| View previous topic :: View next topic |
| Author |
Message |
aquaman
Veteran Member
Joined: 18 Apr 2005
Posts: 807
|
 Posted: Fri Mar 03, 2006 6:36 pm Post subject: |
 |
|
| bsnoel wrote: | | It would also be nice to see the new 17" iMac Core Duo included in the benchmarks too. This would provide us a reasonably close comparison between the Intel GMA950 VS a ATI Radeon X1600. Obviously the ATI would win, but it would helpful for those concerned about the GMA950. |
I would like to see a few scores comparing the top of the line mini and 17" iMac too. I'm pretty sure the revision B Intel Mac mini is going to be my next Mac. I'm pretty impressed with all the new features, and really jonesing for an Apple Remote and Front Row in the worst way. |
|
| Back to top |
|
 |
devo
Veteran Member
Joined: 23 Jan 2005
Posts: 5274
Location: Dunwoody, GA
|
| Posted: Fri Mar 03, 2006 7:41 pm Post subject: Re: World of Warcraft? |
|
|
| Yvan256 wrote: | | What I'd like to see is Handbrake timings (say, with H.264/AAC, Quality 60, DVD resolution settings) and World of Warcraft FPS comparisons. |
I've ripped two movies with my 1.66. Both were at 320x240 for my 5G iPod. H.264/AAC was averaging 44 fps and MPEG-4 averaged 62 fps. It's waaaaay faster than my G4 mini! You should be happy.
I posted my scores last night. I've noticed the scores are all over the board. Wonder why?
http://db.xbench.com/merge.xhtml?doc1=158966&doc2=134443 |
|
| Back to top |
|
|
bsnoel
Senior Member
Joined: 04 Jun 2005
Posts: 355
Location: USA
|
| Posted: Fri Mar 03, 2006 9:01 pm Post subject: Re: World of Warcraft? |
|
|
Xbench scores are affected by lunar gravitational forces, solar winds, Gama rays, and coastal tides. It depends on both when you benchmark and where you live.  No, seriously Xbench scores do vary somewhat even when run on the same machine. I would suggest turning the User interface test off when running XBench on an Intel Mac. This test appears to be very biased toword PPC CPUs.
_________________
Mac Pro Quad 2.66GHz, PowerMac G5 2.3 GHz, iMac Core Duo 17", MacBook Pro 2GHz, MacBook Pro 2.2GHz, Mac mini Core 2 Duo 1.83GHz. Multi-K9 Security System. No false alarms, just lots of sharp teeth.
www.grweather.com |
|
| Back to top |
|
|
devo
Veteran Member
Joined: 23 Jan 2005
Posts: 5274
Location: Dunwoody, GA
|
| Posted: Fri Mar 03, 2006 9:42 pm Post subject: Re: World of Warcraft? |
|
|
| bsnoel wrote: | | I would suggest turning the User interface test off when running XBench on an Intel Mac. This test appears to be very biased toword PPC CPUs. |
Test with UI turned off.
http://db.xbench.com/merge.xhtml?doc1=159210 |
|
| Back to top |
|
|
bsnoel
Senior Member
Joined: 04 Jun 2005
Posts: 355
Location: USA
|
| Posted: Fri Mar 03, 2006 9:57 pm Post subject: Re: World of Warcraft? |
|
|
| devo wrote: | | bsnoel wrote: | | I would suggest turning the User interface test off when running XBench on an Intel Mac. This test appears to be very biased toword PPC CPUs. |
Test with UI turned off.
http://db.xbench.com/merge.xhtml?doc1=159210 |
That is better, now if there was only a way to turn the vecLIB tests off. That is the other component that skews the tests compairsons.
_________________
Mac Pro Quad 2.66GHz, PowerMac G5 2.3 GHz, iMac Core Duo 17", MacBook Pro 2GHz, MacBook Pro 2.2GHz, Mac mini Core 2 Duo 1.83GHz. Multi-K9 Security System. No false alarms, just lots of sharp teeth.
www.grweather.com |
|
| Back to top |
|
|
devo
Veteran Member
Joined: 23 Jan 2005
Posts: 5274
Location: Dunwoody, GA
|
| Posted: Fri Mar 03, 2006 10:18 pm Post subject: Re: World of Warcraft? |
|
|
| bsnoel wrote: |
That is better, now if there was only a way to turn the vecLIB tests off |
Why? AltiVec? |
|
| Back to top |
|
|
Renec
New Member
Joined: 04 Mar 2006
Posts: 1
|
| Posted: Sat Mar 04, 2006 12:06 am Post subject: Intel Mac Mini and Quartz |
|
|
| So does the new mini do quartz effects on the dashboard and such? I just ordered a mac mini and I hope I shouldn't have got an iMac instead. |
|
| Back to top |
|
|
curt
Veteran Member
Joined: 17 Mar 2005
Posts: 1411
Location: Oregon
|
| Posted: Sat Mar 04, 2006 12:46 am Post subject: Re: Intel Mac Mini and Quartz |
|
|
| Renec wrote: | | So does the new mini do quartz effects on the dashboard and such? I just ordered a mac mini and I hope I shouldn't have got an iMac instead. |
Yes, if you mean the ripple effect. The Mac mini can do Core Image now. 
