Tweet
I decided it best to create a new thread on this rather than leave it buried in some comments in the Intro to PG thread.
I have a GTX 970 video card and an I5-6400 cpu which runs TGC great but when running Augusta on PG I'm getting quitee low frame rates. On PG default settings at 1980x1080 full screen, I was getting a lot in the high 20's. Average for the whole course was probably in the low 30s. The 13 tee was 19 fps!
When I examined the GPU and CPU loading on the 13th tee I was surprised to see my 970 was only running about 35% loaded. That typically means the system is CPU limited. Using CPUID HWMonitor my CPU total load was around 50% which didn't seem too bad but I noticed two cores were about 75% and 79% respectively...the others were 50-60%. Values that high can easily be CPU limited (the measurements are software estimates and don't necessary reflect actual saturation limits within the CPU hardware).
At the least it is clear that PG uses the CPU more than TGC does. In fact it appears my system may be CPU limited (not video card limited) which indicates that, while an I5 is fine for TGC, it may not good enough for PG.
I would like to have other people with the 970 video card do this test. Both with I5 and I7s to see if they get different frame rates with the different CPUs (but same 970 video card).
The quick test would be setup the same resolution/quality settings then jump to the 13th tee and record the frame rate and GPU loading.
The exact duplicate of my test (which would give us the best data) is:
1) Resolution = 1920x1080, Full Screen, Quality settings = Default (ie Ultra High, VSync is off)
2) Two screens with CPUID HWMonitor simultaneously running on the extended display, PG running on the main screen (If only one screen then PG gets minmized to read the CPU/GPU data which distorts the rresults).
3) Load Augusta in Practice mode, jump to 13th Tee and measure fps. Mine is 19 fps
NOTE: Once I have everything setup and running I clear the Min/Max values in HWMonitor and then run for a minute or so and record the maximum values for the four cores and gpu.
Either way, tell us the following:
- frame rate on 13th tee
- gpu and cpu loading
- Processor type (ie I5 or I7 and the specific version if you know it)
- how you did the test
Hopefully we can get a better handle on how Perfect Golf works and what types of systems we need to get good results. It may be that a lesser video card with a better cpu is the way to go for PG. We've gotten used to TGC which is different (In TGC my GPU runs at 99% on all courses while my I5 run around 60% so it is completely video card limited even with an I5).
By the way, this isn't about the Augusta course. It is still beta and probably will get tweaked a bit. It is just a good test to investigate and understand Perfect Golf behaviour.
I have a GTX 970 video card and an I5-6400 cpu which runs TGC great but when running Augusta on PG I'm getting quitee low frame rates. On PG default settings at 1980x1080 full screen, I was getting a lot in the high 20's. Average for the whole course was probably in the low 30s. The 13 tee was 19 fps!
When I examined the GPU and CPU loading on the 13th tee I was surprised to see my 970 was only running about 35% loaded. That typically means the system is CPU limited. Using CPUID HWMonitor my CPU total load was around 50% which didn't seem too bad but I noticed two cores were about 75% and 79% respectively...the others were 50-60%. Values that high can easily be CPU limited (the measurements are software estimates and don't necessary reflect actual saturation limits within the CPU hardware).
At the least it is clear that PG uses the CPU more than TGC does. In fact it appears my system may be CPU limited (not video card limited) which indicates that, while an I5 is fine for TGC, it may not good enough for PG.
I would like to have other people with the 970 video card do this test. Both with I5 and I7s to see if they get different frame rates with the different CPUs (but same 970 video card).
The quick test would be setup the same resolution/quality settings then jump to the 13th tee and record the frame rate and GPU loading.
The exact duplicate of my test (which would give us the best data) is:
1) Resolution = 1920x1080, Full Screen, Quality settings = Default (ie Ultra High, VSync is off)
2) Two screens with CPUID HWMonitor simultaneously running on the extended display, PG running on the main screen (If only one screen then PG gets minmized to read the CPU/GPU data which distorts the rresults).
3) Load Augusta in Practice mode, jump to 13th Tee and measure fps. Mine is 19 fps
NOTE: Once I have everything setup and running I clear the Min/Max values in HWMonitor and then run for a minute or so and record the maximum values for the four cores and gpu.
Either way, tell us the following:
- frame rate on 13th tee
- gpu and cpu loading
- Processor type (ie I5 or I7 and the specific version if you know it)
- how you did the test
Hopefully we can get a better handle on how Perfect Golf works and what types of systems we need to get good results. It may be that a lesser video card with a better cpu is the way to go for PG. We've gotten used to TGC which is different (In TGC my GPU runs at 99% on all courses while my I5 run around 60% so it is completely video card limited even with an I5).
By the way, this isn't about the Augusta course. It is still beta and probably will get tweaked a bit. It is just a good test to investigate and understand Perfect Golf behaviour.
Comment