WYP
News Guru
A blower card that's faster than Nvidia's Founders Edition?
Read more about Gigabyte's Aorus RTX 2080 Ti TURBO graphics card.
Read more about Gigabyte's Aorus RTX 2080 Ti TURBO graphics card.
Gigabyte releases high-end Aorus RTX 2080 Ti TURBO Blower-style GPU
- Thread starter WYP
- Start date
If it is anything like the ASUS variant, its only good for users of headphones. The biggest positive is dumping air outside the case. But the ASUS Turbo does create quite alot of noise on the 2080ti.
That said, when I compare my 1080ti reference to the gainward golden sample blower, the difference is exceptional good. Gainward made a fantastic blower, performing better at lower speed.
Perhaps same applies here. I'll always choose a blower cooler over those that dump heat inside the case.
Exactly my opinion. So if bemuzes me when I see people suggesting non blower style cards for SLI/Crossfire.
I like blowers, they have a great many uses in compact or external systems, allows isolation of CPU/GPU heat, and are pretty hard to break, but for a big open system an open cooling style with slower fans makes much more sense. BUT I'd much rather have the heat than noise. For all intents and purposes, the faster your fan has to run, the more likely your card is to have a hardware failure. The actual GPU chip on your card is one of the least likely parts to fail, the fan is the most likely. If you want to go as long as possible between having to do repairs on your parts, prioritise low fan speeds over increased temperatures. High fan speeds means faster ware of the most likely to fail part, more dust, more heat output to the system in total and ultimately far more servicing needed. The useful life span of a consumer chip technologically(How long you'd actually want to run games on it for) is far, far shorter than how long it could last being run at a high temperature without permanent damage. There's a lot of scaremongering I see with regards to heat that has some basis in reality but with issues that engineers mostly solved back in the 90's.
Heat does impact stability *IN THEORY*, but only temporarily; It's caused by increased heat energy in the chip increasing the thermal noise and raising the chance of a low voltage level being read as a high level. However, in modern processors there are so many things to mitigate or correct these errors, and they're so unlikely at temperatures below chip-frying tier, that it's a non-issue. Increased running temperatures *will not* affect long term stability *unless* the heat damages the transistors.
With a modern chip, damage isn't realistically going to occur to a chip below it's Tjunction unless you run it very near that temperature, consistently, for many many years. Tj's are generally in the 95-105C range for logic transistors (But can go far beyond that for MOSFETs in general, most power MOSFETs (Found in VRMs and the like) are rated for 125-150C Tjunction.
The fact is, regardless of what cooling you have, the same amount of thermal energy is being pumped out of your chip (Technically with fast fan based cooling more heat is output to the system in total due to the motors); The cooling just affects how long is takes to reach an equilibrium and where that equilibrium point is. Basically, I'd take a slow speed hot blower over a noisy cool blower any day, for both my short term benefits(ears) and long term(Less cutting fingers on heatsink fins while servicing fans/cleaning dust).
For reference, I'm using a HD7870XT that has a custom power limit buff of 25%(It can about 240W peak power draw). For much of its life it was a mining card (Ran full whack 24/7). It's 6 years old now, still runs as well as day one, can still reach base clock speeds with a 12% unervolt, can still reach 1.25Ghz while overclocked. What did I use for cooling through most of its life? a £3 12cm 3-pin fan zip tied to it. It'll soon be time to retire it, but because it's outlasted its useful lifespan.
Heat does impact stability *IN THEORY*, but only temporarily; It's caused by increased heat energy in the chip increasing the thermal noise and raising the chance of a low voltage level being read as a high level. However, in modern processors there are so many things to mitigate or correct these errors, and they're so unlikely at temperatures below chip-frying tier, that it's a non-issue. Increased running temperatures *will not* affect long term stability *unless* the heat damages the transistors.
With a modern chip, damage isn't realistically going to occur to a chip below it's Tjunction unless you run it very near that temperature, consistently, for many many years. Tj's are generally in the 95-105C range for logic transistors (But can go far beyond that for MOSFETs in general, most power MOSFETs (Found in VRMs and the like) are rated for 125-150C Tjunction.
The fact is, regardless of what cooling you have, the same amount of thermal energy is being pumped out of your chip (Technically with fast fan based cooling more heat is output to the system in total due to the motors); The cooling just affects how long is takes to reach an equilibrium and where that equilibrium point is. Basically, I'd take a slow speed hot blower over a noisy cool blower any day, for both my short term benefits(ears) and long term(Less cutting fingers on heatsink fins while servicing fans/cleaning dust).
