The Source for Performance News, Tips & Reviews
Click for Voyager Hard Drive Docks!
|Memory Performance Benchmarks on New G4/466 vs. Prev. G4/450|
|Last Updated: 1/11/2001, 9:07 PM|
|A reader sent memory benchmark results of tests with a 100mhz bus G4/450 system vs. a new G4/466 system with 133MHz bus. In most cases the new G4 showed substantially higher performance (up to 140%).|
I thought you might be interested in these measurements I made of the memory
throughput of the new machines. Attached is a comparison - the numbers in
the chart are in Mb/s. (The CopyBits scores are attributable to the nVidia
card). [Initial driver release] The measurements were taken with a tool I wrote.
I would be very curious to see what the results would be with a 7450 CPU (256k on-chip L2 cache and 1MB L3 cache) with a Radeon card.
Kent's test results are in the table below. (I'm assuming the numbers are MB/sec rates).
|Vector Writes to
|Vector Writes Cache
|GPR Writes to GDevice
|GPR Writes reversed
|GPR Writes (cache)
|GPR Writes byte reversed
|GPR Writes Cache
|FPR Writes (cache)
|FPR Writes Uncached
|FPR Writes to GDevice
For more info on the new G4s - see our page with comments, features, photos and pricing.
|Return to Main News Page|
See full list!
Mac Help Now!
Freq. Asked Questions
Mac Model Articles
CPU Upgrade Articles
Make a Difference: Be the Difference
Randy Mita, Brad Lau and I came up with a new slogan and theme for Mac advocacy - "Make a Difference: Be the Difference". We're asking that all Mac owners take advantage of every opportunity to let others know the Mac advantage. It's a grass-roots approach that will surely help.
This site is not affiliated with nor supported by XLR8 or Apple Computer.
Thanks to my wife, the ultimate computer widow
for putting up with all the time I spend with the site work, tests and email.
= Top of Page =
Copyright © 1997-2014. All Rights Reserved
All brand or product names mentioned here are properties of their respective companies.
Legal: Site Privacy and terms/conditions of use. of use.