Digital Foundry Article Technical Discussion [2021]
- Thread starter BRiT
- Start date
- Status
Does Tom even know the specs of those machines? Why is he even surprised the PS5 is having such an edge in transparency heavy scenes? We know PS5 has about 18% faster pixel throughput with color ROPs and about 130% advantage with depth/stencil ROPs against the heavily downgraded ROPs inside XSX. As I understand it (I could be wrong) those depth/stencil ROPs could be very uselful in scenes just like this one with a dense forest.
While we have seen XSX having logically an edge in many compute limited games, we also have seen PS5 having logically an edge in those ROPs bound scenes since one of the first game tested (AC Valhalla), up to the very pixel limited The Touryst with a 77% resolution advantage on PS5. Another game where the XSX had unexpected and big drops was in Hitman 3 in a very similar scene with plenty of vegetation.
While we have seen XSX having logically an edge in many compute limited games, we also have seen PS5 having logically an edge in those ROPs bound scenes since one of the first game tested (AC Valhalla), up to the very pixel limited The Touryst with a 77% resolution advantage on PS5. Another game where the XSX had unexpected and big drops was in Hitman 3 in a very similar scene with plenty of vegetation.
Last edited:
Forgive me playing catch up - what's up with the ROPs in XB?
Lets see if I can recall correctly. They use RDNA2 rops just like the desktop GPUs, but are cut in half and then double-pumped. This was for some silicon savings.
Some folks like to color all their posts with hyperbole adjectives to describe technology from a company that is not the one they worship. So take all their posts on the matter with the appropriate levels of salt.
snc
Veteran
"against the heavily downgraded ROPs inside XSX" - I mean isnt this a little exageration ?:d we have rdna 1 vs rdna2 tests on computerbase and there wasn't any performance differencesDoes Tom even know the specs of those machines? Why is he even surprised the PS5 is having such an edge in transparency heavy scenes? We know PS5 has about 18% faster pixel throughput with color ROPs and about 130% advantage with depth/stencil ROPs against the heavily downgraded ROPs inside XSX. As I understand it (I could be wrong) those depth/stencil ROPs could be very uselful in scenes just like this one with a dense forest.
While we have seen XSX having logically an edge in many compute limited games, we also have seen PS5 having logically an edge in those ROPs bound scenes since one of the first game tested (AC Valhalla), up to the very pixel limited The Touryst with a 77% resolution advantage on PS5. Another games where the XSX had unexpected and big drops was in Hitman 3 in a very similar scene with plenty of vegetation.
https://www.computerbase.de/2021-03/amd-radeon-rdna2-rdna-gcn-ipc-cu-vergleich/
tough its possible there are some situations that deifference could be found edit: there are some differences but small and change from title to title
RDNA2 ROPs (Xbox Series) have as much as color ROPs but twice less depth/stencil ROPs as RDNA1 ROPs (PS5). PS5 also has 18% higher clocks on top of that. On PS5 the ROPs take almost twice the die area as the XSX ROPs (but there is a bit more of them for yield on PS5).
XSX GPU is not a fully RDNA2 GPU. For instance there is no infinite cache on Xbox Series. So the comparisons can't be applied for XSX.
Last edited:
snc
Veteran
still if this changed rops would be such a big deal probably we would notice bigger differences rdna 1 vs rdna 2
davis.anthony
Veteran
Thank you!! It's the one thing about DF that really pisses me off, every time PS5 is faster in a game or in a specific scene they say that they don't understand why, for people who are supposed to know about GPU's, game engines and GPU pipelines they sure don't understand a lot.
davis.anthony
Veteran
snc
Veteran
ps5 has rops from rdna1 xsx from rdna2, imo valid comparison but we can agree to disagree
The ROPs in XSX are RDNA2 RB+. Meaning they are designed to do an effective 2X push with color calculations. So effectively you could get an output of 16 RBEs with just 8. But depth calculations it does only 1X.
PS5 shipped with 16 RBEs, XSX shipped with 8 relying on double pumped nature of the RBEs; but that doesn’t apply to everything. As per the tweets above, silicon savings were large.
Last edited:
Riddlewire
Regular
A game in which XSX is pushing 44% more pixels.
Also, DF's test for Hitman 3 was back in January. That game has gotten patches all year that no one has bothered testing. They added a Performance Mode option to Xbox One S/X back in March which I didn't see anyone bothering to test, The point being that IOI has continued improving the game throughout 2021.
I'm not saying XSX doesn't have a disadvantage when it comes to alpha, just that using the results of old tests for one game to make definitive statements about expected performance in a different game is stupid.
davis.anthony
Veteran
davis.anthony
Veteran
And PS5 could have a 44%+ higher frame rate, it was very strange PS5 running at the resolution it did in Hitman as it clearly has the head room to go higher.
Silent_Buddha
Legend
And your evidence that it could without potentially suffering numerous dips below the target framerate if resolution were increased or framerate were unlocked is...?
IOI just arbitrarily decided that it should just sacrifice up to 44% more performance for laughs?
Regards,
SB
Riddlewire
Regular
Absolutely it could. The point I was making was that unless FROM Software has created Elden Ring using IO Interactive's Glacier Engine, there's no reason to use results from Hitman 3 to form opinions about the expected performance of the consoles in Elden Ring. Doubly so, since those Hitman 3 measurements are potentially outdated by now.
davis.anthony
Veteran
And your evidence that it could without potentially suffering numerous dips below the target framerate if resolution were increased or framerate were unlocked is...?
IOI just arbitrarily decided that it should just sacrifice up to 44% more performance for laughs?
Regards,
SB
1. A difference of 44% is higher then any paper difference between the two machines (Largest is memory bandwidth at 25%)
2. PS5 doesn't have a history filled full of suffering massively when running at the same resolution as XSX
3. IOI were happy enough to let the XSX dip in to the low 40's in places but not PS5?
Last edited:
- Status
Similar threads
