Does VC++ (I assume we’re only talking about compilations on Windows platforms given what you do and the other info) take advantage of multithreaded compiles? If so, the most cost effective path is probably upgrading to an X2 3800+ which is both faster and multi-thread capable.
Exactly how long does an average compile take for you, out of curiosity. Is it really that lengthy of a process? (All I ever compile are small programs <20k lines or so, and that seems more or less instantaneous.)
It would help if you provided more details about the types of programs you compile, what your compile speeds are like now, how much faster you want them to be (2x? 3x? 10x?), and/or how much you’re willing to spend (now or in the future). Any upgrade will help - the question is how much of an improvement do you want to see and how much are you willing to spend to get it?
As for Vista: officially, its release seems to be confirmed for end of 2006; unofficially, of course, it’s “whenever the hell Microsoft finishes with it.” :)
It takes about a minute to do a full recompile and relink on my current project - maybe 20 seconds for an incremental. But lots of other stuff is kinda slow-ish, too (opening Photoshop or Firefox, etc. Rebooting takes me about ~2 minutes).
VC doesn’t utilize hyperthreading, though supposedly the brand new version that came out last week might (but it’s kinda buggy otherwise, apparently).
For me, if it’s just a matter of spending $500 and 60 minutes of time (finding, ordering, and replacing a CPU), then a 15% improvement would justify it (maybe even less).
If it’s the matter of a whole new PC, with all the re-installing, etc (usually takes me at least a day to get set up), then I’d like to see a 50-80% improvement.
How does one even know what options I have to replace my current AMD CPU? I looked at AMD’s site, but there’s no info like “If you have this, you can upgrade to any of these…” Is there some reference source for that?
Hard drive access and read speeds.
Compiling and memory paging during debugging are very IO intensive, and faster hard drives will noticably increase the apparent speed of the machine. So, use SATA1 or SATA2 instead of IDE.
As stated previously, the more processors you can throw at a build, the faster it will finish. Also, your compiler has to support concurrent compiles.
64-bit dual-processor dual-core 2.8GHz (11.2 GHz of processing power)
64-bit single-processor dual-core 2.8GHz (5.6GHz)
64-bit single-processor 3.8 GHz
I’m not sure how much it will help during the linking stage. I don’t even know which win32 compilers support concurrent linking.
bigger on-die cache = better compiler performance
Stusser can probably tell you more about optimizing memory speeds than I can. I’ll just say that it’s preferable to have PC3200 with tight timings rather than PC4000 with loose timings.
As red guy points out, the fastest Athlon XP CPU is the 3200+, which is only a modest improvement over your 3000+. So you’d be talking a new CPU and MB to upgrade from that, which would be at least $250 - $300; more if you want a serious performance improvement.
A faster HD would certainly help: you could buy the fastest 120GB IDE HDD you can find, move your system partition(s) onto it, then turn your original drive into spare storage. [Given the age of your system, it probably doesn’t have native SATA support, but I could be wrong.]
More RAM never hurts - an extra gig should be no more than $100 - but you need to know more about your MB and RAM configuration to know what RAM you should buy. Compaq / HP should have that info on your system; or go someplace like Crucial and see if they list your system and what RAM it can use.