Existing users, 
 Xbench - 1.3benchmarking utility |
|
||||||||||||||
|
|||||||||||||||
Feedback Summary:
| This Version: | |||||
| Overall Rating: | Not rated (0.0) | Features: | Not rated (0.0) | Support: | Not rated (0.0) |
| Ease of Use: | Not rated (0.0) | Quality / Stability: | Not rated (0.0) | Price: | Not rated (0.0) |
Key to Types of Feedback:
Reviews 
Troubleshooting 
Usage Tips 
Developer Notes 
Commentary 
Featured Reviews
Overdue for an update - Version: 1.3, 4/2/2009 03:34PM PST
BradMacPro87
it's getting to be two and a half years old. Not optimized for Mac OS 10.5 or Nehalem processors.
1.3 Still featurless and uninformative!!! 
- Version: 1.3, 9/19/2007 04:09AM PST
(0 of 2 users found this comment useful)
Tesselator
XBench seems to be all that there is for Mac as benchmarking goes.
It's clearly the worst benchmarking application I've ever used on ANY platform. Even ancient Amiga or Windows 2 benchmarking utilities kick the crud out of this. It supplies you with no interesting information and only runs stupidly simple non-real-world tests. It doesn't know and can't test multiple CPU cores (single core tests ONLY) so it's not even good for that really. It might be ok for testing various drive configurations but that would be about it.
XBench 1.3 is just another in a series of let-downs. When will Mac users get a REAL benchmarking suite?
Most Recent Replies: View All 2 Replies
- 1.3 Still featurless and uninformative!!!
I can't understand the results - Version: 1.3, 6/27/2007 01:50AM PST
(1 of 1 users found this comment useful)
davidasta1
If you take a look in the "Results Comparison Site" (http://db.xbench.com/) and you order by "Average Score", the results are a bit unbelievable.
How is possible that a Mac mini (Intel) is more powerful than a MacBook Pro and MacBook?
How is possible that an AppleTV is more powerful than a iMac G5 and a Mac mini (PPC)?
And looking inside the models, how is possible that a MacBook with similar results than mine, even worse in some cases, have a better performance, more than 30.68%?
Compare, for example "Macbook" MacBook vs "DAG Computer" MacBook.
Are the results absolutely trustworthy?
How is possible that a Mac mini (Intel) is more powerful than a MacBook Pro and MacBook?
How is possible that an AppleTV is more powerful than a iMac G5 and a Mac mini (PPC)?
And looking inside the models, how is possible that a MacBook with similar results than mine, even worse in some cases, have a better performance, more than 30.68%?
Compare, for example "Macbook" MacBook vs "DAG Computer" MacBook.
Are the results absolutely trustworthy?
Most Recent Replies: View All 1 Replies
Nice results, GUI needs some work - Version: 1.3, 8/14/2006 08:31AM PST
Paul_Vail_674
It's a decent little app, and if one needs to do localized speed comparisons, RELATIVE tests may help point out where systems would benefit from:
killing off unnecessary widgets, apps, trinkets and such
differences between hard drive I/O issues
bad cables and oddities inside the machine
differences between OS versions.
But don't go completely by these numbers from this app. There's a lot to making an optimized and stable machine. This is only one tool.
With that said, it's fun to run the test and compare different machines around the office. Kind of gives a visual idea of why certain machines 'feel' faster.
My only complaint thus far is based on a GUI bug: do NOT select the Help menu items. Selecting the Help:Read me displayed a welcome dialog that wouldn't dismiss when OK was selected. Can't get rid of it. Can't quit out of Xbench because there's a dialog that hasn't been dismissed. Very ugly to have to force-quit the app. Perhaps this is system-specific, but it is such a basic GUI element, I would have expected it to be rock solid. Many users would force restart their machine if this occurs... and no app should appear to get stuck on something this simple.
