Disclaimer: It's Friday! The difference Rod is seeing is because he "forgot the floating city!" $5 to the person who can name the obscure, horrible movie that I actually paid money years ago to go see that has that outburst. It was so bad, I remember making fun of the line during the movie, after the movie, and even remembering the stupid name of the movie. It will be a test of your Google-Fu (I think). If I give more details it becomes easier to find.
For Rod's stuff, my guess would be optimized drivers, not that it matters as long as it's working now. Dave Lum - Systems Engineer [EMAIL PROTECTED] - (971)-222-1025 "When you step on the brakes your life is in your foot's hands" -----Original Message----- From: Ben Scott [mailto:[EMAIL PROTECTED] Sent: Friday, April 25, 2008 6:37 AM To: NT System Admin Issues Subject: Re: 64Bit Vista on laptop? On Fri, Apr 25, 2008 at 2:37 AM, Ken Schaefer <[EMAIL PROTECTED]> wrote: > Again - I doubt this is caused by CPU. Besides the almost 99.9% instruction compatibility > between the two, there would be /so/ many bench marks showing how poorly AMD CPUs > compare to Intel CPUs out there, especially in the driver forums etc. Something else is up. While the instruction sets are compatible, there are significant architectural differences between AMD and Intel CPUs. That can affect performance. A given microarchitecture may be good at one thing, but not so good at another. For example, one design may have good floating point performance, while another may excel at integer calculations. Another scenario is that a given executable/library was built with compiler optimization for a particular microarchitecture, and running on a different microarchitecture yields a performance penalty, because one design's optimization is another design's worst-case. In the days of the P4, a lot of programs ran slower on a P4 than a P3 or AMD, because most executables were optimized in a way that was worst-case for the P4's really deep pipeline. I'm not saying that's what's going on in Rod Trent's case, just that there's more to a CPU than the instruction set. Indeed, I think that drawing the conclusion "Intel rocks; AMD sucks" from Rod Trent's experience is premature. Way too much changed. For one, the AMD and Intel chips are in no way compatible at the board level, so switching CPUs also means you're switching memory controllers, disk controllers, bus controllers, and who knows what else. And that can mean not just silicon, but quality of device driver code, too. He also switched from 32-bit Windows to 64-bit Windows. AMD64 has 16 general-purpose registers, as opposed to the i386's 4/8. Given compiler optimization, that can make a big difference for function calls, as it's more likely all arguments can be stored in registers instead of pushed on to the stack. I don't know anything about architectural differences in the Windows code between i386 and AMD64, but there may well significant reworking of the kernel, too. There may also have been issues with the software configuration his old laptop, or a change in the architecture of his anti-virus software, or who-knows-what. In short, far too much changed to simply blame it on the CPU. Without profiling the systems to find out where the bottlenecks are, we really can't know. Of course, for Rod's purposes, that doesn't matter -- all that matters is laptop Y is much faster than laptop X. But that doesn't help the rest of us, unless we also happen to have laptop X. :) -- Ben ~ Upgrade to Next Generation Antispam/Antivirus with Ninja! ~ ~ <http://www.sunbelt-software.com/SunbeltMessagingNinja.cfm> ~ ~ Upgrade to Next Generation Antispam/Antivirus with Ninja! ~ ~ <http://www.sunbelt-software.com/SunbeltMessagingNinja.cfm> ~
