Unreal Engine 5, [UE5 Developer Availability 2022-04-05]

pretty harsh performance on a 1070 -- building the project now. lol almost threw up on drone mode, controls exactly like IRL drone, was thinking it was going to respond like a standard ghost camera in a game
 

E2VUHXCXsAUdHnD
 

So this is how it's handled in the editor if you have a slow HDD like I do. The editor can't instant flash, so it's just going through loading everything into VRAM before continuing.

I had a couple stutters like this as well when first loading, but once it's in memory, stuttering will go away generally. Performance is still bad on a 1070 however. By bad, as in, I don't like it=P Probably less than 30 it feels like, but I could be wrong. IT's also very blurry in motion, the temporal is pretty intense.

A lot of etc, effects load JIT the first time they appear.
I'm setup is struggling here. I don't know if you can get away with a slow HDD and play through fully without seeing a hiccup
 
Last edited:
https://www.resetera.com/threads/unreal-engine-5-dev-workflow-stream-today-5-26.431627/post-65799137

gofreak reduce RAM available and this is how Nanite looks with a low speed HDD(125 MB/s)

Well... I do wonder if the windows filecache is having an effect here. If there's enough RAM available and the files had been loaded due to a previous run, or editor run, or from a write-cache during installation, you might not be seeing what really happens when data has to come in off the hard drive.

I was seeing the same behaviour on low IO drives, but, when I reduced system RAM to <8GB free (should be no problem if the footprint is 3-4GB), and again used my lowest speed drive (125MB/s), this was the result after a bit of traversal:

wRMF4Ce.png
 
I was arguing just a few pages back in this thread (and for months before) that additional system RAM could be used to mitigate the need for an ultra fast IO in many scenarios and this seems to prove that point (with which not everyone agreed). Its not ideal obviously from a cost, or initial load point of view, but its clearly a viable solution for those without fast SSD's or until DirectStorage is available. DDR5 hitting the market later this year with double the density of DDR4 should make these RAM capacities much more commonplace moving forwards too.

These specs scream lack of Direct Storage, high RAM to mitigate the slower IO and additional CPU cores to mitigate the lack of a hardware or GPU based decompressor.

I'd love to hear what kind of SSD utilisation the demo uses both on initial load and then once the RAM is fully loaded up. My assumption would be that on high RAM systems, once the RAM is filled the IO requirements are pretty reasonable. But the less RAM you have, the higher they get.
 
I'm trying to download the sample but I have "no compatible engine installed" despite having just installed UE5ea..

Any ideas?

edit:
Nevermind I restarted and it worked lol
 
Last edited:
I was arguing just a few pages back in this thread (and for months before) that additional system RAM could be used to mitigate the need for an ultra fast IO in many scenarios and this seems to prove that point (with which not everyone agreed). Its not ideal obviously from a cost, or initial load point of view, but its clearly a viable solution for those without fast SSD's or until DirectStorage is available. DDR5 hitting the market later this year with double the density of DDR4 should make these RAM capacities much more commonplace moving forwards too.

These specs scream lack of Direct Storage, high RAM to mitigate the slower IO and additional CPU cores to mitigate the lack of a hardware or GPU based decompressor.

I'd love to hear what kind of SSD utilisation the demo uses both on initial load and then once the RAM is fully loaded up. My assumption would be that on high RAM systems, once the RAM is filled the IO requirements are pretty reasonable. But the less RAM you have, the higher they get.

Well, im having no problems with a samsung 970, which is half the speed of whats available today (3400mb/s vs 7000mb/s). Seems direct storage isnt even a need since im enjoying true next gen graphics as in PS5 tech demo to its fullest now. Obviously, DS will ease things up and improve things even further. My CPU isnt even taxed so much by this ue5 demo, maybe due to fast nvme storage.

My system has 32gb of main ram, im too lazy to try with 16 (its almost midnight here) myself but found this video on someone with a nvme/16gb main ram system (and a i5 cpu/RTX2060 lol).

Another thing, older GPUs struggle more (pascal), proof next gen graphics still require compute and processing power.

 
I sure did, it is now corrected, thanks.

That was your video? Great comparison! They are strangely close together given the speed difference (125MB/s vs 7000MB/s). How much system RAM do you have and did you happen to note how much of it was in use when the transition began? I suspect large amounts of the required data were already cached thus reducing dependency on raw IO speed. If this was pulling the dark world entirely from disk at the point of transition I'd expect the difference to be much greater.
 
Can I throw a random request into the ether for someone to play around with Series S levels of base resolution upscaled to 1080p. Really curious as to how that looks.
 
I was comparing Hardware-RT vs Software RT with Lumen and this is just ridiculous.

OlHPTYv.jpg


SjYRVhW.jpg


Sorry for the different perspectives but it is kind of obvious to see that there are minimal improvements to image quality with Hardware-RT if any.

Yet, using Hardware RT completely destroys performance. From 34 FPS (top pic) to 5 FPS (bottom pic)

Either this is a bug or Epic single handedly made Triangle Based Raytracing and dedicated RT acceleration obsolete. Because Lumen looks damn good.

Maybe if you really search for differences and your name is Alex Battaglia, Hardware RT is still better in some scenarios, but seriously, if it tanks performance that much for no visual boost the average Joe can notice, Hardware-RT is completely useless when coupled with Lumen in UE5.

Really hope this is a bug and RT cores/ Ray accelerators will still have an important role to play in next gen titles.
 
Would it be something like 480p scaled to 1080p or is that taking it too far?

Consoles were said to target native 1080p and then use temporal super resolution(tsr) to scale to 4k. This is something very similar dlss tries to do. Will be interesting to compare dlss versus tsr in ue5.
 
Back
Top