all trying to force air out of the rear, and three of them are doing it at a perpendicular angle to the exhaust (since there's not top vents to push out air).

 

But that doesn't mean this case NEEDS an exhaust fan. This case scores exceedingly well in all of the non-torture GPU intensive tests including Fire strike and Blender GPU renders.

 

A quote taken directly from the video:

Quote

"the GPU averaged 21º when using it to render instead, which is tied for the best result on the chart with the Lancool 2 Mesh and the P500A digital. GPU temperatures were good in the original [torture] test, but the firestrike and blender tests show that when loading the GPU in isolation, the Torrent excels"

Timestamped Video:

The CPU tests also were in the top 1%, if not chart topping (In the case of the blender CPU Render test) of all the metrics as well.

 

When both are working at full power simultaneously,  the Torrent seems to diminish a bit in it's overwhelming brute forcing of air, in part to what I think is a lack of an exhaust fan to help the hot air and the two walls of Fresh air perpendicular to each other to escape. I haven't seen anyone do a test with an exhaust fan, but for most users I'd say it's not needed as the case does well enough on it's own. 

[Jeppes]

GN uses an old gpu with their tests. Throw in a 450W 3090 instead of that 200W 1080 and you will see a larger difference and most likely slightly different order too on the charts.

[Dogzilla07]

@TVwazhere Yes, airflow performance is only there if a case flows out (exhaust), as much air as it flows in (and the fans on the CPU+GPU are in sync with the intake and exhaust fans (not exact RPM, but just not going too much faster or too much slower). However u don't need exhaust fans, to exhaust air outside of a case. If there is enough area opposite from intake, that's more than enough. With meticulous planning, some black tape and a proper setup exhaust fans are unnecessary outside small form-factor builds.

 

Brute-forcing it is actually the most efficient way, but there's some caveats. As you said have too many fans in the case at different angles is just gonna mess up the airflow and produce pockets re-circled air (air vortex). Also having places where air can escape near and around intake fans is not good (And clear front to back airflow is the best option).

 

That's the reason the Torrent does so well without an exhaust, the amount of holes and their size at the back is huge. That free-flowing exhaust area is more than enough of a replacement for an exhaust fan for the CPU. However there's still a bit more obstructions around the GPU and the bottom fans are not the best static pressure, so removing an additional above, and 1-2 below PCI Express metal brackets would probably equalize the case with the others.

 

What's probably happening is the combination of not exactly stellar static pressure and than first of the 3 fan closest to the front, translates into weird flow around the gpu, and the gpu recycles it's own air before it can escape out the back fast enough (hence removing PCI-E brackets, to see if that helps with transfering the heat out of the nearby air.

 

However the front fans, while a really good piece of technology are pretty loud and probably don't flow air properly in combination with the bot fans (for the GPU), they flow air more than adequately for the CPU. I got a suspicion that even those fan fast 180mm would be behind 3x120/140 static pressure optimised LCP fans like the (Phanteks T30-120/140), Noctua A12x25 and Scythe Wonder Snail). 3 of any of those in the front, and 1-2 in the bottom maybe (but possible even redundant). It would be imperative that the CPU heatsinks had similar fans at similar speeds, and ideally a de-shroud non-invasive mod on the GPU with same fans zip-tied (but some of the biggest 3080TI/3090/6900XT with 100+mm stock fans should be fine).

 

The case is a marvel of technology, it's the first of it's kind with a proper huge unobstructed back area (the extremely clean area of the back fan exhaust mounting place, and a proper focus on the future of front intake. (It's just that Fractal needs these new technology super-material constructed fans ASAP, their current stuff will just not cut it for the second part of the current decade).

 

I would have expected the lack of restriction of the area around the GPU to narrow the gap more i mean look at these holes, even the PCi-E bracket ones are huge. And I'm very interested how much of a difference removing brackets would be, but an optimization with different fans and fan configuration might be more productive:

 

[Milos Sladic]

Just wanted to thank you all for your knowledge. I've learned a lot and I will definitely add 1 more fan to the exhaust for the best possible thermals this case can provide!

[thrasher_565]

12 hours ago, TVwazhere said:

Firstly, this is just one chart of many tests, so this isnt the "be all end all" for the case being "Bad GPU temps". My guess has nothing to do with volume of the case or it's materials (Seriously?), but more to do with the airflow principle. air must come in and flow out. 

