It was 5 cents cheaper per $2500 GPU not to?
It was 5 cents cheaper per $2500 GPU not to?They could've just opted for 48V rails but noooo.
More actually, 48V pdn ain't cheap. Still my point stands.It was 5 cents cheaper per $2500 GPU not to?
loud incorrect buzzer noiseThese guys seem to be on the right track:
It's the end user problem, like fuel consumption of V12. As long as people want 'em and pay for 'em it's ok...Honestly, the fact that we have >500W GPUs in desktops is the main problem.
That analogy works, but it’s more like if the V12 supercar had a special high-flow fuel nozzle that sometimes would overflow the tank and spill out if it was not properly inserted into the fuel tank, and the manufacturer of that car then decided to use that same nozzle for the rest of their lineup, including more economical cars which don’t need a high-flow nozzle to begin with. Nor did any other car that came before it.It's the end user problem, like fuel consumption of V12. As long as people want 'em and pay for 'em it's ok...
Now what's bad and should not be happening is allowing wattage go 50% up for a minor 10%-ish overlock, that stuff should stop but I can't see it happening without legislation, which will never happen in the US.
If it was 600w but with optimal clocking, then it would be fine to get more perf - only minor problem in this otherwise fine plan: it will require wider bigger chips, so more expensive to produce in the first place.
Supercars especially, got a lot of special very expensive non-standard stuff that can go wrong if not "properly inserted"if the V12 supercar had a special high-flow fuel nozzle that sometimes would overflow the tank and spill out if it was not properly inserted
No, it's just a segmentation thing. Halo is halo and is basically its own thing.I'm guessing they have a glut of unsold 4090s. Hence the gaping VRAM hole between 5080 16GB and 5090 32GB. Or they plan to release a 5080 Super 24GB later.
They do not. Go check the market in the US. Still $2000 for a 2 year+ old card about to be replaced. Supply has to be low.I'm guessing they have a glut of unsold 4090s. Hence the gaping VRAM hole between 5080 16GB and 5090 32GB. Or they plan to release a 5080 Super 24GB later.
That's the retailers themselves holding onto stock in their warehouses. They have gotten clever. Remember how Newegg was selling some brand new RX 5600s just a year ago? There are retailers in UAE holding onto even RX 6800 XT, 6900 XT and 6700 XT stock and trying to sell them at launch prices instead of facing reality. And 3090 Ti selling for over $1000.Still $2000 for a 2 year+ old card about to be replaced.
Everybody knows that 48V DC would be better for both GPU and CPU VRMs. There is just a serious transition problem here. NV could introduce the 12-pin connector by putting an adapter that eats 8-pin connectors in the box. You can't do the same for 48V, everyone would have to switch PSUs and GPUs at the same time.it's not fine PCI-SIG literally had to amend the POS.
But "we're gonna pump a zillion amps over tiny ass pins on a microscopic connector" is truly one for the road.
They could've just opted for 48V rails but noooo.
These guys seem to be on the right track:
Everybody knows that 48V DC would be better for both GPU and CPU VRMs.
Perhaps Nvidia could bundle new PSU with 2 grand cards? If that makes them perform better (even if it's just lower power) than what's the problem? Get top models on board of thise - whoever is buying 5090 will be ok with it, as long as it brings real benefits thoughYou can't do the same for 48V, everyone would have to switch PSUs and GPUs at the same time.
at least client Blackwell has boxes leaked. not even that for RDNA4.Less than 24 hours left till CES and still ZERO leaks on 5080/5090 perf???
That's because it is expected to sell more than 1at least client Blackwell has boxes leaked. not even that for RDNA4.
not a box! just a teaser.Powercolor!
AMD doe the exact same thing, except worse (signed PSP blobs).I'm not surprised at all that there are no leaks; it was enough for me to read on Guru3D years ago about how Nvidia treats its AIB partners.
Guru3D: "I'll let you in on that secret. The AIB partners have all been prepping their cards for months now. They have the products, engineering boards for a while. NVIDIA however, has not released a driver that works with anything other than the test software they supply. So get this, I am writing this article on September 1st, hours before the presentation, and still, the board partners have no idea what the performance is going to be like. We need to advance on that as the board partners even do not know the thermal capacity effect of their products. NVIDIA has provided them with test software that will work with the driver. Basically, these are DOS-like applications that run stress tests. No output is given other than PASS or FAIL. We know the names of these test applications: NVfulcrum test and NVUberstress test. For thermals, there is another unnamed stress test, but here again, the board partners can only see PASS or FAIL. Well, we assume they have tested with thermal probes. What this paragraph, well, to show you the secrecy that NVIDIA applied for this Ampere project."