RTX 5090 - $2000 - 2 Slot Design - available on Jan. 30 - 12VHPWR still

mullet said:
That's funny, ohh ok fine let that massive heat go through your upper radiator then. I'm sure the saturation will happen faster now.
My air cooling does just fine. Super easy to clean as well. no worries about pump failure.
Click to expand...
It's not really an issue, a single 140mm radiator can handle between 300-400w depending on the thickness and your fan static pressure before you saturate it, assuming somebody using a double or triple fan radiator for their CPU has nothing to worry about unless their cases internal temp is pushing into the 60+ degree C range at which point there are a multitude of other issues going on.

The reality is any case with decent airflow will be moving enough air that having the GPU exhaust being taken into the CPU radiator would have a marginal impact at worst, and likely result in nothing more than a slight increase in fan speeds with no change in temperatures.
 
Lakados said:
It's not really an issue, a single 140mm radiator can handle between 300-400w depending on the thickness and your fan static pressure before you saturate it, assuming somebody using a double or triple fan radiator for their CPU has nothing to worry about unless their cases internal temp is pushing into the 60+ degree C range at which point there are a multitude of other issues going on.

The reality is any case with decent airflow will be moving enough air that having the GPU exhaust being taken into the CPU radiator would have a marginal impact at worst, and likely result in nothing more than a slight increase in fan speeds with no change in temperatures.
Click to expand...

Temps will easily increase, as the air is carrying more heat, reducing the capability of the air to absorb more heat. Considering most modern hardware overclocks itself based on temperature, it quickly becomes a issue. Also the longer you play the more the ambient air starts to heat up as well, which will once again increase overall temps for the cpu. If your in a huge office building than likely not a big deal, but a little office in a house or small room is going to become a issue. A big part is how much wattage is your cpu producing, if your using 300 watts for the cpu that cooler is being pushed already.
 
Gideon said:
Temps will easily increase, as the air is carrying more heat, reducing the capability of the air to absorb more heat. Considering most modern hardware overclocks itself based on temperature, it quickly becomes a issue. Also the longer you play the more the ambient air starts to heat up as well, which will once again increase overall temps for the cpu. If your in a huge office building than likely not a big deal, but a little office in a house or small room is going to become a issue. A big part is how much wattage is your cpu producing, if your using 300 watts for the cpu that cooler is being pushed already.
Click to expand...
Not really, the CPU AIO push or pull debate was done a while back by a few different YouTube sites and they all pretty much agreed that there was no noticeable performance difference, they found sound differences but system temps changed by less than a degree in most cases.
Any decent PC case is moving more than enough air that it isn't a realistic problem unless the case itself has some fundamental airflow problems.
 
Lakados said:
Not really, the CPU AIO push or pull debate was done a while back by a few different YouTube sites and they all pretty much agreed that there was no noticeable performance difference, they found sound differences but system temps changed by less than a degree in most cases.
Any decent PC case is moving more than enough air that it isn't a realistic problem unless the case itself has some fundamental airflow problems.
Click to expand...

You know when they do those tests, they keep the room the same exact temperature, you are not going to be able to do that in a small space and game for hours. If ambient temps rise, than guess what your AIO temps will rise as well and so will your cpu temps.
 
Gideon said:
You know when they do those tests, they keep the room the same exact temperature, you are not going to be able to do that in a small space and game for hours. If ambient temps rise, than guess what your AIO temps will rise as well and so will your cpu temps.
Click to expand...
OK but if your ambient room temperature rises to a place where it could be a problem, that problem exists regardless of your chosen cooling solution and fan configuration so I am not sure what argument you are trying to make.
 
mullet said:
And I have never had to RMA a ASUS board for 25 years.
Click to expand...
I had to RMA my first, an A7V333-X. PATA controller died on the southbridge. Was the only mobo I've ever had that died. Turned around in less than two weeks when they still had their Louisville service center. Only part I've ever had go bad.
That, and a BFG 6800GT, but that may have been me over tightening a Danger Den block.
 

View: https://youtu.be/rCwgAGG2sZQ

Apparently the giant 750-760mm and core count that come with it even if it is diminishing rewards... are needed to make 512 bits possible, 100 of the border being used for the memory controller connections to make this work.

i.e. trying to come up with well calculated marketing-business type decision when looking at a sku line up is always dangerous to over analysis, yes when they decide price lot of thinking goes into those, but core count, die size, etc... a lot of them will be decided by force we do not think about....

They still have hardware optical flow engine it seem (near the NVENC, maybe some encoder-decoder use them) so people that have them on Lovelace, they will probably still be used, they seem quite small regardless.

That picture of avrey expensive cpu complete CCD relative to the 5090 die size is a good reminder of how much more you are buying there
 
Last edited:
LukeTbk said:

View: https://youtu.be/rCwgAGG2sZQ

Apparently the giant 750-760mm and core count that come with it even if it is diminishing rewards... are needed to make 512 bits possible, 100 of the border being used for the memory controller connections to make this work.

i.e. trying to come up with well calculated marketing-business type decision when looking at a sku line up is always dangerous to over analysis, yes when they decide price lot of thinking goes into those, but core count, die size, etc... a lot of them will be decided by force we do not think about....

They still have hardware optical flow engine it seem (near the NVENC, maybe some encoder-decoder use them) so people that have them on Lovelace, they will probably still be used, they seem quite small regardless.
Click to expand...

Optical flow is a component of the NVENC engine, it is and has always been used by Nvidia for a few of their processes. They use it in their video encoding engine when doing changes in frame rates, be it taking something from 30-60 fps or smoothing out slow motion video.
The accelerators just make it a hell of a lot faster to do, and the initial DLSS frame gen engine was built ontop of that software stack, so it used it too. Maxwell, Pascel, and Volta did what Nvidia now refers to as Motion Estimation Only Mode which analyzed the frames and tried to guess the direction of motion. It wasn’t always accurate and video footage where the subject was naturally motion blurred because of its capture medium had some interesting results, you needed a lot of filtering, and it took a fair while to analyze and render the middle frame. The accelerator is what made it viable to use in game, but a nice feature if you are only doing video work.
 
Last edited:
You must log in or register to reply here.
Back
Top