 

The Torrent has 5 Pc fans; all of them are MASSIVE intakes. What the case lacks (that nearly every other PC above it on the list has) is a exhaust fan. Ejecting hot air out of the case is just as important as pulling in fresh air. Sometimes you can brute force exhaust by just putting in more fans or spinning them faster forcing air out the back, but this method is inefficient at best, especially when you have 5 fans all trying to force air out of the rear, and three of them are doing it at a perpendicular angle to the exhaust (since there's not top vents to push out air).

 

But that doesn't mean this case NEEDS an exhaust fan. This case scores exceedingly well in all of the non-torture GPU intensive tests including Fire strike and Blender GPU renders.

 

A quote taken directly from the video:

The CPU tests also were in the top 1%, if not chart topping (In the case of the blender CPU Render test) of all the metrics as well.

 

When both are working at full power simultaneously,  the Torrent seems to diminish a bit in it's overwhelming brute forcing of air, in part to what I think is a lack of an exhaust fan to help the hot air and the two walls of Fresh air perpendicular to each other to escape. I haven't seen anyone do a test with an exhaust fan, but for most users I'd say it's not needed as the case does well enough on it's own. 

i have a problem with this guess. the augment here is the o11 XL vs the torrent and gpu temps. both case intake from the bottom but the 011xl the front intake as to do a 90 degree well the torrent dose not have to do that so...the torrent should win the cpu cooling test.

 

the torrent should have so much air from the front and the back being so open i dont think a back fan is need but i guess that could be a test done.

 

i think both dust filters on both of of the case should be fine i think? but i think he would remove them  for testing?

 

not only that the 011xl comes with no fans so in that case he will use his 3 fans 2 front 1 back.

Edited September 14, 2021 by thrasher_565

[thrasher_565]

12 hours ago, Jeppes said:

GN uses an old gpu with their tests. Throw in a 450W 3090 instead of that 200W 1080 and you will see a larger difference and most likely slightly different order too on the charts.

but he use the same gpu in every case so... i do agree he should get in the the 3090 and case testing but i dont no if he even knose about these problems.

[thrasher_565]

10 hours ago, Dogzilla07 said:

@TVwazhere Yes, airflow performance is only there if a case flows out (exhaust), as much air as it flows in (and the fans on the CPU+GPU are in sync with the intake and exhaust fans (not exact RPM, but just not going too much faster or too much slower). However u don't need exhaust fans, to exhaust air outside of a case. If there is enough area opposite from intake, that's more than enough. With meticulous planning, some black tape and a proper setup exhaust fans are unnecessary outside small form-factor builds.

 

Brute-forcing it is actually the most efficient way, but there's some caveats. As you said have too many fans in the case at different angles is just gonna mess up the airflow and produce pockets re-circled air (air vortex). Also having places where air can escape near and around intake fans is not good (And clear front to back airflow is the best option).

 

That's the reason the Torrent does so well without an exhaust, the amount of holes and their size at the back is huge. That free-flowing exhaust area is more than enough of a replacement for an exhaust fan for the CPU. However there's still a bit more obstructions around the GPU and the bottom fans are not the best static pressure, so removing an additional above, and 1-2 below PCI Express metal brackets would probably equalize the case with the others.

 

What's probably happening is the combination of not exactly stellar static pressure and than first of the 3 fan closest to the front, translates into weird flow around the gpu, and the gpu recycles it's own air before it can escape out the back fast enough (hence removing PCI-E brackets, to see if that helps with transfering the heat out of the nearby air.

 

However the front fans, while a really good piece of technology are pretty loud and probably don't flow air properly in combination with the bot fans (for the GPU), they flow air more than adequately for the CPU. I got a suspicion that even those fan fast 180mm would be behind 3x120/140 static pressure optimised LCP fans like the (Phanteks T30-120/140), Noctua A12x25 and Scythe Wonder Snail). 3 of any of those in the front, and 1-2 in the bottom maybe (but possible even redundant). It would be imperative that the CPU heatsinks had similar fans at similar speeds, and ideally a de-shroud non-invasive mod on the GPU with same fans zip-tied (but some of the biggest 3080TI/3090/6900XT with 100+mm stock fans should be fine).

 

The case is a marvel of technology, it's the first of it's kind with a proper huge unobstructed back area (the extremely clean area of the back fan exhaust mounting place, and a proper focus on the future of front intake. (It's just that Fractal needs these new technology super-material constructed fans ASAP, their current stuff will just not cut it for the second part of the current decade).

