.jpg "Raja"){kind=avatar}
cekeu wrote:
ROG-MATRIX-RTX2080TI-P11G-GAMING :
https://www.asus.com/ROG-Republic-Of-Gamers/ROG-MATRIX-RTX2080TI-P11G-GAMING/overview/
Yes, that's very nice looking GPU.
Flat nice edges, big card. I don't understand still how she work.
Looks like she have small dual radiator instead heatsinks.
For sure revolutionary model of graphic card.
Without Poseidon and in some different situation, with a lot of cheaper Turing, I would be happy with this card, very happy.
But that was dream, when I prepared 800 euro with plan to add max 100 euro more for newer and much powerful GPU than Pascal and better for 4K. NVIDIA had different plans.
Some video clips say 2018 is one of worse year for PC enthusiasts. I don't know for others but that's year when I couldn't afford any more to buy
after 2-3-4 years new premium GPU as I used on more than 15 years before.
I think owners will be very happy if you can afford.
I must satisfied with my Poseidon. Now I see price of waterblock drop to 110 euro for him but he is more beutiful in default condition.
I can't say who is nicer Poseidon or Matrix.
I need 150-180 euro more to finish loop.
Second radiator GTS 360 and three more Noctua NF-A12x25 PWM.
But I'm so impatient... very soon I will have rest of this stuff.
Noctua Fans are expensive but worth.
Almost every year ASUS bring some innovation and many brands later adopt that silently next year.
We will see how this model work in next months. But It would be nice to someone build full cover waterblock.
I can bet temps are much better than Strix and I suppose that's task to leave space for higher boost.
Second radiator GTS 360 and three more Noctua NF-A12x25 PWM.
But I'm so impatient... very soon I will have rest of this stuff.
Noctua Fans are expensive but worth.
Almost every year ASUS bring some innovation and many brands later adopt that silently next year.
We will see how this model work in next months. But It would be nice to someone build full cover waterblock.
I can bet temps are much better than Strix and I suppose that's task to leave space for higher boost.
What processor coolers do is take the heat from a very concentrated source of thermal energy (the processors) and rapidly transport it to a large surface area that can better convect the energy away with the surrounding air. There are a few ways by which this is accomplished:
- conduction cooling through a solid heat sink - used in cheap, low power solutions, conduction is the most efficient means to transport thermal energy but is subject to thermal resistance that results in thermal gradients within the material that end up with higher temperatures in high power applications;
- convection cooling with fluid phase change through heat pipes or vapor chambers - structures with cavities in the (heat pipes and vapor chambers) have a working fluid in them. Around the heat source, the fluid is vaporized and buoyancy of the fluid moves it away from the heat source through the structure to areas that conduct energy to fins to increase the surface area with the surrounding air (all of this is by design);
- convection cooling with water (without phase change) through a thermal engine (water block) - while a phase change is more efficient way to extract heat, a pumped fluid forces it through the thermal engine, replacing the water that has absorbed some thermal energy with cooler water, keeping the thermal gradient between the thermal engine and working fluid at a maximum.
Also to consider is where the heat is being transferred to the surrounding air. In an air-cooled system with a conduction or convection with heat pipes or vapor chamber, the fins are located directly over the processor. That requires air to be directed toward those particular airs. Because these are the least expensive cooling options and are most prevalent, motherboards and cases are designed with such provisions in mind. However, this is still not the most ideal set-up, as there are compromises with signal integrity of electrical traces in the motherboard and increasingly aesthetic trade-offs with case design (e.g. solid tempered glass side glass immediately over the coolers) where packaging results in recycled hot air or limited cool air supply to begin with.
A water cooling loop, on the other hand, allows for transport of the thermal energy to remote locations, like the periphery of the case, where radiators transfer the thermal energy to the surrounding air. They still have to contend with the lower efficiencies of heat transfer from the processor to the working fluid, but can gain in overall performance with better working fluid temperatures and heat dispersion to the surrounding air.
The new Matrix cooling solution requires the air to flow to the graphics card, just as air cooled cards do. However, they are running a less efficient closed (water) loop cooler (CLC) in that space. That CLC is quite large, filling up three slots and extra tall, further impeding air flow to the cooler and restricting air flow through the case. There may still be performance seen out of the card, but it would lack the efficiency and effectiveness of a Poseidon card.
