I'm looking for some enlightened guidance regarding vnc expectations of the new Intel-based Mac hardware.
Currently, I have OSXvnc running on a Powermac G4 Dual1.42 and am accessing it using CotVNC from a Powerbook G4 1.67. It's reasonably useable for the occasional time I need it but, for regular use, it's dog-slow, even on a LAN!!
If I replaced the above hardware with, say an Intel Core Duo Mini (OSXvnc) and MacBook Pro (CotVNC), all other things being equal, would you expect that a future Intel-native or Universal binary release (of both apps) to yield a significant boost in speed/response?
I'd be very grateful for a frank answer, as this is the only application improvement (if any) that would entice me to upgrade at the moment!
If I'm missing the point entirely, and someone knows how to improve the responsiveness within a LAN (using RAW encoding option), then I'd be very happy to hear it!!!
Thanks all in advance.
Thanks in advance.
On the responsiveness issue (not using Intel yet) I find CotVNC / OSXvnc combo very usable setting a CotVNC profile to 256 colors. It is a huge difference over the default, which does seem too slow to use practically.
I don't think using Intel will change things much, but as someone else said here if you change the profile to use 256 colours you'll get quite an improvement on slow or remote networks. On LAN's (100baseT) I use thousands of colours fairly easily.
Damn useful tool...
Thanks for the replies. I totally agree, the 'damn useful tool' factor overides the latency/responsiveness I'm experiencing. I shall try experimenting with 256 colours. Does this offer an improvement irrespective of the server (OSXvnc) display setting?
Regarding the 'Intel' aspect to this thread, I wondered how much of the "5 times faster" (!!) marketing blurb might actually make it's way through to real life apps, specifically vnc related, particularly as a result of the much higher FSB compared to the PowerPC machines. I guess not much then :-)
Processor speed is irrelevant when talking about VNC (well, almost). The bandwidth of the network is what slows things down (as well as other non-speed related issues).
I have used a Mac to PC connection for two years and discovered that 256 colors will vastly improve the speed ("vastly" is, of course, a YMMV term).
After my housemate complained about CotVNC eating large amounts of CPU in random spikes I profiled CotVNC using shark and traced part of the responsiveness problem with non 8bpp modes to really badly written endian-safe copying functions in the framebuffer drawing code.
I've got a preliminary fix to this which has smoothed out the CPU usage significantly. A patch will be sent to the developers when I've finished cleaning up bits and pieces in a few days time (hopefully).
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.