The price is right, the numbers are interesting, the rest of the interface is sparten and workable (I don't need eyecandy in a 'Mac Way' for everything). It would be very cool if down the road the app developed internal comparitive charts like the old MacBench did -- or rather than have the developer waste their time on writing that code, just have Xbench read results files and build a web page for comparitive analysis similar to the pages used on their website. This would allow comparison across multiple results so two, three, a dozen machines could all be graphed together. Just a thought.
thank you for providing this. Hope the GUI bugs get fixed fast. The help menu would seem to be an important solution for FAQ and 'stupid question' avoidance.
killing off unnecessary widgets, apps, trinkets and such
differences between hard drive I/O issues
bad cables and oddities inside the machine
differences between OS versions.
But don't go completely by these numbers from this app. There's a lot to making an optimized and stable machine. This is only one tool.
With that said, it's fun to run the test and compare different machines around the office. Kind of gives a visual idea of why certain machines 'feel' faster.
My only complaint thus far is based on a GUI bug: do NOT select the Help menu items. Selecting the Help:Read me displayed a welcome dialog that wouldn't dismiss when OK was selected. Can't get rid of it. Can't quit out of Xbench because there's a dialog that hasn't been dismissed. Very ugly to have to force-quit the app. Perhaps this is system-specific, but it is such a basic GUI element, I would have expected it to be rock solid. Many users would force restart their machine if this occurs... and no app should appear to get stuck on something this simple.
The price is right, the numbers are interesting, the rest of the interface is sparten and workable (I don't need eyecandy in a 'Mac Way' for everything). It would be very cool if down the road the app developed internal comparitive charts like the old MacBench did -- or rather than have the developer waste their time on writing that code, just have Xbench read results files and build a web page for comparitive analysis similar to the pages used on their website. This would allow comparison across multiple results so two, three, a dozen machines could all be graphed together. Just a thought.
thank you for providing this. Hope the GUI bugs get fixed fast. The help menu would seem to be an important solution for FAQ and 'stupid question' avoidance.
Hard Locks my machine - Version: 1.2, 8/7/2005 10:46PM PST
(0 of 3 users found this comment useful)
mackopes
This version hard locks my PowerMac. 1.1.3 works fine.
Most Recent Replies: View All 1 Replies
Que mesure XBench? - Version: 1.1.3, 7/20/2004 06:03AM PST
(8 of 15 users found this comment useful)
Frederic Poncet
XBench mesure la performance de chaque composant de l'ordinateur ou de quelques opérations élémentaires. Mais la note globale ne reflète pas les performances réelles de l'ordinateur.
Par exemple, si vous ajoutez de la RAM à votre ordinateur, il devient sensiblement plus rapide mais XBench ne voit aucune différence. Avec 256 Mo ou 768 Mo de RAM, la valeur moyenne mesurée par XBench est toujours la même...
Par exemple, si vous ajoutez de la RAM à votre ordinateur, il devient sensiblement plus rapide mais XBench ne voit aucune différence. Avec 256 Mo ou 768 Mo de RAM, la valeur moyenne mesurée par XBench est toujours la même...
Most Recent Replies: View All 2 Replies
- Que mesure XBench? - Translation (1 replies)
Ahem. Get the facts straight. - Version: 1.1.3, 11/5/2003 08:18PM PST
(22 of 23 users found this comment useful)
pwargo
Uhhhh, blueskymining I hate to burst your bubble, but you've made a few less-than-correct statements.
First, The G5 is cycle-per-cycle qute a bit faster than the G4. It has little to do with 64-bit or not - while a 64-bit OS and applications will be somewhat faster (and a whole lot more efficient for large-memory apps), the real gain in in processor efficiency. The G5 can execute four instructions per clock cycle (in a perfect world), the G4 is limited to three. In addition, the G5 has a much faster frontside bus, as well as considerably more I/O and memory bandwidth.
Second, MP can be quite a bit more efficient than you state. I've been working with MP systems for 10 years now, and I've seen some systems that hit linear (and in one weird case, super-linear) performance curves. A *lot* has to do with the OS and application, and how well they thread. Typically, on the G5, I'm seeing about 80-85% improvement by turning the second processor on.
Finally, have you actually worked on a G5? I'm lucky, I have - both at work and at home I have dual G5's. At work, we've seen some truly impressive results on a heavy mix of I/O and CPU. Results that put a dual G4 to shame. (Not to mention a dual 2.4GHz P4 Xeon and a Sun 280R.) As for home, I work on a variety of A/V applications and I am impressed with the results. (To be honest, I used the systems at work, got hooked, and then lucked out and found a way to afford a dual G5.)
