On Sat, 2010-06-19 at 12:51 -0400, Felix Miata wrote:
On 2010/06/19 11:03 (GMT-0400) Adam Tauno Williams composed:
But spending dollars to save *TIME* [and equivalent frustration] is always a worthwhile; _Not_ always, not when supply of time is abundant and/or time saved is contextually small, or when initial investment funding is limited.
Wow, abundant time. I don't even know how to respond to that. It is more cost-effective to use a slower computer than a faster one? Eh...
Agree. But replace a P4 system with an i7 system and there is no Likely the preponderance of P4 users are not i7 candidates at this point in time.
Well, Macs are eating up market share (and they are all i5/i7s). And PC sales... they don't suck. Analysts [IDC] project a 10% increase in units shipped for 2010. Others [Cohen] estimate as high as 20%. So somebody disagrees with you.
confusing that for mere change; its *progress* when procedures that took minutes complete in seconds. I can't imagine a typical I/O bound process on a Cedar Mill P4 taking magnitudes less time on an i7 absent costly I/O cost premium to thwart the bottleneck, which would be comparing beets to raisins. For a process taking mere seconds on an i7 using mainstream I/O systems to take minutes I'd expect the elder CPU to be sub-1GHz or a first generation Duron or P3 Celeron or slower, and its HD bus to be sub-ATA3.
Bah. I've been a corporate UNIX and network admin for ~15 years. These arguments are always trumpeted by the anti-progress crowd [and they appear technically sound]. But its crap. New systems arrive and tests show they smash the old system for actual real-world performance. The processors are faster, the disks are faster, the busses are faster, the chipsets are more efficient, and the firmware smarter. Not to mention more RAM per adjusted $1. New wins, every single time, and always by a easily measurable margin.
Comparing differences of bazillions of megaflops among different CPUs instead of the much more common I/O bound operations is sales speak, not practical
Nope, its 100% practical. Whether it is a server running an RDBMS system or mom's PC sorting 5,000 photos or a workstation editing an 18 chapter manual - the performance difference is huge and very real.
sense spoken by cost/benefit analysts. When funds are unlimited, sure more is better, but mainly only out of control governments and special situations have that, not mainstream individuals and businesses.
Seriously? Your political anti-government bent doesn't have anything to do with the argument. It is more convincing to readers if you leave that out. Mainstream individuals and business benefit significantly from current hardware. A Gartner study showed that the single biggest improvement for office worker productivity - give them a second monitor [more hardware]. At which point, of course, they will be running more applications concurrently....
They are not interchangeable. Nope. But newer systems are clearly progress. I can assure you plenty of cost accountants would disagree.
Ok, and I can assure you that many don't.
Newer systems are dissipating more heat, not what I call progress when paying for more dead dinosaurs to be consumed to run both puters and air conditioners. Who's paying your electric and hardware bills?
Bills are paid via the sale of the worker's productivity.
And modern hardware can successfully suspend/hibernate - which older
hardware *very* often fails to do.
--
Adam Tauno Williams