_________________
1.42GHz Mac mini
1.66GHz Intel Mac mini
2.0GHz 2009 Mac mini
46" Sony LCD
50" Sony KDS-50A2000
EyeTV 250 Plus
My Setup |
|
| Back to top |
|
|
atty
New Member
Joined: 02 Mar 2006
Posts: 3
Location: UK
|
| Posted: Sat Mar 04, 2006 10:33 am Post subject: Re: Intel Mac Mini and Quartz |
|
|
| Renec wrote: | | So does the new mini do quartz effects on the dashboard and such? I just ordered a mac mini and I hope I shouldn't have got an iMac instead. |
Hi,
Yeah, the new Mini's and the GMA950 graphics chipsets have full core image support, so stuff like the ripple effects and other openGL accelerated things should run beautifully.  |
|
| Back to top |
|
|
bsnoel
Senior Member
Joined: 04 Jun 2005
Posts: 355
Location: USA
|
| Posted: Sat Mar 04, 2006 12:37 pm Post subject: Re: World of Warcraft? |
|
|
| devo wrote: | | bsnoel wrote: |
That is better, now if there was only a way to turn the vecLIB tests off |
Why? AltiVec? |
Both Intel and PPC have their own implementations of vector processing or SIMD (Single Instruction Multiple Data) Apple also refers to vector processing as AltiVec or Velocity Engine, while Intel calls it SSE. Both Intel and PPC are similar in function, but there are differences that make them tricky to benchmark. From what I have seen the vecLib in Xbench seems to favor PPC CPUs. It is possible that AltiVec is superior to SSE, but I suspect that what we are seeing is partly due to the vecLib test itself.
_________________
Mac Pro Quad 2.66GHz, PowerMac G5 2.3 GHz, iMac Core Duo 17", MacBook Pro 2GHz, MacBook Pro 2.2GHz, Mac mini Core 2 Duo 1.83GHz. Multi-K9 Security System. No false alarms, just lots of sharp teeth.
www.grweather.com |
|
| Back to top |
|
|
devo
Veteran Member
Joined: 23 Jan 2005
Posts: 5274
Location: Dunwoody, GA
|
| Posted: Sat Mar 04, 2006 1:02 pm Post subject: Re: World of Warcraft? |
|
|
| bsnoel wrote: | | Both Intel and PPC have their own implementations of vector processing or SIMD (Single Instruction Multiple Data) Apple also refers to vector processing as AltiVec or Velocity Engine, while Intel calls it SSE. Both Intel and PPC are similar in function, but there are differences that make them tricky to benchmark. From what I have seen the vecLib in Xbench seems to favor PPC CPUs. It is possible that AltiVec is superior to SSE, but I suspect that what we are seeing is partly due to the vecLib test itself. |
Half of that stuff I beyond me, but I sort of understand where you're heading. I also think it's possible that Xbench may need a little more tweaking for x86 Macs. What type of scores are you getting with your 17" iMac? |
|
| Back to top |
|
|
ejohnson
New Member
Joined: 05 Mar 2006
Posts: 2
|
| Posted: Sun Mar 05, 2006 3:49 pm Post subject: |
|
|
So what kind of preformance can i expect out of the new low end model that i cant get out of my silent upgraded g4. i mean, all the specs are the same except the new mini takes 2gb ram, has a smaller harddrive, and has 2 more usb ports. mine was 599$, came with 1.5ghz, 64 meg video ram, 80 gig 5400rpm harddrive. the new mini is the same price and seems to me that the 2 extra usb ports and upgradable 2 gigs ram isnt enough to get me to shell out another 599 or to even consider getting on, look on ebay, now that the new mac came out you can buy up old ones for cheap.
this isnt a bash mac rant but rather they should of thought their bottom of the line mac out better. I love my mac mini midline model, and i havnt upgraded anything on it yet.
if anything in that pile of words makes sense to anyone then reply otherwise just know i am very tired so things wont make sence. |
|
| Back to top |
|
|
James T Kirk
Veteran Member
Joined: 23 Aug 2005
Posts: 757
Location: The Netherlands, Europe
|
| Posted: Sun Mar 05, 2006 7:05 pm Post subject: Re: Core Solo Xbench results |
|
|
| ongnoi wrote: | I had the chance last night to run a test with my 1.5GHz CS.
http://db.xbench.com/merge.xhtml?doc1=159011
The computer overall is very snappy, open UB apps in one or two bounces mostly. I was also able to play a few HD clips. 720p played flawlessly fullscreen, not so true with 1080p though. Boot time is under 30 sec. |
I should have waited buying my 1.5 G4. Best overal result I had was 45.03, even when I had 1 gig stick I did not go over 45.. Should I sell my mini?  The new core solo is only 20 euro's more in my country than the old one was! 
_________________
___________________Jim©
Now on Hackintosh in G5 case! |
|
| Back to top |
|
|
resuna
Member
Joined: 15 May 2005
Posts: 215
|
| Posted: Mon Mar 06, 2006 11:52 am Post subject: |
|
|
"Shading (OpenGL Hardware Lighting) 518 543"
That's the vertex shader on the 9200 running as fast as the software emulation of the vertex shader on the Core Duo.
I assume the vertex shader emulation is single-threaded, so that's only one of the cores running that... so it's *possible* that a single-threaded application on a core duo will be able to get comparable OpenGL performance to the Mac mini, but anything that's doing 3d multithreaded will basically lose half its CPU to software OpenGL.
That means the new machine will probably not be any worse for 3d than the old one. It's not going to be any "4x faster" though. |
|
| Back to top |
|
|
resuna
Member
Joined: 15 May 2005
Posts: 215
|
| Posted: Tue Mar 07, 2006 12:35 pm Post subject: Question for someone with a Mac mini Duo... |
|
|
Does the vertex shader emulation run on the same core as the application (that is, it's in the OpenGL library) or is it a separate process/thread that can be sceduled on the other core?
That is, if you run one copy of a single-threaded application that's using the vertex shader and also doing CPU-intensive work, what's the load on the two CPUs? |
|
| Back to top |
|
|
|
FAQ 
Search 
Memberlist 
Photo Gallery
Register
Profile 
Log in to check your private messages 
Log in 