So, artificial benchmarks aside, I have no problem stating that the G5 is the fastest desktop I've ever had the pleasure to use.
First, The G5 is cycle-per-cycle qute a bit faster than the G4. It has little to do with 64-bit or not - while a 64-bit OS and applications will be somewhat faster (and a whole lot more efficient for large-memory apps), the real gain in in processor efficiency. The G5 can execute four instructions per clock cycle (in a perfect world), the G4 is limited to three. In addition, the G5 has a much faster frontside bus, as well as considerably more I/O and memory bandwidth.
Second, MP can be quite a bit more efficient than you state. I've been working with MP systems for 10 years now, and I've seen some systems that hit linear (and in one weird case, super-linear) performance curves. A *lot* has to do with the OS and application, and how well they thread. Typically, on the G5, I'm seeing about 80-85% improvement by turning the second processor on.
Finally, have you actually worked on a G5? I'm lucky, I have - both at work and at home I have dual G5's. At work, we've seen some truly impressive results on a heavy mix of I/O and CPU. Results that put a dual G4 to shame. (Not to mention a dual 2.4GHz P4 Xeon and a Sun 280R.) As for home, I work on a variety of A/V applications and I am impressed with the results. (To be honest, I used the systems at work, got hooked, and then lucked out and found a way to afford a dual G5.)
So, artificial benchmarks aside, I have no problem stating that the G5 is the fastest desktop I've ever had the pleasure to use.
Most Recent Replies: View All 2 Replies
- Ahem. Get the facts straight. (1 replies)
For the G5 clueless - Version: 1.1.3, 10/8/2003 06:26PM PST
(3 of 25 users found this comment useful)
blueskymining
Xbench is a good benchmark under the tests that I've done with it on a very wide spectrum of Macs from G3's to G5's, single and dual processor. Jeez, it's free folks.
For the Clueless G5 owners, your G5's AREN'T that much faster than a G4 if at all, because G5's ONLY show a performance increase on true 64-bit apps, which there aren't any to benchmark with. Using your current versions of Photoshop, Premier or After Effects, etc., you've spent a lot of money for hardware that offers NO advantage to the app! So, what is Cinebench supposed to do? Benchmark for something that's not available? The G5 is that mythical "I have some swamp property to sell you" by Apple, UNTIL 64-bit apps are prevalent.
The same goes for AMD's new 64-bit processor, UNLESS you have the $40,000 video editing software that was written for it, or you travel to the future and bring back Windows 64-bit! One caviat, versions of Linux for 64-bit does see the advantages.
Ps. Dual processors DON'T make things "twice as fast"; try about a 18 -25% additional processing throughput. Look it up for yourself.
For the Clueless G5 owners, your G5's AREN'T that much faster than a G4 if at all, because G5's ONLY show a performance increase on true 64-bit apps, which there aren't any to benchmark with. Using your current versions of Photoshop, Premier or After Effects, etc., you've spent a lot of money for hardware that offers NO advantage to the app! So, what is Cinebench supposed to do? Benchmark for something that's not available? The G5 is that mythical "I have some swamp property to sell you" by Apple, UNTIL 64-bit apps are prevalent.
The same goes for AMD's new 64-bit processor, UNLESS you have the $40,000 video editing software that was written for it, or you travel to the future and bring back Windows 64-bit! One caviat, versions of Linux for 64-bit does see the advantages.
Ps. Dual processors DON'T make things "twice as fast"; try about a 18 -25% additional processing throughput. Look it up for yourself.
Strange results. - Version: 1.1.3, 10/8/2003 01:12AM PST
(0 of 1 users found this comment useful)
rsanderm
I have a hard time believing that my 1.8 GHz G5 scores only slightly higher overall compared to the base dual 800. I have an even harder time believing that it scored lower on the hard drive and some processor tests.
Most Recent Replies: View All 1 Replies
there's a server issue.
Maybe I'll donate when they get it up and running.