Going beyond the 4 GB of memory
#141
Posted 23 November 2017 - 03:36 AM
I think that for Csantucci FPS isn't improved, because GPU triangles/second is bottleneck here.
#142
Posted 23 November 2017 - 03:47 AM
Carlo, try MG 3.5, it is without any AA. It will better performence for low level HW.
But it has no sense, when almost all of users, I hope, have a better HW then GT120M.
But it has no sense, when almost all of users, I hope, have a better HW then GT120M.
#143
Posted 23 November 2017 - 04:08 AM
#145
Posted 23 November 2017 - 08:03 AM
#146
Posted 23 November 2017 - 08:17 AM
RYZEN 7 1700X (3,4 GHz + XFR)
16GB DDR4 2993MHz very low latency
GTX 1080Ti 11GB
The same scene:
XNA OR - 86 fps
MG3.6 OR - 159 fps
It seems that the more powerful HW, the higher gap in performance have these two platforms.
16GB DDR4 2993MHz very low latency
GTX 1080Ti 11GB
The same scene:
XNA OR - 86 fps
MG3.6 OR - 159 fps
It seems that the more powerful HW, the higher gap in performance have these two platforms.
#147
Posted 23 November 2017 - 09:36 AM
Csantucci, on 23 November 2017 - 07:59 AM, said:
Do you have a 32 or a 64 bit system?
It's a 64-bit system, i7 Quad, W10.
We should consider that some programs are more CPU dependent (XNA) and others are more GPU dependent (MG).
In this case, I think Goku is right about the bottleneck, our weak GPU can not manage MG 3.6 efficiently.
#148
Posted 23 November 2017 - 10:48 AM
I think that what Icik and mbm_OR are saying is true.
In the meantime, starting from the hints of Icik, I found out what makes the difference in FPS between 3.5 versions and 3.6 versions: it's the fact that only with 3.6 multisampling is really used. When disabling multisampling in the last OR MG versions I get the higher FPS I got with MG 3.5. So I have added a video option "Enable multisampling", which is true by default. In older, slower computers this option it may be convenient to disable this option. Picture quality in my opinion is still better than OR plus XNA.
Here http://www.interazio...e/OR3989_MG.zip an OR MG version can be downloaded where such video option is present.
I'm interested in the results that Mauricio gets with this version, disabling multisampling.
I'm quite happy with this: I have again interesting FPS with OR MG, plus the other pluses of MG.
In the meantime, starting from the hints of Icik, I found out what makes the difference in FPS between 3.5 versions and 3.6 versions: it's the fact that only with 3.6 multisampling is really used. When disabling multisampling in the last OR MG versions I get the higher FPS I got with MG 3.5. So I have added a video option "Enable multisampling", which is true by default. In older, slower computers this option it may be convenient to disable this option. Picture quality in my opinion is still better than OR plus XNA.
Here http://www.interazio...e/OR3989_MG.zip an OR MG version can be downloaded where such video option is present.
I'm interested in the results that Mauricio gets with this version, disabling multisampling.
I'm quite happy with this: I have again interesting FPS with OR MG, plus the other pluses of MG.
#149
Posted 23 November 2017 - 11:09 AM