Date/Time: Wed, 17 Jan 2018 15:04:10 +0000
Post From: Graphics performance regression on Wine1.6 to Wine1.4
|vegasfoster - Posts: 444|
I just set up SC on Wine 1.4 on my second computer, AMD A64 3800+, 2 Gigs of ram, NVIDIA graphics, and 5400rpm hard drive. Not a speed demon by any measure.
The 5 min load average after waiting 6 minutes after each change was as follows:
- Reboot and nothing opened, load average was 0.04 (out of 2).
- Open SC with no live data feed, not doing anything and just sitting there, load average was 0.13.
- Open a single chart with 15 days YM on a 4-2-2 FR chart with half a dozen studies, load average was 0.38.
- Run replay at 200x, single YM chart, load average was .73.
- Load 3 charts tile vertically, all YM charts, doing nothing, load average was 0.57.
- Run replay at 200x, all YM charts, load average was 0.74.
- Load 3 charts tiled vertically for YM, ES, 6E doing nothing, load average was 0.67.
- Run replay at 200x for YM, ES, 6E, load average was 0.89.
So what's interesting to me is that just having charts open seems to suck a lot of juice even though it's just sitting there and supposedly not doing any calculations??? And the more charts the more juice it sucks. But what's really interesting is that running a replay with 3 charts vs 1 chart didn't increase the load at all for just YM and only .05 for YM, ES and 6E (.15 replay minus .10 resting).
When I load the same set up as my trading computer (6 charts), then it does start to feel sluggish, but my trading computer is Core2Duo, 4gigs of ram, and SSD. But a couple charts even on a really slow computer IMO gives totally acceptable results. It should be even better with a native linux version.
The downside is I keep getting an alpha bend error and encountered an improper argument error, where the replay no longer functions and I have to restart SC.
I can install Wine 1.6 later and do the same thing and see how it compares, but I dunno if that's a good way to test it or not. Can I ask how you have it set up (i.e. did you optimize anything?) and how are you testing it?