Skimmed through without sound. I've jumped to the loading part of the video now
and heard it. It's nice to know they are comparing against a PS4 Pro with SSD, it's my main complaint with loading tests these days because it compares the best scenario on PS5 and XSS/XSX (albeit that being the drive they ship with).
So the loading up from start menu to dashboard is:
PS5: 4 seconds
PS4 Pro w/ SSD: 7,8s
PS4 w/ HDD: 9,1s
loading up from dashboard to a race:
PS5: 4 seconds
PS4 Pro w/ SSD: 19s
PS4 w/ HDD: 38,9s
If I understood correctly, base PS4 is HDD which is a shame as PS4 Pro has Sata 3 and PS4 base has Sata 2, which can change the results a bit (300 mb/s vs 600 mb/s throughput), should be comparable enough though because that's actually not the bottleneck usually.
Some of the PS5 advantages over the PS4 are clearly laid out:
- Faster CPU (directly affects loading times)
- Hardware decompression
- Direct access from storage to vram
Of course I know that, but it depends on the game. Usually on base PS4 the improvement is roughly 30% so not lifechanging but nice, sometimes there's no change for one reason or the other and sometimes there's a huge bump, cutting loadings in half. PS5 SSD is always impossible to reach, but sometimes double the PS5 loading times on SSD is not bad at all. Also improves asset streaming a lot, but I don't think GT7 relies heavily on that.
There's also other recurrent bullet point which is similar to decompression but not decompression which is real time asset generation. I wouldn't be surprised if GT7 does the pre-baking for the tracks on the fly at some time during loading.