new case

It's not so much about limited space but more about as technology evolves, and things get smaller and thinner

Well if you've been paying attention you'll know that's not true. Just for fun try comparing a GTX 1080 to a GT 630.

Components get smaller (transistors, etc.) But as the whole gets more complicated it gets bigger and less efficient, which requires more space and cooling (a 630 doesn't even have a cooler). That's why full towers exist.
 
I'm positive a GTX970 runs cooler than a GTX280, and a 970 is a lot more complicated + a hell of a lot more powerful than a 280. So complexity does come into it but not so much.

ATX is indeed a better choice for cooling, and bigger will always be more powerful (you can always fit more components into a bigger space). For most people though it comes down to requirements. Hence why so many people now have laptops instead of a desktop case that comes up to their hip!
 
I'm positive a GTX970 runs cooler than a GTX280, and a 970 is a lot more complicated + a hell of a lot more powerful than a 280. So complexity does come into it but not so much.

The 280 uses 60% more power (max). Of course it'll run hotter.

The GTX 200 family is also particularly hot. Grab the 550 and put it against a 1080ti. There is a considerable power consumption increase (150W to 250W). We could spend all day pulling examples from the Internet. Pointless.

Just think about it. The GT 630 used a 40nm process, and had 97 CUDA cores. The 1080ti uses a 16nm process, and has 3500 CUDA cores. I'll leave the math to you. It's fair to assume complexity means more than cientific breakthroughs when it comes to size.
 
Well hopefully you learned a lesson pretty cheaply. It could have cost you thousands to make that mistake on the wrong hardware.
I work on 750,000 dollar machines. Imagine the repercussions of doing that there.
 
Ouch!!! I don't mind with 40buck board,but what you said...omg!!!!!

---------- Post added at 04:22 PM ---------- Previous post was at 04:19 PM ----------

I may notbe here much for tne next 24hours. We are under servere tbunder stor s and flood warnings. May loose power that will kill internet. Later wd65
 
Well if you've been paying attention you'll know that's not true. Just for fun try comparing a GTX 1080 to a GT 630.

Components get smaller (transistors, etc.) But as the whole gets more complicated it gets bigger and less efficient, which requires more space and cooling (a 630 doesn't even have a cooler). That's why full towers exist.

I pay attention to each and every detail and those that many would never think of. Full towers don't exist because of a fanless GT-630, each of those have zero to do with each other though they exist in the same technology stream. Full towers are there for space and airflow for larger PCB and sometimes multi CPU-GPU workstations.

A simple GT-630 even fanless can be just as cool in an ITX/MATX situation-case housing as long as the cooling is sufficient. It does not need a huge anvil full ATX case with 10 fans to keep it cool
 
Last edited:
The 280 uses 60% more power (max). Of course it'll run hotter.

The GTX 200 family is also particularly hot. Grab the 550 and put it against a 1080ti. There is a considerable power consumption increase (150W to 250W). We could spend all day pulling examples from the Internet. Pointless.

Just think about it. The GT 630 used a 40nm process, and had 97 CUDA cores. The 1080ti uses a 16nm process, and has 3500 CUDA cores. I'll leave the math to you. It's fair to assume complexity means more than cientific breakthroughs when it comes to size.

On the contrary. As the die shrinks, less power is required and the chip runs colder.
 
A simple GT-630 even fanless can be just as cool in an ITX/MATX situation-case housing as long as the cooling is sufficient. It does not need a huge anvil full ATX case with 10 fans to keep it cool

That's my point.
On the contrary. As the die shrinks, less power is required and the chip runs colder.

That would be true if they had the same amount of components.

For each CUDA core, the power would be less. But there are so many more it doesn't matter.
 
Back
Top Bottom