 

I would have expected the lack of restriction of the area around the GPU to narrow the gap more i mean look at these holes, even the PCi-E bracket ones are huge. And I'm very interested how much of a difference removing brackets would be, but an optimization with different fans and fan configuration might be more productive:

 

hmm so probably stopping the front bottom fan will probably fix the problem.

 

i guess if you got front air flow for the gpu that can mess with the bottom intake. i would think bottom intake would be better then front intake.

Edited September 14, 2021 by thrasher_565

[thrasher_565]

not only that i dont see a case standardize fan test for the torrent witch means its not apples to apples

[TVwazhere]

10 hours ago, thrasher_565 said:

ii think both dust filters on both of of the case should be fine i think? but i think he would remove them  for testing?.

Filters are left on for stock tests, the Torrent review doesnt specify the filter being removed despite its ability to be removed, however the full fornt panel removed drooped temps 3ºC so we can assume most of that restriction came from the filter. They made no indication of bottom filter removal for the GPU tests, so they may drop some as well. Same could be said for the O11D XL, however both cases came with exceptionally low restriction dust filters, so as long as both have them on or both have them off, it's a more fiar competition (but not totally apples to apples as you note below:

10 hours ago, thrasher_565 said:

not only that i dont see a case standardize fan test for the torrent witch means its not apples to apples

They did that test (It's in the video), but had a lengthy preamble saying "these are gonna suck because 1/2 the cost and reason you buy this case are for the fans, so using any other fan configuration is just silly"

 

In both instances, the case performed better with the stock layout. The O11XL 's stock fans were placed at the bottom, while the Torrents were placed at the front. Hence the torrents "dominance" over the o11XL and the O11 XL's "dominance" over the Torrent. Both cases are good, it just depends on whether you want fan flexibility or control, or you want a wall of air out of the box (and there's definitely markets for both)

[Dogzilla07]

10 hours ago, thrasher_565 said:

i would think bottom intake would be better then front intake.

Front Intake is still better for the GPU, because a sufficient high flow of air will push out the hot air faster above adn below the gpu, and stop the GPU fans from recycling the hot air exhaused by the gpu (and yes i know air doesn't get pushed, and that it's heat displacement, but it's easier to explain this way xD)

 

But yeah too many fans in general, and too many bottom fans with weak static pressure is not ideal and messes with the airflow.

[Dogzilla07]

33 minutes ago, TVwazhere said:

They did that test (It's in the video), but had a lengthy preamble saying "these are gonna suck because 1/2 the cost and reason you buy this case are for the fans, so using any other fan configuration is just silly"

Specifically what that means, is a problem with the GamersNexus test is  that it's only valid with the same CPU cooler + Case combination they're doing. A different Cooler with different amount of fans and speed is gonna behave differently and produce different heat vortices around the CPU heatsink itself and the GPU.

 

And the front 2x Noctua intake is paramount for a standardized proper methodology, but while it make work good in one case with one GPU, it probably won't in another. 

 

CPU temps you can take at face value, that you won't be able to brute-force, or differently optimize much better, but GPU temp is another matter. Taking GPU temp for the standardized + noise-normalized test and cross-referencing it is a good clue what to expect, but figuring out the pros and cons of your own setup you can optimized it way further. It's just that it is physically impossible to standardize a methodology that replicates how much each user can optimized their setup with better static pressure fans, and better taped around and parts removed from their case, there's just too many variables.

 

Any high GPU temp, or high discrepancy between GPU and CPU temps you see in GamersNexus results for a case does not in any way mean that that is the best that case can do, or that it's immutable. There is room to lower GPU temps by 3-10 degrees in almost any situation, and only with soft modding without altering the case permanently.

 

 

[TVwazhere]

1 minute ago, Dogzilla07 said:

Specifically what that means, is a problem with the GamersNexus test is  that it's only valid with the same CPU cooler + Case combination they're doing.

That's.... exactly the point of the tests. They're testing the case against case, not the case with a certain CPU and GPU combo. 

2 minutes ago, Dogzilla07 said:

And the front 2x Noctua intake is paramount for a standardized proper methodology, but while it make work good in one case with one GPU, it probably won't in another.

