As mentioned on <a href="http://www.xlr8yourmac.com" target="_blank">www.xlr8yourmac.com</a> both yesterday and today, one lucky person with a dual 1GHz G4 reports as high as 7625MFLOPS on the Altivec Carbon Fractal test.
My puny G4 500MHz manages 1.6GFLOPS. According to initial 1Ghz G4 benchmarks, a new dual 1GHz is 4.75 times as fast as my 500MHz G4 at Altivec Carbon Fractal Demo.
How does your machine compare? Please post your system and the number of MFLOPS you score and which OS you are under.
**Note: This benchmark can be run on any PPC based Macintosh. For non-G4's, Altivec will automatically be disabled.
<strong>My dual 450 got about 2860 last night (but I was running a couple other programs at the same time).
I was curious what my iBook would get, but I had no idea how bad it would be.... about 380 megaflops! Yikes! That test really does depend on Altivec.</strong><hr></blockquote>
The dual 1GHz PowerMac is faster than your iBook by a factor of 20 at this test.
The dual 1GHz PowerMac is faster than your iBook by a factor of 20 at this test.
</strong><hr></blockquote>
Yeah, and my own Powermac is faster than my iBok by a factor of 8 or so. It's weird, you wouldn't think, stacking up a G3-600 versus a G4-450DP, that the G4 would be that much faster, but I guess in an Altivec-reliant benchmark, it would be.
I recall that disabling the Altivec capability on my G4 seriously crippled it on this benchmark (less than 1/2 the score).
Well, I reran the test under OS 9 and this time I got 194.3 megaflops which is a little closer to yours cosmo. I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
<strong>Well, I reran the test under OS 9 and this time I got 194.3 megaflops which is a little closer to yours cosmo. I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> </strong><hr></blockquote>
By adjusting the various options, there can be a huge range of results in either OS. Plus, you need to upgrade to 10.1.
<strong>I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> </strong><hr></blockquote>
Well, I read somewhere that the current scheduler in Mac OS X is not ideally suited for compute-bound applications (such as this fractal benchmark thingy), and slows them down.
Preemptive multithreading doesn't really help here, either - in OS 9, the benchmark can just hog the CPU as long and as much as it sees fit, whereas preemptive multitasking prevents just that.
<strong>Well, I reran the test under OS 9 and this time I got 194.3 megaflops which is a little closer to yours cosmo. I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> </strong><hr></blockquote>
I was in 10.1.2 when i did the test, all the settings were set to default and no other apps were running.
<strong>MaCommentary: Those dual 450's and 500's really were a great deal. They'll be useful for a long time.</strong><hr></blockquote>
My dual 450 is the only computer I've ever owned that was became faster over the course of my first year using it!
I started of in OS9 all the time, then moved over to OSX and it sped up (due to utilizing the 2nd CPU), then when 10.1 came out it sped up again due to the optimizations. Too bad all computers can't do that.
I'm still very pleased with how fast this thing is for most uses.
Comments
<a href="http://daugerresearch.com/fractaldemos/altivecfractalcarbon.html" target="_blank">http://daugerresearch.com/fractaldemos/altivecfractalcarbon.html</a>
My puny G4 500MHz manages 1.6GFLOPS. According to initial 1Ghz G4 benchmarks, a new dual 1GHz is 4.75 times as fast as my 500MHz G4 at Altivec Carbon Fractal Demo.
How does your machine compare? Please post your system and the number of MFLOPS you score and which OS you are under.
**Note: This benchmark can be run on any PPC based Macintosh. For non-G4's, Altivec will automatically be disabled.
[ 01-30-2002: Message edited by: Nostradamus ]</p>
I was curious what my iBook would get, but I had no idea how bad it would be.... about 380 megaflops! Yikes! That test really does depend on Altivec.
<strong>My dual 450 got about 2860 last night (but I was running a couple other programs at the same time).
I was curious what my iBook would get, but I had no idea how bad it would be.... about 380 megaflops! Yikes! That test really does depend on Altivec.</strong><hr></blockquote>
The dual 1GHz PowerMac is faster than your iBook by a factor of 20 at this test.
[ 01-31-2002: Message edited by: Nostradamus ]</p>
Results: 1255.5 Megaflops
Is that..good?
Yippee, I won...........oh wait
<strong>
The dual 1GHz PowerMac is faster than your iBook by a factor of 20 at this test.
</strong><hr></blockquote>
Yeah, and my own Powermac is faster than my iBok by a factor of 8 or so. It's weird, you wouldn't think, stacking up a G3-600 versus a G4-450DP, that the G4 would be that much faster, but I guess in an Altivec-reliant benchmark, it would be.
I recall that disabling the Altivec capability on my G4 seriously crippled it on this benchmark (less than 1/2 the score).
<strong>Well, I ran it on my 333mhz Powerbook G3 and got 155.7 Megaflops.
Yippee, I won...........oh wait </strong><hr></blockquote>
Odd my 333 iMac got 206megaflops.
<strong>
Odd my 333 iMac got 206megaflops.</strong><hr></blockquote>
PS 2 : 6152 Megaflops, haha
Pretty cool demo, but Cinebench looks even cooler.
<strong>Well, I reran the test under OS 9 and this time I got 194.3 megaflops which is a little closer to yours cosmo. I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> </strong><hr></blockquote>
By adjusting the various options, there can be a huge range of results in either OS. Plus, you need to upgrade to 10.1.
<strong>I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> </strong><hr></blockquote>
Well, I read somewhere that the current scheduler in Mac OS X is not ideally suited for compute-bound applications (such as this fractal benchmark thingy), and slows them down.
Preemptive multithreading doesn't really help here, either - in OS 9, the benchmark can just hog the CPU as long and as much as it sees fit, whereas preemptive multitasking prevents just that.
Bye,
RazzFazz
<strong>Well, I reran the test under OS 9 and this time I got 194.3 megaflops which is a little closer to yours cosmo. I wonder why it would do 194 under OS 9 and 155 under OS X. Of course I haven't upgraded to 10.1 yet so that may be it. <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" /> </strong><hr></blockquote>
I was in 10.1.2 when i did the test, all the settings were set to default and no other apps were running.
<strong>MaCommentary: Those dual 450's and 500's really were a great deal. They'll be useful for a long time.</strong><hr></blockquote>
My dual 450 is the only computer I've ever owned that was became faster over the course of my first year using it!
I started of in OS9 all the time, then moved over to OSX and it sped up (due to utilizing the 2nd CPU), then when 10.1 came out it sped up again due to the optimizations. Too bad all computers can't do that.
I'm still very pleased with how fast this thing is for most uses.