edit: I came back to this comment a year after it was posted because of a reply on it and found it almost has 2k likes... wow, that's the most of seen on any post anywhere
oliverthecat
2:08 look at this Vulkan steps :D Udp: are u crazy guys?) 1.3K likes, ty
Xth
3:22
the horse in d12 : "guess I'll die now"
wtfdarus
first 10 seconds: I imagined people tapping on tables waiting for food xd
anonyme
2:10 The run animation on the Vulkan API looked silly LMFAO
evergaolbird
I have seen that cutscene so many times playing with my settings. It’s weird because it seems to be randomized. Like when he shoots the cops and sometimes he shoots the second cart while it approaches and other times it’s stopped. One time one of the horses survived the dynamite and ran in front of the cart as he rides away causing him to hit a tree. It’s hard to get a very accurate comparison because the randomization causes slight changes to the FPS average.
CouchMan
Let me just say that your channel is one of the best on youtube. No BS, just pure benchmarking, it is brilliant. Never stop.
Paul
Guys with these explicit APIs like DX12, Vulkan and Metal, CPU usage and RAM usage could vary because of how they are implemented in the engine itself. These APIs lend themselves much better to multithreading, custom memory allocation, etc. Saying Vulkan used more RAM or DX12 uses more CPU doesn't mean anything since we're looking simply at how Rockstar used the API in their game engine. Another developer might do something different in the way they implement their rendering system. Given just how much more work is put on the engine devs, I would take any of these numbers with a grain of salt when benchmarking the API itself. Remember that it's not JUST the API that is changing between these benchmarks, there are subsystems that operate those APIs as well and handles all memory operations with regard to the GPU, as well as use many CPU cycles preparing data for the GPU. Those subsystems will change depending on the API, but are still part of the game itself, not the API (driver). This is grossly simplified obviously.
NickEnchev
Vulkan has more ram usage, dx12 more cpu usage
valerio
Vulkan shpuld be default option. When dx12 works only on win10, Vulkan supports all of platforms and works on older version of windows like win 7 and etc
kaciorx
DX12 has a straighter, more consistent line in the fps/frame timing, while Vulkan saw some better performance in some scenes. The real question is: which won't completely mess up my ability to even play the damn game?
lescarpio
The framerate is much more stable with directx 12 but a bit lower, depends on which just feels better to you
Mart-E
2:40 Wow, Vulkan look so much more smooth
clx
0:00 - 0:10 me hitting the table waiting for my mom to finish dinner.
pruna
Pretty comparable. Looks at least for nvidia you get a stability edge in DX12. That said, it shows a good implementation in either gives almost identical results. Which is fucking great news for Linux and just the industry in general.
Lets hope we see future games putting these 2 engines as their primary focus moving fowards.
jackpowell
More FPS with Vulkan, But smoother with DirectX12
Try 5700 XT with Vulkan.
rlifeh
2:10 >>> Vulcan look at screen
also nice FPS difference at 2:25
Gomcio
I was dealing with some major pop in on Vulkan and then tried out DX12 and that basically solved the issue.
omerali
02:10 ROFL ... that Vulkan run should have been added to the game trailer
aelaraji
Either something is wrong with my eyes, or the Vulcan has a smoother picture, but according to monitoring - 1% and 0.1% low - lower than DX12. At the same time, DX12 feels stuttering.
Judging by the graph what walking on the Internet (I do not know how much it correct) - RTX 2060 > GTX1080 in this game (without rtx of course). Freaking Nvidia.
