Nvidia GeForce RTX 4080 Reviews

The scalping and instant sell-outs of these GPUs is just dumb. I'm tired of it. 4090, 4080, can't get because there aren't enough made at the very moment they go on sale and for weeks afterward. CPUs and other related products generally don't have this problem.

Enough of this silliness, I'm buying a Voodoo 2 and calling it a day...
 
There's been GPU since then too with ATI having some DX8 and early DX9 cards with external PSU's.

They're just not needed so never took off.
Huh? ATi's first card with any sort of power plug was Radeon 9700 with standard Floppy-power from computers PSU. VoodooVolts is literally the only external PSU for videocards (servers possibly omitted) ever, and it wasn't an issue ever.
 
The ATI All-In-Wonder Radeon 8500 was released ~10 months prior to the 9700 and had a standard Floppy-power from computers PSU.
Can't see floppy power on either AIW 8500 nor AIW 8500DV?
1668611358251.png
1668611384394.png
But regardless, floppy isn't "external psu" claimed earlier
 
Can't see floppy power on either AIW 8500 nor AIW 8500DV?

Easy to find simply by using Google so I'm not sure how you missed the countless images and websites.




So the 9700 wasn't the first ATI card with an auxiliary internal power connecter as you claimed.

But regardless, floppy isn't "external psu" claimed earlier

The floppy power connector wasn't the external power plug I was talking about, you made an incorrect assumption here.
 
Easy to find simply by using Google so I'm not sure how you missed the countless images and websites.




So the 9700 wasn't the first ATI card with an auxiliary internal power connecter as you claimed.



The floppy power connector wasn't the external power plug I was talking about, you made an incorrect assumption here.
Looks like some AIW 8500's have it and others don't, not sure what's the difference between those models, so you were right there.
However your original argument was that external power supplies have been used with GPUs before, which is false (except for VoodooVolts, but it never came to market and wasn't the issue preventing V56000 from coming to markets, and you specified you didn't mean it)
 
However your original argument was that external power supplies have been used with GPUs before, which is false.

Sigh, you seriously need to brush up on your GPU knowledge.

32.jpg


And unlike the Voodoo card this GPU was actually released to market as I used to own one.
 
Sigh, you seriously need to brush up on your GPU knowledge.

32.jpg


And unlike the Voodoo card this GPU was actually released to market as I used to own one.
I seriously was not aware of this very limited edition dual GPU 7800 GT, which wasn't 'official model' in NV lineup. You hardly can judge something as failure as tech from that, can you?
 
I seriously was not aware of this very limited edition dual GPU 7800 GT, which wasn't 'official model' in NV lineup. You hardly can judge something as failure as tech from that, can you?

This is my last reply on the matter as this is an RTX4080 thread.

This wasn't the only GPU to have the option of a external power brick.
 
Actually it's one of the reasons why I decided not to wait for 4080, when I heard about its 256 bits memory interface.
4080 did use faster memory to mitigate the problem a bit, but that also potentially making it harder to overclock the memory.
On the other hand, 4090 actually have the same memory bandwidth as a 3090 Ti, but it performed quite a bit better. So it's entirely possible that either Ada is more bandwidth efficient or Ampere does not have enough computation power to use all these memory bandwidth.
Don't forget the 40xx have a massive L2 cache compared to 30xx, I guess it's helping a lot with b/w and memory access.
 
Last edited:
Don't forget the 40xx have a massive L2 cache compared to 30xx, I guess it's helping a lot with b/w and memory access.

Yes, actually 4080 is not that bandwidth deficient compared to 4090, but it's more deficient in the computation department.
Comparing various numbers between 4090 and 4080, 4090 has ~60% more pixel rate, and ~70% more "TFLOPS", but only ~40% more memory bandwidth. Also, 4090 only has ~12% more cache so 4080 actually has more cache per bandwidth/computation/fillrate.
So looking from this perspective it's quite as possible what 4080 lacks (relative to 4090) is not bandwidth but computation power.
 
Sigh, you seriously need to brush up on your GPU knowledge.

32.jpg


And unlike the Voodoo card this GPU was actually released to market as I used to own one.
That was the FX 5800 right? I see the AGP "hook" on the back of the connector... Damn that's some old, sad memories right there.
 
That was the FX 5800 right? I see the AGP "hook" on the back of the connector... Damn that's some old, sad memories right there.
Nope, FX 5800 had molex. That's Asus Dual 6800 GT Limited Edition, "custom model" in the sense NV official line up didn't have such things. And it's PCIe, not AGP.
 
Ah yeah, I see my reply wasn't very timely in terms of you already called out it was the dual 6800. Interesting...
 
Ah yeah, I see my reply wasn't very timely in terms of you already called out it was the dual 6800. Interesting...
I'll end the offtopic on my part now too, but looks like the power brick is optional in that, it just leads to 6-pin PCIe-plug, meaning the card would work just fine without it as long as you have 6-pin PCIe from your PSU

--
More on topic, this is either genius or someone getting fired
 
Back
Top