It was late at night when I wrote that last post, and I did not have much time to tinker.
Playing with the system today reveals that it is indeed much, much faster and smoother at everything in "real life" tasks. Office 2008 is very bearable. Scrolling is smoother and faster, and while there there is still some small lag, it is much less pronounced than before. Safari is much snappier, and the fading effects and such which draw on Quartz are much less ragged. It's no Mac Pro, but it'll do me for the purpose I have in mind for another couple of years at least.
I also just cross-checked the Geekbench score I got last evening with the published score for the Dual 500 Gigabit Ethernet machine from which the daughter card came. The Dual 500 Cube is actually faster with a Geekbench score of 512 than the donor Gigabit Ethernet tower was at 489 — possibly due to the graphics card, as my Cube has a Radeon 7500 — and checks in just under a single processor 933MHz Quicksilver at 523. These are, however, average scores: no doubt the single processor Quicksilver would outshine the dual 500 on something like the Memory scores, whereas the dual G4 would presumably handily beat the Quicksilver in threads. I think threads the more important of the two in many ways.
The upshot is that this really is a great upgrade and one that I would recommend to Cube owners, as it is very cheap (maybe $25), though it requires a little mechanical know-how and some soldering to pull off. The temperature in the upgraded Cube seems to be, on average, about 7ºC hotter than it was in the machine stock, now standing at at about 36ºC under normal operating conditions. I am just doing web and text work, with a little MP3 encoding and such, nothing more. I saw it peak when running the benchmarking utilities at 43ºC.
These temperature numbers are not, I should add, based on the temperature of the core itself, but on the temperature sensor in the WD Hard Drive (my utilities do not detect any other temperature sensor in the machine). Assuming this is about right — and I am assuming that it is, given that the whole top cover of the HD physically touches the heat sink in a Cube, and the the HD has heatsinks that contact the core's heatsink — 35ºC should be kosher, should it not? I have never bothered to measure temperatures before, know nothing about it, and would like not to fry my "new" G4 Cube, which I have treasured since I bought it when new.
Playing with the system today reveals that it is indeed much, much faster and smoother at everything in "real life" tasks. Office 2008 is very bearable. Scrolling is smoother and faster, and while there there is still some small lag, it is much less pronounced than before. Safari is much snappier, and the fading effects and such which draw on Quartz are much less ragged. It's no Mac Pro, but it'll do me for the purpose I have in mind for another couple of years at least.
I also just cross-checked the Geekbench score I got last evening with the published score for the Dual 500 Gigabit Ethernet machine from which the daughter card came. The Dual 500 Cube is actually faster with a Geekbench score of 512 than the donor Gigabit Ethernet tower was at 489 — possibly due to the graphics card, as my Cube has a Radeon 7500 — and checks in just under a single processor 933MHz Quicksilver at 523. These are, however, average scores: no doubt the single processor Quicksilver would outshine the dual 500 on something like the Memory scores, whereas the dual G4 would presumably handily beat the Quicksilver in threads. I think threads the more important of the two in many ways.
The upshot is that this really is a great upgrade and one that I would recommend to Cube owners, as it is very cheap (maybe $25), though it requires a little mechanical know-how and some soldering to pull off. The temperature in the upgraded Cube seems to be, on average, about 7ºC hotter than it was in the machine stock, now standing at at about 36ºC under normal operating conditions. I am just doing web and text work, with a little MP3 encoding and such, nothing more. I saw it peak when running the benchmarking utilities at 43ºC.
These temperature numbers are not, I should add, based on the temperature of the core itself, but on the temperature sensor in the WD Hard Drive (my utilities do not detect any other temperature sensor in the machine). Assuming this is about right — and I am assuming that it is, given that the whole top cover of the HD physically touches the heat sink in a Cube, and the the HD has heatsinks that contact the core's heatsink — 35ºC should be kosher, should it not? I have never bothered to measure temperatures before, know nothing about it, and would like not to fry my "new" G4 Cube, which I have treasured since I bought it when new.