Opinion – A long time ago, in a galaxy far, far away, when gas cost 25 cents a gallon, my good self and another venerable TGD writer working together on another mag coined the phrase Megahurtz Madness. These were the golden days when Intel and AMD were vying for supremacy in terms of who hit the magical – and meaningless – milestone of producing the first 1GHz CPU.
History records that AMD beat its considerably-larger rival to the mark by a couple of weeks. Intel then attempted to leapfrog the young upstart contender by launching a 1.1GHz Pentium III, which, rather tragically, didn’t work.
Anyway, our point was that measuring system performance in terms of raw megahertz was a complete waste of time. Countless other factors had a more serious impact on overall performance and normal users simply wouldn’t know or care how fast their chip was running.
The whole affair was a sad indictment of publications with an unhealthy oscilloscope and benchmark fixation, who genuinely seemed interested in whose processor was the faster. Needless to say, Via did not trouble the scorers.
Perhaps Intel listened to what we said. These days it appears to be making every effort to obfuscate the clock speed of its chips, preferring to give them hip and groovy names almost bereft of numbers.
The company also flirted briefly with the concept of NURBS, which it claimed would improve internet performance by compressing everything to make it download faster. Thankfully, this absurd notion was soon dropped and consigned to the dustbin of history along with other daft Intel ideas such as the Camino Rambus chipset and the Timna system-on-a-chip.
Naturally, in our caring, sharing manner, we poured scorn on NURBS at the time, but, rather chasteningly, Intel was on the right track after all.
Reading claims from rival browser makers about how their client outperforms its rivals in various esoteric rendering benchmarks is an unhealthy return to the ‘mine’s bigger than yours’ days we rather hoped were long gone.
End users – at least those without oscilloscopes and benchmark suites – are, quite reasonably, interested in how fast their favorite sites load. But I contend that the speed of Java or Ajax handling is of no consequence to the average user.
They will be constrained not by the performance of their FireFox, IE8 or Opera client, but by their line speed, the power of the servers at the far end, how many people are trying to access the site and, most critically, by the bandwidth contention imposed on the user by their ISP in terms of the number of people it attempts to cram down a single pipe.
I therefore put it to you, members of the jury, that paying a little extra to move to an ISP with a 100 to one contention ratio rather than 10,000 to one and doesn’t throttle bandwidth at peak times will reap more benefit in real terms than agonising over whether you should be using Safari, IE8, Netscape or Mosaic.
To use a really bad car analogy, a Bugatti Veyron may be capable of 250mph. But on a 50 mile run on a normal country road, it won’t get there any faster than my 145mph diesel Jaguar. Even the editor’s 100mph Subaru could give it a run for its money.