|
About VirtualGL Downloads Developer Info Library |
TigerVNCThe VirtualGL Project is proud to announce our affiliation with the TigerVNC Project. TigerVNC was founded by some of the former TightVNC developers and aims to provide a VNC solution with similar performance to TurboVNC but based on the newer RealVNC 4 and X.org code bases. The existing TurboVNC code base will continue to be supported for the foreseeable future, but we will be working to migrate to TigerVNC over the long term. Throughout 2009, extensive work was done to improve the performance of the low-level JPEG codec (libjpeg-turbo) used by TigerVNC. libjpeg-turbo is now an 80-90% solution in the aggregate, when compared to TurboJPEG/IPP, but it is actually faster than the latter in some cases. There are still, however, some buffer copies and additional overhead which cause TigerVNC to fall short of its full potential. I am seeking sponsorship to continue the work of optimizing and packaging TigerVNC. Comparison of the Performance of TurboVNC 0.6 and TigerVNC r3939This brief study uses the same techniques which are described in the VirtualGL 2.1 Performance Study. In all cases, JPEG quality 92 with no chrominance subsampling (quality level 8 in TigerVNC) was used. Machine 1: Dell Precision Workstation 470, 1 x 2.8 GHz Intel P4 Xeon (Hyperthreading enabled), NVIDIA QuadroFX 370 (driver 190.42), CentOS 4.8 (64-bit) + updates 11/9/09, VirtualGL 2.1.4 Machine 2: HP Pavilion Slimline s5100z, 1 x 2.6 GHz AMD Athlon 64 X2 5050e (dual-core), NVIDIA GeForce 6150SE nForce 430 (driver 190.40), CentOS 5.4 (64-bit) + updates 11/9/09, VirtualGL 2.1.4 Machine 1 --> Machine 2
Machine 2 --> Machine 1
|
|
![]() | All content on this web-site is licensed under the Creative Commons Attribution 2.5 License. Any works containing material derived from this web-site must cite The VirtualGL Project as the source of the material and list the current URL for the VirtualGL web-site. |