The Poseidon has a very elegant solution for cooling. It uses a high efficiency vapor chamber to carry thermal energy away from the processor. There are fins for the air cooling, as you would expect on an air-cooled card, but there is also a water tube from the terminal block on top of the card, flowing cool water from an open water loop, to augment the cooling of the fins.
Therefore, the card uses the space that is designed for air-cooled cards to use for cooling. But that cooling is supplemented by the cooling of an open-loop water cooling system. Both cooling solutions can pick up the slack if the other ever falters, but in conjunction they work sublimely together.
Asus has already done the hard part of the engineering to get this solution to work. It would be a shame if they were to put it out to pasture and not implement it anymore. They could improve on the previous design by taking the larger vapor chambers that Nvidia uses in their reference design and their competitors are using in their air-cooled solutions. Here is a vapor chamber used in a competitors solution:
If they could combine a large vapor chamber like that with a cooler that remains within 18mm of the PCB, including the terminal block on top for the open-loop water cooling attachments, and no thicker than 2-slots, allowing for excellent air flow to and around the card, that would probably be the most ideal configuration for a performance GPU.
- conduction cooling through a solid heat sink - used in cheap, low power solutions, conduction is the most efficient means to transport thermal energy but is subject to thermal resistance that results in thermal gradients within the material that end up with higher temperatures in high power applications;
- convection cooling with fluid phase change through heat pipes or vapor chambers - structures with cavities in the (heat pipes and vapor chambers) have a working fluid in them. Around the heat source, the fluid is vaporized and buoyancy of the fluid moves it away from the heat source through the structure to areas that conduct energy to fins to increase the surface area with the surrounding air (all of this is by design);
- convection cooling with water (without phase change) through a thermal engine (water block) - while a phase change is more efficient way to extract heat, a pumped fluid forces it through the thermal engine, replacing the water that has absorbed some thermal energy with cooler water, keeping the thermal gradient between the thermal engine and working fluid at a maximum.
Also to consider is where the heat is being transferred to the surrounding air. In an air-cooled system with a conduction or convection with heat pipes or vapor chamber, the fins are located directly over the processor. That requires air to be directed toward those particular airs. Because these are the least expensive cooling options and are most prevalent, motherboards and cases are designed with such provisions in mind. However, this is still not the most ideal set-up, as there are compromises with signal integrity of electrical traces in the motherboard and increasingly aesthetic trade-offs with case design (e.g. solid tempered glass side glass immediately over the coolers) where packaging results in recycled hot air or limited cool air supply to begin with.
A water cooling loop, on the other hand, allows for transport of the thermal energy to remote locations, like the periphery of the case, where radiators transfer the thermal energy to the surrounding air. They still have to contend with the lower efficiencies of heat transfer from the processor to the working fluid, but can gain in overall performance with better working fluid temperatures and heat dispersion to the surrounding air.
The new Matrix cooling solution requires the air to flow to the graphics card, just as air cooled cards do. However, they are running a less efficient closed (water) loop cooler (CLC) in that space. That CLC is quite large, filling up three slots and extra tall, further impeding air flow to the cooler and restricting air flow through the case. There may still be performance seen out of the card, but it would lack the efficiency and effectiveness of a Poseidon card.
The Poseidon has a very elegant solution for cooling. It uses a high efficiency vapor chamber to carry thermal energy away from the processor. There are fins for the air cooling, as you would expect on an air-cooled card, but there is also a water tube from the terminal block on top of the card, flowing cool water from an open water loop, to augment the cooling of the fins.
Therefore, the card uses the space that is designed for air-cooled cards to use for cooling. But that cooling is supplemented by the cooling of an open-loop water cooling system. Both cooling solutions can pick up the slack if the other ever falters, but in conjunction they work sublimely together.
Asus has already done the hard part of the engineering to get this solution to work. It would be a shame if they were to put it out to pasture and not implement it anymore. They could improve on the previous design by taking the larger vapor chambers that Nvidia uses in their reference design and their competitors are using in their air-cooled solutions. Here is a vapor chamber used in a competitors solution:
If they could combine a large vapor chamber like that with a cooler that remains within 18mm of the PCB, including the terminal block on top for the open-loop water cooling attachments, and no thicker than 2-slots, allowing for excellent air flow to and around the card, that would probably be the most ideal configuration for a performance GPU.
