Friday, December 9th 2022
NVIDIA GeForce RTX 4070 and RTX 4070 Ti Detailed Specs Sheet Leaks
It turns out that NVIDIA has not one, but two new GeForce RTX 40-series "Ada" SKUs on the anvil this January. One of these is the RTX 4070 Ti, which we know well to be a rebranding of the RTX 4080 12 GB in the face of backlash that forced NVIDIA to "unlaunch" it. The other as it turns out, is the RTX 4070, with an interesting set of specifications. Based on the same 4 nm AD104 silicon as the RTX 4070 Ti, the new RTX 4070 is significantly cut down. NVIDIA enabled 46 out of 60 streaming multiprocessors (SM) physically present on the silicon, which yield 5,888 CUDA cores—the same count as the previous-gen RTX 3070, when compared to the 7,680 that the maxed-out RTX 4070 Ti enjoys.
The GeForce RTX 4070, besides 5,888 CUDA cores, gets 46 RT cores, 184 Tensor cores, 184 TMUs, and a reduced ROP count of 64, compared to 80 of the RTX 4070 Ti. The memory configuration remains the same as the RTX 4070 Ti, with 12 GB of 21 Gbps GDDR6X memory across the chip's 192-bit memory interface, working out to 504 GB/s of memory bandwidth. An interesting aspect of this SKU is its board power, rated at 250 W, compared to the 285 W of the RTX 4070 Ti, and the 220 W of its 8 nm predecessor, the RTX 3070.
Source:
harukaze5719 (Twitter)
The GeForce RTX 4070, besides 5,888 CUDA cores, gets 46 RT cores, 184 Tensor cores, 184 TMUs, and a reduced ROP count of 64, compared to 80 of the RTX 4070 Ti. The memory configuration remains the same as the RTX 4070 Ti, with 12 GB of 21 Gbps GDDR6X memory across the chip's 192-bit memory interface, working out to 504 GB/s of memory bandwidth. An interesting aspect of this SKU is its board power, rated at 250 W, compared to the 285 W of the RTX 4070 Ti, and the 220 W of its 8 nm predecessor, the RTX 3070.
93 Comments on NVIDIA GeForce RTX 4070 and RTX 4070 Ti Detailed Specs Sheet Leaks
How in the hell is it possible that a simple case fan that doesn't earrape you costs 40 bucks? A stupid peace of plastic for 40 bucks. And there is enormous competition in the fans market, and they still treat us like fools. So what's so surprising about GPU giants shenaningans, where there is almost non-existant competition.
Only two things can mitigate the problem:
Extreme competition to break down prices (which is unlikely because it's a niche market and very few entities can enter the market)
Reaching a point where it's not necessary to upgrade hardware anymore, where you can run a card for 15 years so a one time investment makes the prices look much better
TL DR-your expectations weren't wrong, but you did buy the wrong card to do it with. When you're looking to last 4+ years with a GPU, you want VRAM headroom, and enough bandwidth. The exact thing Nvidia is cutting down since Turing.
That's in a nutshell what has been happening across the Nvidia stack at large. Today, we see a 4090 that on release struggles already on two notable games; Cyberpunk at full tilt and Portal, the latter being a horribly simplified full path traced application, after all, its geometry and texture simplicity offers major optimization chances. We're like those donkeys chasing the carrot. We'll never eat it, but we'll sniff it from time to time if only we keep running to the latest greatest. I've never been a donkey like that, it just doesn't feel right to me. I feel like being taken for a ride.
Innovation, progress, I guess so, all I see is pretty limited progress for three generations worth of innovation... for an extreme performance cost. Even with DLSS3 on the FPS gets a factor 4-5 worse in Portal. Without DLSS3, its a factor 20 worse or even more. Is it a better game for it? I'm really not seeing it tbh...
Its worth questioning this push.
I thought the XTX targets 4090, not the pitiful and lower grade 4080.
Navi 21 was relatively slower at 2160p exactly because of the limited memory throughput which the Infinity Cache of only 128 MB couldn't compensate. Not really. They will be very fast even for 2160p because you can set the in-game settings to medium, high, very-high and ultra-high depending on the demands of the particular game engine.
Hell, you will be even able to play CS:GO at 4320p, no worries :D
You vote with your wallet - could buy a faster $0.9k card from AMD - 7900 XT 20 GB. You also get 4 GB VRAM more.
I'm glad I bought a 3070 now....
I'm not saying you're wrong, and if I were a betting man, I'd bet you'll be right - but it's way too early to know at the moment.
Legit benchmarks is the only way to judge and eventually the perf/$-€ and/or from features every individual user is interested in.
Bit bus is just a single number on a equation. As speeds, cores/stream processors/MBs count, and all other bits and pieces of modern PCs are.
Wait until all of them are put to real usage together and then make up your minds.
Its only logical and rational.