R37 on OS X is a step backwards in both performance and stability

The idea of this release, IIUC, was to sacrifice some stability (that certainly happened) temporarily in the interests of improving performance. From my POV, running on OS X, the latter has not been delivered. Instead of a slow, but reasonably steady scene loading as I move through the game, I get a 5 - 10 minute delay starting up the first scene, as the client hangs up with no feedback in the final stages of loading the first scene ('equipping avatars'). I just spent over 10 minutes getting loaded into a POT (Rift's End). Yesterday I gave up after 15+ minutest trying to get into Owl's Head, when the OS X Activity monitor finally declared SotA 'not responding'. (The latter happens frequently during scene transitions of all sorts, SotA client is not well behaved from an OS X guidelines point of view.)

This issue was well known in QA. @Attenwood at that time suggested I wait up to 10 minutes before declaring a crash, IIRC. It hasn't changed or improved during public release. With all due respect, this is unacceptable for a public release, and one that was designed for improve perceived performance.

 

@Chris @DarkStarr So far waiting 18 minutes to zone into Owl's Head. This. Is. Ridiculous. And it's gotten worse since the start of R37, at least then I eventually got into scene.

 

I've noticed (on Linux) that the progress indicator often stops at 98% when entering Owl's Head and I have to kill the client and restart to get in.

 

I had a similar experience (see linked thread below). Fortunately, things got better for me, but it is really hit or miss.

As far as housing loading, sometimes it happens to work and in very good time. Other times, I have to get up close to let the game know to populate the home in question. There are times there is a slow to no response while at other times it seems fast. Hit or miss.

Another issue are what appear to be strong memory leaks. I suspect this due to how sluggish combat becomes over time, clearing up completely after a restart.

And there are also strange flashes and graphic glitches (slow loading of avatar on paper doll; cloak showing up in front on same).

None of us should have to go through this, especially knowing that these issues did not exist at all w/ r36 and only appeared during a publish (r37) meant to fix or smooth over the very problems it is now seemingly creating. I don't say these things to be smart, funny or biting with my commentary. Its just the facts on the ground for my OS X experience as of r37.

OS X is advertised as a supported game client. To be fair, not all OS X users have experienced the issues we have. My hope is that people w/ like experiences will speak up and this will get some attention (and correction) from the devs. Meanwhile, also in the interest of fairness, you should post your system specs.

https://www.shroudoftheavatar.com/forum/index.php?threads/r37-and-mac-os-x-client.74498/

 

Yes, that's where the OS X client stalls as well.

 

I believe the client (at least on Linux) is consuming a lot more memory than before. I used to be able to get into Parrot's Roost using the 32 bit client but now it takes over 6GB of memory.

 

I am also freezing up on the loading screen on Mac but mine stops at 50%. If it ever gets passed, it continues to freeze usually stopping in the 70's and at 90-92% again.

The few times I've made it in, the game almost immediately crashes.

Also getting crazy strobe light screen for a minute or 2 before the title screen pops up.

 

I have most of the issues you mention above (strobe light screen on start up and slow loading between scenes). I do get a long pause at 50% as the client hangs there for a good while; the difference for me is that it does eventually move along and I do not experience the crashes you mention.

On a side note, I do get PEACHPIT crashes which usually occur when there is another player in the same area as me in MPO, but this has been going on since prior to r37. Interestingly, and also prior to r37 and currently, should I be in FO the presence of other players is a non issue and no crashes ensue. This leads me to believe that player presence induced crashing is not an issue on my end nor an actual player presence issue, but rather an issue w/ how MPO is configured that for some reason affects me particularly. This was one of the main things I was hoping would be resolved w/ r37.

I welcome Portalarium to invite me for any necessary testing if needed.