Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2009 Mac Pro in hand - unboxing etc.
- Thread starter WonderSausage
- Start date
- Sort by reaction score
Awesome, so now after 6 months of hearing "don't buy the 2008 mac pro, the 2009 will be out any day now and be MUCH better",
the 2009 finally arrives and isn;t really better,
but wait, salvation is around the corner
"just wait for a new OS to be released (any day now, of course) and the 2009 will really fly"
with an extra helping of "future applications like CS6 will really take advantage of the 2009 model"
Do you listen to yourself when you say this stuff?
Snow Leopard helps developers get to the point that GeekBench is already at. It simply makes it easier to have an application multithread correctly via GC
Think of it this way, in the context of languages:
For the sake of argument, let's pretend everyone in the world grows up speaking English (single-thread code). To be a real badass, you have to learn Russian (good multithreaded code). For most people, learning new languages is a huge pain in the ass, so they just learn the bare minimums to get by in an unfamiliar situation.
What Grand Central does is tag along like your little translator. Sure, you have to do a few minutes of work to learn how to make the translator work, but it'll handle all your communication with Russian (multi-core or multi-processor) computers.
However, it won't do you a damned bit of good if you're already fluent in Russian.
Make sense?
Here are some of the numbers in histogram format. I'm not sure how meaningful the data is at this time but the 2.93 is certainly fast and way more expen$ive too
Attachments
You don't understand what Grand Central is if you think that.
Grand Central does not make it possible to use multiple cores. That can already be done. What Grand Central does is make accessing multiple cores simpler and more efficient. It is going to greatly improve the organizing, assigning, and scheduling of tasks amongst available cores.
Add OpenCL to the mix and parallel processing on the Mac gets pretty darned exciting.
S-
Did you even read what I just wrote? FFS… I tried to use a non-technical analogy to convey the idea of it making life easier, but it won't be significantly faster than someone writing out the code the long way if they already know how to do it in the context of GeekBench
I'm not claiming it's a perfect comparison, but it conveys the general concept.
People who dissapointed by new pros single threading performance! Just think: with hyperthreading turned on every single thread use only half of single physical core.
Just turn hyperthreading off (is it possible on mac pro? If no, just wait for snow leo) and single threading performance will grow twice. And this new doubled score will be true single threading score of new pros.
Just turn hyperthreading off (is it possible on mac pro? If no, just wait for snow leo) and single threading performance will grow twice. And this new doubled score will be true single threading score of new pros.
What makes you think that by turning off HT will double the score of single thread benchmark...?
My Intel Pentium 4 experience.What makes you think that by turning off HT will double the score of single thread benchmark...?
Why don't you try benchmarking your P4 with HT turned off, then again with HT turned on, see if it doubles the performance...?
And try real-world performance test.... not just synthetic benchmarking...
Can someone with logic pro and the new 2,26 run some benchmarks with http://www.evan.se/logicprobenchmark/ and post how many tracks they where able to run?
-Thanks
P.S. If someone with the previous 3,2 can do the same it would be great too. According to the post on gearslutz a 2,8 can do up to 50 tracks.
-Thanks
P.S. If someone with the previous 3,2 can do the same it would be great too. According to the post on gearslutz a 2,8 can do up to 50 tracks.
Not sure if this makes any sense. As far as I'm aware Logic can run as many tracks as it can handle. It really depends on the CPU draw from each track in regards to soft synths and effects. In theory you could probably have 100 tracks if the soft synth or effect pulled very little CPU power. The number of tracks can't really equate to any sort of benchmark. You'd have to be running exactly the same project on each computer to gauge the CPU usage like the Ableton Live performance test (see their forums).
Just a thought
Im sorry Im a little bit confussed. I read in one post
Processor : Dual Nehalem Xeon
MHz : 2933
Number of CPUs : 16
Operating System : OS X 32 BIT 10.5.6
16 core? I read this computer has 2 quad core. Please explain me that.
Thanks!
Multithreading.
This will be the definitive scores for many. Including me. Depending on this I might bite the bullet and get the 2,26. If not I'll keep on waiting to see if there is an improvement with snow and/or a logic update. Other wise just wait on a used 2,8 deal...
Seems like were in the same boat here. Problem is thou that Ive ordered a 2.8, but if 09 is ALOT faster it might be worth the extra $! Maybe even a Quad is faster then a 2,8 in Logic and That one coast exactly the same as I have to pay for the 2.8.
What to do what do do! MORE: The Evan Logic test is just this. The same Logic setup, just for compasion. I guess it doesnt show how "fast" the computer is, but still its serves as a compasion between diffrent systems.
// Johan
Seems like were in the same boat here. Problem is thou that Ive ordered a 2.8, but if 09 is ALOT faster it might be worth the extra $! Maybe even a Quad is faster then a 2,8 in Logic and That one coast exactly the same as I have to pay for the 2.8.
MORE: The Evan Logic test is just this. The same Logic setup, just for compasion. I guess it doesnt show how "fast" the computer is, but still its serves as a compasion between diffrent systems.
// Johan
I doubt the quad is faster than the dual quad. Logic is multi-threaded but I'm not sure how it deals with virtual threads.
Buffersize.. ? scrub response, solo/mute mode, buffer range..
its a stupid benchmark
btw, 8x2.8 is great for audio.
+logic has a crappy multithread implementation..
right now im maxing 3 cores, 3 are at 50-60%, 1 is at 75%, and 1 is IDLING.
+update, 50tracks, one core idling. (other max)
+update 55, one core 10%.
i have to close itunes&safari