08-05-2012, 01:04 PM (This post was last modified: 08-05-2012, 01:06 PM by avih.)
(08-05-2012, 12:46 PM)miseru99 Wrote: As much as I would like it to be true(who wouldn't?;D), I guess it counts too many zeros:
(used RC7 without any additional commands, NTSC game, same problem with 1.0.0)
This is obviously an issue. It seems it either failed to read the FPS, or your GSdx window showed temporary (even for a split second) this very high fps (6433), and since this tool counts in "PS2 Time", it calculated that in 6000 fps (that's 10x speed) - 15s in "ps2 time" passed after a split second (15/100 = 150ms).
Does it happen always? or just a one-time glitch? and if it's reproducible, can you tell if your GSdx window shows such very high FPS (even very shortly)?
Hmm.. no, I don't think anything's changed with pcsx2. It's just that on my system the decimal symbol is '.' (dot), while on your system it's a comma. It appears that my code interprets 65,44 as 6544 fps instead of just less than 65 and a half. I'm quite sure it uses the systems locale to interpret the numeric readings, but I'll have to look at it again.
In the meantime, if you'll change your system display to use '.' as a decimal point, I'm quite sure it'll work. Cute bug. Thanks
08-05-2012, 03:31 PM (This post was last modified: 08-05-2012, 03:34 PM by omnikam.)
Avih"Bottom line is, that you can modify pcsx2's settings (e.g. speed hacks, EE/VU advanced recompiler options, GSdx resolution, etc) and use this tool to give rating of how good the settings are in 'real-world play'. This takes into account not only the average speed, but also how bad it does when it can't reach full speed. So you want to find settings which gives you the highest rating on your system at this specific measured section."
I like your program Avih, im sure persons with slower Pc`s will use this to squeeze every fps from hacks. But that also brings me to the question, When messing with Vu cycle stealing the FPS displayed are inaccurate, so wont this make the results from this program false?
Msi GF-62VR
Intel core i7 7700hq @2.8(3.5ghz turbo) Nvidia 6gig GTX1060 16 gigs DDr5 windows 10
*base 64 images don't work here - ref*
08-05-2012, 03:50 PM (This post was last modified: 08-05-2012, 03:59 PM by avih.)
(08-05-2012, 03:31 PM)omnikam Wrote: ...
But that also brings me to the question, When messing with Vu cycle stealing the FPS displayed are inaccurate, so wont this make the results from this program false?
Yes, it will show false results if your settings cause false FPS. I was sure I mentioned it before, but can't seem to find it right now...
- edit
Ahh.. found it:
(11-15-2010, 04:37 PM)avih Wrote: ...
The final result, i.e. 'pcsx2bench rating' at the bottom is a metric I invented that tries to measure how good the game can be played in practice (speed-wise, regardless of visual glitches or inaccurate FPS due to excessive VU steal, etc).
...
Maybe not too clear, but that's what I meant, that false fps will cause false readings/results.
Also, it can't detect if there are visual glitches due to bad settings (like rounding modes), etc.
It just measures the FPS as displayed at the window title bar, and then provides overall average speed, "usable speed rating" and few more statistics.
08-05-2012, 05:56 PM (This post was last modified: 08-05-2012, 05:57 PM by avih.)
Updated 2012-08-05:
---------------------
pcsx2bench 0.1RC9
- Fix: Should now compensate for different numeric formats by locales (e.g. "FPS: 60,00" vs "FPS: 60.00").
- NOTE: _Might_ break if FPS goes beyond 1000. If that happens, just limit the FPS.
08-06-2012, 06:46 AM (This post was last modified: 08-06-2012, 06:47 AM by soratidus999.)
"NOTE: _Might_ break if FPS goes beyond 1000. If that happens, just limit the FPS"- bahahahahah!
also, how hard would it be to use the hook that prints FPS to the title to also print into a text document every .5 second, or even feed the info directly into the benchmarker when a flag is set?
just seems to me it may have a more accurate reading
EDIT: also thanks for the benchmarker - its quite useful XD
(08-06-2012, 06:46 AM)soratidus999 Wrote: ...
also, how hard would it be to use the hook that prints FPS to the title to also print into a text document every .5 second, or even feed the info directly into the benchmarker when a flag is set?
just seems to me it may have a more accurate reading
...
On a scale of 1-10: not hard, but not sure it's worth it either. As I mentioned, IMO it's surprisingly usable, accurate and consistent even with this amateur-ish reading process.
Also, this tool didn't attract much attention, judging by the number of comments by other users, which reinforces the conclusion that such improvement is not worth the time spent on it.
And finally, I wrote this tool before I ever looked at pcsx2's source code. I just threw it together quickly as a stand-alone thingy, and then it just worked, so I didn't feel an urge to fix it...
i was looking into it and i found FpsManager.GetFramerate but it doesnt get referred to enough, so i don't think thats the function i want
the pcsx2 code is a bit too complicated/huge for an uninitiated C++'er like me, i understand it, but i spent my time in HTML and PHP instead... and even that was a long time ago
if i can wrap my head around visual C++ 2010 i might give it a shot further down the line...