That's covered EXTENSIVELY in their "we didnt want to do this test but people wouldnt stop asking for it" standardized test video

5 minutes ago, Dogzilla07 said:

Any high GPU temp, or high discrepancy between GPU and CPU temps you see in GamersNexus results for a case does not in any way mean that that is the best that case can do, or that it's immutable. There is room to lower GPU temps by 3-10 degrees in almost any situation, and only with soft modding without altering the case permanently.

I don't believe anyone has said or indicated otherwise, but thank you for confirming. 

[Dogzilla07]

16 minutes ago, TVwazhere said:

I don't believe anyone has said or indicated otherwise, but thank you for confirming. 

I know but I wanted to write it out anyway for all the lurkers too xD, and specifically mention that there's much more room to improve for the GPU temp, than the CPU temp, and that it's more important to pick out a case by its CPU temp ranking than the GPU ranking (because it's counter-intuitive a bit, cause u'd think the GPU would be a priority) , but u can instead just tinker the GPU temp away.

 

And yes, I absolutely agree with both previous quote points, how they're doing it is the best u can realistically do it (And Liquid-Crystal Polymer fans make a sizeable difference, but it's not viable to change their platform), every few years the amount how much more (or temporary less due to thermal density/chiplets) u can get out of a case changes due to new technology but it won't show up in the test because they have to use what they started with.

 

End-game is a Sigma 6 accuracy-like AI Super-Computer Real-TIme Fluid Dynamics Simulation of all possible combinations of hardware in a "closed" system (case), for cheap available to everyone with a intuitive UI where u can just move stuff around in AR, and get instant temps for a chosen game/scenario+climate+ambient temp/time-frame. But that's gonna take a while xD

[Chunky Moe]

What I don't understand is how the Phantek P500A Digital is scoring so high on GPU when it has no bottom airflow?
 

[freeagent]

Is this noise normalized? If so than that sucks

 

Moving air makes noise

 

Typing this out my PC is quiet ish but when its busy the fans spool right up and I honestly don't care how loud it is.. as long as it performs the way I expect it to 

 

But its about 8 feet from me so.. in line with my tv sitting on my sub

[thrasher_565]

according to the graph they used standardized case fan test witch is 2 fans in front and on in back witch would faver a tower cooler and no bottom mounted fans for the gpu. weather this is tru or not i dont no. you wouldn't by this case to then rip out its fans and only mount 3 fans in there unless you can buy the case with out fans?

[thrasher_565]

On 9/14/2021 at 5:46 AM, Dogzilla07 said:

Front Intake is still better for the GPU, because a sufficient high flow of air will push out the hot air faster above adn below the gpu, and stop the GPU fans from recycling the hot air exhaused by the gpu (and yes i know air doesn't get pushed, and that it's heat displacement, but it's easier to explain this way xD)

 

But yeah too many fans in general, and too many bottom fans with weak static pressure is not ideal and messes with the airflow.

no its not... bottom intake is the best for gpus as that's how gpu cooling works... the torrent dose have an advantage that the back being so open and so much front air can push trapped hot air out but with it having a bottom front fan this will interfere with the flow.

 

the air coming from the front the has to do a 90 degree in to the gpu. then it exhaust out the sides at the side panel and mb some will get sucked back in to the gpu well some makes its way up and normally out the back fan but the case being open it can just go out the pci slots.

 

the Cooler Master MasterCase SL600M being the king when it comes to gpu air flow as you get 2 massive fans blowing on the gpu. but dose vary pore with a cpu tower cooler and the top is closed off mostly... so an top aio would be ideal in that case and passably mod the top panel.

 

 

Edited October 1, 2021 by thrasher_565

[Bombastinator]

41 minutes ago, thrasher_565 said:

no its not... bottom intake is the best for gpus as that's how gpu cooling works... the torrent dose have an advantage that the back being so open and so much front air can push trapped hot air out but with it having a bottom front fan this will interfere with the flow.

 

the air coming from the front the has to do a 90 degree in to the gpu. then it exhaust out the sides at the side panel and mb some will get sucked back in to the gpu well some makes its way up and normally out the back fan but the case being open it can just go out the pci slots.

 

the Cooler Master MasterCase SL600M being the king when it comes to gpu air flow as you get 2 massive fans blowing on the gpu. but dose vary pore with a cpu tower cooler and the top is closed off mostly... so an top aio would be ideal in that case and passably mod the top panel.

 

 

If the PSU isn’t in the way, anyway.