Once more into the breach (or, where I’m headed with the speed tests)...
I’ve had some interesting conversations with folks about the benchmarks over the last couple of days. Several have made suggestions and, even better, sent me some data. I’m looking forward to updating the Cinebench benchmarks with AMD 2400+ data and G5 1.8 data once I get a complete set of each. I may even wander down to the local Apple store soon to see if they’ve got a G5 I can use for a few moments.
One reader who claimed to be discussing the subject with Maxon employees claimed that I had “used Cinebench incorrectly” because I used the time values instead of the raw CPU scores in the testing. Frankly, while some people may prefer the CPU benchmarks, I find the time and frames per second more meaningful. I can understand them right away. The argument that they’re not meaningful leads to the question: “If they’re not meaningful and don’t mean anything, why did the testers choose to display them?” I feel I must also point out that Rob-ART over at Barefeats.com uses the time values as well, and he has a lot more experience than I do. Frankly, I’m on the same page he is about why using the time values is a valid tactic. I believe folks are fussing at me about what they’re seeing because they don’t like the result. It appears to be a “shoot the messenger” approach.
Another reader suggested I run benchmarks against AMD’s Opetron. That’s a fair suggestion but not something I’m going to pursue right now. I’m a home and small business user and want to confine myself to the market I know. For the moment, the Opteron is really a server CPU, just like the Xeon. Secondly, there are other folks out there already gathering that kind of data (Barefeats.com is one), and I don’t have access to the assets I’d need to do that. So, for now, I’m going to constrain myself to running G5’s and some G4’s against Athlons and Pentium 4’s, cpu’s I would be more likely to buy.
Several readers made me aware of what I believe to be a pretty big trend. Many folks are treating the website like it’s a Playboy magazine: they claim to be reading the articles, but they’re really only looking at the pictures. Please read what I have to say and then if you have a comment or suggestion, send me an e-mail. It’ll save us both time and worry.
I’m going to generate some Photoshop files in the next week or so and run some informal tests to at get some G4 and AMD performance data using Photoshop 7 filters and tasks. If anyone in the Houston area would be interested in working with me on testing their G5, send me an e-mail and we’ll discuss what we can do.
One reader who claimed to be discussing the subject with Maxon employees claimed that I had “used Cinebench incorrectly” because I used the time values instead of the raw CPU scores in the testing. Frankly, while some people may prefer the CPU benchmarks, I find the time and frames per second more meaningful. I can understand them right away. The argument that they’re not meaningful leads to the question: “If they’re not meaningful and don’t mean anything, why did the testers choose to display them?” I feel I must also point out that Rob-ART over at Barefeats.com uses the time values as well, and he has a lot more experience than I do. Frankly, I’m on the same page he is about why using the time values is a valid tactic. I believe folks are fussing at me about what they’re seeing because they don’t like the result. It appears to be a “shoot the messenger” approach.
Another reader suggested I run benchmarks against AMD’s Opetron. That’s a fair suggestion but not something I’m going to pursue right now. I’m a home and small business user and want to confine myself to the market I know. For the moment, the Opteron is really a server CPU, just like the Xeon. Secondly, there are other folks out there already gathering that kind of data (Barefeats.com is one), and I don’t have access to the assets I’d need to do that. So, for now, I’m going to constrain myself to running G5’s and some G4’s against Athlons and Pentium 4’s, cpu’s I would be more likely to buy.
Several readers made me aware of what I believe to be a pretty big trend. Many folks are treating the website like it’s a Playboy magazine: they claim to be reading the articles, but they’re really only looking at the pictures. Please read what I have to say and then if you have a comment or suggestion, send me an e-mail. It’ll save us both time and worry.
I’m going to generate some Photoshop files in the next week or so and run some informal tests to at get some G4 and AMD performance data using Photoshop 7 filters and tasks. If anyone in the Houston area would be interested in working with me on testing their G5, send me an e-mail and we’ll discuss what we can do.
posted by Andy Foster at
3:21 AM
0 Comments:
Post a Comment
<< Home