Hi,
I've just performed some testing on this before buying a new system myself.
It seems that on a modern system the bottleneck is the CPU and/or FSB, while I always assumed this was the harddisk. It's sad that Visual Studio only uses 1 core at 100% so increasing cores doesn't help (yet). I found a link on multiple core compilation but these results couldn't be displayed in the Error List, which I found to big of a trade off.
http://www.hanselman.com/blog/FasterBuildsWithMSBuildUsingParallelBuildsAndMulticoreCPUs.aspx
System A:
Q6600 2.4Ghz,
FSB 800Mhz,
RAM 8Gb DDR2 800MHz,
Raptors in RAID 0
System B:
Pentium D 3.2Ghz,
FSB 800Mhz,
RAM 4Gb DDR2 333MHz (the bottleneck?),
normal HD in RAID 0
Comparison A - B
From harddisk the compilation of the project on B took 92% more time than on A.
(26seconds vs 50seconds)
Comparison A Harddisks - A RAMDisk
The RAMdisk compilation took 12% less than from the harddisks
(26seconds vs 23seconds)
Comparision B Harddisks - B RAMDisk
The RAMdisk compilation took 12% less than from the harddisks
(50seconds vs 44seconds)
When the new system is in I will post some details on performance. It's 1333FSB & RAM.
Cheers,
Gab