Every generation this happens... PC GPU owners will of course be fine.
Back when the PlayStation 4 launched, it was a 1.84 TFLOPS machine while PCs had 5 TFLOPS GPUs available (780 Ti).
The XSX is going to launch with a 12 TFLOPS APU while the current fastest GPU you can buy today is 13 TFLOPS - for $1200+.
The PlayStation 4 launched with a weak 1.6 GHz netbook-class CPU, which was dwarfed by even older CPUs that had much higher IPC and 4.5–5.0 GHz clocks.
The XSX is going to launch with a modern 3.5 GHz 8c16t Zen 2 CPU, while mainstream PCs today often have… 4.0–4.2 GHz 8c16t Zen 2 CPUs.
It's not exactly the same situation.
Yes, there are going to be differences. The TDP is likely tuned lower than you'd get with an equally-clocked Ryzen CPU, and there are probably other cost-saving measures like using less cache.
TFLOPS are a bad way to measure gaming performance, with NVIDIA having higher game performance than their raw compute performance implies when compared against AMD.
We have no idea what AMD's RT implementation is going to be like compared to NVIDIA's, and it will probably be going up against NVIDIA's second-generation RT implementation by the time the consoles launch.
But the pace of hardware improvements has slowed considerably, and the gap in performance seems far closer than it's ever been.
By the time the consoles launch, we'll probably have Zen 3 CPUs, and a new series of GPUs from NVIDIA. With the shrink to 7nm, those new GPUs will hopefully have a big leap in performance compared to what is currently available.
So the sky is not falling, but I think a lot of people with current PC hardware, expecting it to last for much of next generation, are going to be disappointed.
Even though you cannot compare AMD TFLOPS with NVIDIA TFLOPS, my GTX 1070 with its 6.5 TFLOPS is going to fall quite short of next-gen with 12 TFLOPS and RT hardware. At best, my 4.0 GHz Ryzen 1700X is probably going to match next-gen CPUs.
I'm not expecting there to be PC hardware—except maybe at the very highest-end of what is available by launch—which will be capable of running games built for 30 FPS on next-gen hardware at ≥60 FPS with equivalent or better settings. Maybe you'll be fine with cross-gen games, but not "next-gen" games that are really pushing the hardware.
Storage could get pretty interesting once Sony announce details of the PS5 hardware too.
For me, it's not a question of whether the consoles are going to be faster than my existing PC or not—I'm sure they will be.
It's whether it will be worth it for me to spend what it takes to build a better PC than these consoles.
There is a lot that I value about the experience of gaming on PC, but I spent >$1000 on hardware upgrades this generation only to be left disappointed with how poorly many games ran.
Last generation, a mid-gen upgrade of similar cost blew consoles out of the water and would run games stutter-free at 60 FPS or higher with ease. This generation the games run at higher frame rates but rarely smoothly.
I went into more detail about it with posts in this topic recently.
i really don't see how Microsoft and Sony aren't going to be taking a sizable loss on each console even at $499, regardless how good their relationship with AMD is
It's because AMD and NVIDIA realized they could screw over PC gamers once NVIDIA released the Titan at $999 and it wasn't a complete failure, and people kept buying hardware even though prices were inflated due to GPU mining.
What Sony and Microsoft are paying will be far closer to the "real" cost of that hardware.
Yeah I notice 0 difference in gaming on my Sata SSD limited to 500 MB/s versus my NVMe drive that has 3200 MB/s read speed.
That's because game loading isn't optimized for it.
To quote an older post of mine:
Without making things too complicated, let's just say that there are only two things that have to happen when loading a game: getting data off the drive, and decompressing it into memory.
With an HDD being slow, the time it takes to decompress the data into memory might only be 5% of the loading time - so a developer may not see the point in optimizing it, and it only runs on a single CPU thread because that is easier to build.
With a SATA SSD the data access is significantly faster, so it speeds things up a lot - to the point that most of the load time is now waiting on decompression to happen.
With an NVMe SSD the data might be loaded in 10x faster than the SATA SSD, but because most of the loading time is still spent waiting for decompression, the difference is negligible.
With a game optimized for loading off an SSD, the developer now focuses on speeding up the data decompression. Instead of only using one CPU core they build a loading system which scales to as many CPU cores as the system has.
Now that the loading uses all 8 cores/16 threads on a Ryzen CPU (and more in higher-end PCs) instead of only one of them, loading is significantly faster - and we start to see that difference between SATA and NVMe emerge.
I mean, even an RTX 2080 can't run RDR2 today at 4K/60...
It can. Just not on "ultra" settings.
People
really need to stop judging performance by how games run when maxed-out.