For reference, I'm using a HD7870XT that has a custom power limit buff of 25%(It can about 240W peak power draw). For much of its life it was a mining card (Ran full whack 24/7). It's 6 years old now, still runs as well as day one, can still reach base clock speeds with a 12% unervolt, can still reach 1.25Ghz while overclocked. What did I use for cooling through most of its life? a £3 12cm 3-pin fan zip tied to it. It'll soon be time to retire it, but because it's outlasted its useful lifespan.
Last edited:
looz
Active member
Thinking of heat and high load as killers is a remnant from days with no thermal protections and poor quality caps.
Blower fans make sense for OEMs due to compatbility. They kinda make sense for multi-GPU but multi-GPU itself makes little sense in gaming - imo enthusiasts running multiple GPUs are better off just watercooling those setups.
For the chip itself a constant high load and temperature is better than constant thermal cycling caused by gaming load.
Blower fans make sense for OEMs due to compatbility. They kinda make sense for multi-GPU but multi-GPU itself makes little sense in gaming - imo enthusiasts running multiple GPUs are better off just watercooling those setups.
For the chip itself a constant high load and temperature is better than constant thermal cycling caused by gaming load.
Sorry tgrech, that was alot of waffle talk for something so simple.
I have yet to experience a faulty fan on a blower as opposed to the countless MSI 560ti/760 cards with their twin frozr III (or maybe iv) fans. Others can claim opposite.
As a headphone user who doesnt even notice the sound, I would take a blower easily, especially in SLI which has been proven to be a better solution than twin/triple fan solutions dumping heat onto card above. (even if the performance output is the same regardless)
In the end its personal preference. Both styles are being produced. However the popular choice is a single card solution and multi fan offer better aesthetics and more attractive designs. Blowers, well.. limited looks.
I am quite sure that the blower solution was intentionally designed long ago as Nvidia marketed SLI. Nowadays we all know its hardly required or beneficial and 3/4way is even dropped. And for that, Nvidia opted for twin fan.
There is even tests done showing m.2 drives throttling sooner and more often, due to 2 gpu as opposed to 1, or 2blowers when m2 is placed in between the cards.
dammit.. even im waffling on.
I have yet to experience a faulty fan on a blower as opposed to the countless MSI 560ti/760 cards with their twin frozr III (or maybe iv) fans. Others can claim opposite.
As a headphone user who doesnt even notice the sound, I would take a blower easily, especially in SLI which has been proven to be a better solution than twin/triple fan solutions dumping heat onto card above. (even if the performance output is the same regardless)
In the end its personal preference. Both styles are being produced. However the popular choice is a single card solution and multi fan offer better aesthetics and more attractive designs. Blowers, well.. limited looks.
I am quite sure that the blower solution was intentionally designed long ago as Nvidia marketed SLI. Nowadays we all know its hardly required or beneficial and 3/4way is even dropped. And for that, Nvidia opted for twin fan.
There is even tests done showing m.2 drives throttling sooner and more often, due to 2 gpu as opposed to 1, or 2blowers when m2 is placed in between the cards.
dammit.. even im waffling on.
AlienALX
Well-known member
Yup that is actually crazy. I saw my 670s (Jetstream) hitting 94c in the first TR refresh. I had to move the bottom card right down to the bottom, use a weird SLI ribbon and then hack it to work because that lane was only X4. It did sort the temps out though, but never again would I run multiple GPUs like that.
I've had one come a bit loose on the pole (3870x2) but it was easily fixed. You know why they're so reliable, right? because they are usually made my Delta or one of those other nutters who make fans that could take a finger off. They're as solid as they are reliable.
Last edited:
Yeah "heat leak" is definitely the most annoying part of an open design, but that mostly depends on your setup and can often be mitigated. But having a larger surface area for air to flow in and out of will always result in a higher cooling efficiency from the much lower static pressure required to get a given amount of airflow across the fins.
More motors/fans will always result in a higher chance of failure(The larger component count is the reason most high end products have higher failure rates than any lower end but equivalently manufactured counterparts), but not all open cooler GPUs are multi-fan and you can often find great single-fan open coolers(But multi-fan models can often run at slightly lower RPMs for the same airflow for lower noise per temp). The biggest reliability issue with a single open fan vs blower after RPM is peoples fingers, assuming a similar quality of motor.
More motors/fans will always result in a higher chance of failure(The larger component count is the reason most high end products have higher failure rates than any lower end but equivalently manufactured counterparts), but not all open cooler GPUs are multi-fan and you can often find great single-fan open coolers(But multi-fan models can often run at slightly lower RPMs for the same airflow for lower noise per temp). The biggest reliability issue with a single open fan vs blower after RPM is peoples fingers, assuming a similar quality of motor.
