Reference cool vs. Aftermarket

[ripper101]

So I was watching some videos on the MSI GTX 980ti Gaming 6G and I noticed when comparing it to a titan x with it's reference cooler the titan x was noticably quiet. So is reference coolers from nvidia better than aftermarket coolers in the sense that they're more quiet?

 

 

 

[suchamoneypit]

yeah and how hot does that Titan X run compared to the 980 ti.

[Guest]

The reference is a blower design (pushes air out the back of the card). Blower designs are often much louder than open ones.

 

The video is awful in all honesty as the static sound makes it impossible to actually hear the cards.

 

[MegaDave91]

That Titan X is nowhere near full RPM. I never rely on default fan mode because it makes my Titan X uncomfortably hot.

[drewjn]

@ripper101

3 major issues with the comparison:

 

1. If you are comparing reference with aftermarket, you should be using the same card type.

2. Different aftermarket coolers are different with the same card. The differences are based on sound, efficiency and design.

3. What matters most in a comparison is also under what load the card is on, before any comparison can be made. 

 

The question isn't just noise, but also what temp it runs, what sale is currently going on, and preference. So you can't really claim one is better on such an arbitrary point.

[ripper101]

1 hour ago, drewjn said:

@ripper101

3 major issues with the comparison:

 

1. If you are comparing reference with aftermarket, you should be using the same card type.

2. Different aftermarket coolers are different with the same card. The differences are based on sound, efficiency and design.

3. What matters most in a comparison is also under what load the card is on, before any comparison can be made. 

 

The question isn't just noise, but also what temp it runs, what sale is currently going on, and preference. So you can't really claim one is better on such an arbitrary point.

Yeah because the Titan X and the 980ti are SO different

[David_K]

Still, fan profiles are different. and the noise of the cards needs to be compared with sound measurement tools, so you could compare how actually one is louder than the other, as in a video you can really know. 

[drewjn]

11 minutes ago, ripper101 said:

Yeah because the Titan X and the 980ti are SO different

When you can actually compare two cards quite easily of the same chip set, you should. Just because they are somewhat similar in performance is an extremely poor way to argue whatever arbitrary point you are trying to use to defend a poorly conceived topic.

[ripper101]

52 minutes ago, drewjn said:

When you can actually compare two cards quite easily of the same chip set, you should. Just because they are somewhat similar in performance is an extremely poor way to argue whatever arbitrary point you are trying to use to defend a poorly conceived topic.

The 980ti is yes, not only the same chipset but, basically the same chip as the Titan X. Not to mention you said because they have similar performance it means they're not comparable, too bad they have the same Maxwell architecture in both cards resulting in performance between cards with the same architecture and cores is a VERY comparable topic. But this therefor sets a solid foundation for getting a rough estimate of what a 980ti with a reference cooler would perform like. I wouldn't name this a "poorly conceived topic" due to the fact that this post was out of curiosity because of the videos posted show the aftermarket cooler being louder then a reference cooler of almost the same chip. But yes, now I know that it was the video's problem because of the microphone used in them and the aftermarket coolers are usually quieter then the reference cooler designs.

[drewjn]

51 minutes ago, ripper101 said:

The 980ti is yes, not only the same chipset but, basically the same chip as the Titan X. Not to mention you said because they have similar performance it means they're not comparable, too bad they have the same Maxwell architecture in both cards resulting in performance between cards with the same architecture and cores is a VERY comparable topic. But this therefor sets a solid foundation for getting a rough estimate of what a 980ti with a reference cooler would perform like. I wouldn't name this a "poorly conceived topic" due to the fact that this post was out of curiosity because of the videos posted show the aftermarket cooler being louder then a reference cooler of almost the same chip. But yes, now I know that it was the video's problem because of the microphone used in them and the aftermarket coolers are usually quieter then the reference cooler designs.

 

Yes, and do you compare the 950, 960, 970 and 980 as the same exact card? They are all optimized differently, with different cuda core counts. Just because they are of the same architecture is not the same as making a direct comparison when regards to efficiency and cooling when comparing the heatsink and coolers attached. Hell, maybe we should say that all cpus of the same class are entirely identical. Disregarding any changes as they obviously should benchmark the exact same, with the same temps.

 

The 980 Ti is 6gb onboard with 22 sm units and 2816 cuda cores, the Titan X is 24 sm units with 3072 cuda cores with 12gb memory. Yes they must be absolutely identical and oc the exact same way with the exact same temps. We aren't arguing between an apple and an orange, sure; however, you must realize that when making an actual comparison between aftermarket and reference, you should do so using the same card. There isn't a reason to attack another person on a forum because of whatever personal issues. Keep in mind, you started the flaming.

[Guest]

Compared at the same fan speed/temperature the aftermarket cooler will blow (lulz) the reference cooler away. 

[ripper101]

3 hours ago, drewjn said:

 

Yes, and do you compare the 950, 960, 970 and 980 as the same exact card? They are all optimized differently, with different cuda core counts. Just because they are of the same architecture is not the same as making a direct comparison when regards to efficiency and cooling when comparing the heatsink and coolers attached. Hell, maybe we should say that all cpus of the same class are entirely identical. Disregarding any changes as they obviously should benchmark the exact same, with the same temps.

 

The 980 Ti is 6gb onboard with 22 sm units and 2816 cuda cores, the Titan X is 24 sm units with 3072 cuda cores with 12gb memory. Yes they must be absolutely identical and oc the exact same way with the exact same temps. We aren't arguing between an apple and an orange, sure; however, you must realize that when making an actual comparison between aftermarket and reference, you should do so using the same card. There isn't a reason to attack another person on a forum because of whatever personal issues. Keep in mind, you started the flaming.

200 cuda cores is nothing. Amount of vram has no effect on GPU temps. What I meant by the same architecture it is the same GPU as the Titan X with the same architecture just with less things like cuda cores. Basically we are looking at a very small difference in cores with the SAME architecture which will probably have a few degrees in temperature. But what I don't get is how I said I wanted a ROUGH estimate of what the noise and temps would be like, and I got it, thanks.