• Welcome to the Off-Topic/Schweb's Lounge

    In addition to the Mac-Forums Community Guidelines, there are a few things you should pay attention to while in The Lounge.

    Lounge Rules
    • If your post belongs in a different forum, please post it there.
    • While this area is for off-topic conversations, that doesn't mean that every conversation will be permitted. The moderators will, at their sole discretion, close or delete any threads which do not serve a beneficial purpose to the community.

    Understand that while The Lounge is here as a place to relax and discuss random topics, that doesn't mean we will allow any topic. Topics which are inflammatory, hurtful, or otherwise clash with our Mac-Forums Community Guidelines will be removed.

How does your Mac stack up - Geekbench thread.

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
I wrote it for my use but I can certainly fine tune it such that it works for everyone else. I'll report back.
 
Joined
Feb 25, 2009
Messages
2,112
Reaction score
71
Points
48
Your Mac's Specs
Late 2013 rMBP, i7, 750m gpu, OSX versions 10.9.3, 10.10
Because geekbewnch only takes into account processor and RAM the places where intel excels. It's also not compatible with the newer Power Architecture so it can't show where x86-64 fails. Geekbench shows Mac's in the best light no matter whether it's Intels best light or Apples .

I guess I just don't feel that is a complete test of a system as graphical power directly plays into the overall functionality of a modern day system. So much of the GUIs now are enhanced by the GPU that only looking at the CPU and RAM is only looking at a smaller picture of the total environment. A simple example would be two systems side by side, one with a dedicated GPU and one without, but everything else the same - Geekbench would show them as basically the same system in terms of performance but that would be inaccurate because most likely the one with a dedicated GPU would outperform the non dedicated on many tasks that incorporate the GPU as well as the CPU.
 
Joined
Oct 31, 2011
Messages
283
Reaction score
4
Points
18
I guess I just don't feel that is a complete test of a system as graphical power directly plays into the overall functionality of a modern day system. So much of the GUIs now are enhanced by the GPU that only looking at the CPU and RAM is only looking at a smaller picture of the total environment. A simple example would be two systems side by side, one with a dedicated GPU and one without, but everything else the same - Geekbench would show them as basically the same system in terms of performance but that would be inaccurate because most likely the one with a dedicated GPU would outperform the non dedicated on many tasks that incorporate the GPU as well as the CPU.

More my concern is IO speed only because when have no really good way of cross platform benching between Nvidia and ATI. Cinebench is nice but doesn't let CUDA come into play. and back an fourth. I'm still not a Geekbench lover, it's just not a system test like xbench, but x bench has proven to unreliable and nor has it been updated in half a decade.
 

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
So much of the GUIs now are enhanced by the GPU that only looking at the CPU and RAM is only looking at a smaller picture of the total environment.
I really wish this trend towards GPU accelerating everything would stop because it went overboard (and continues to do so). For example, why do I need a GPU accelerated web browser? I don't. Do I need a fully accelerated desktop environment (I'm looking at you Windows and *nix)? No. I'm willing to guess that 90% of what is accelerated doesn't need to be and does little but showcase programming skill. This is above and beyond the major problem with GPU acceleration - it degrades the performance for anyone who doesn't have a decent GPU.

During the Opera 12 release cycle, they were testing GPU acceleration which, as a company that officially supports four platforms, is a tricky endeavour. During the development process, I tried the builds with it on and it was a disaster with no discernible improvement in performance. Thankfully, it isn't enabled by default. You've got me thinking though what kind of performance difference I might see with it on and off. Hmmm...off to testing I go!
 

dtravis7


Retired Staff
Joined
Jan 4, 2005
Messages
30,133
Reaction score
703
Points
113
Location
Modesto, Ca.
Your Mac's Specs
MacMini M-1 MacOS Monterey, iMac 2010 27"Quad I7 , MBPLate2011, iPad Pro10.5", iPhoneSE
Agreed Van and well said.
 
Joined
Apr 2, 2012
Messages
106
Reaction score
1
Points
18
Location
Ohio
Your Mac's Specs
Macbook Pro 2.8 GHz I7 4GB mem version 10.10 Yosemite // Ipad3 // iPhone 4S
Im not real sure what these numbers mean but is this decent?


ScreenShot2012-07-15at81642AM.png
 

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
Im not real sure what these numbers mean but is this decent?
By themselves, absolutely nothing. They serve no purpose other than a comparative one.
 
Joined
Feb 25, 2009
Messages
2,112
Reaction score
71
Points
48
Your Mac's Specs
Late 2013 rMBP, i7, 750m gpu, OSX versions 10.9.3, 10.10
I really wish this trend towards GPU accelerating everything would stop because it went overboard (and continues to do so). For example, why do I need a GPU accelerated web browser? I don't. Do I need a fully accelerated desktop environment (I'm looking at you Windows and *nix)? No. I'm willing to guess that 90% of what is accelerated doesn't need to be and does little but showcase programming skill. This is above and beyond the major problem with GPU acceleration - it degrades the performance for anyone who doesn't have a decent GPU.

During the Opera 12 release cycle, they were testing GPU acceleration which, as a company that officially supports four platforms, is a tricky endeavour. During the development process, I tried the builds with it on and it was a disaster with no discernible improvement in performance. Thankfully, it isn't enabled by default. You've got me thinking though what kind of performance difference I might see with it on and off. Hmmm...off to testing I go!

Sometimes it pays off though - for those doing work with video or 3D the GPU is optimized for many tasks on that level where the CPU is not. If I can transcode something or render something in 30 minutes that would take my CPU 2+ hours to do, or provide me better real time display (for example, decoding H.264) I'd rather that be put on the GPU where the GPU can shine - and I'd want to see a comparative of how GPU's perform in their capabilities.

Now, I'm not saying that in some cases GPU acceleration has gone overboard. A lot of it is eye candy (like semi-transparent windows or 3D desktops) that would bog down a CPU but don't bog down a GPU near as much (simple example is Aero desktop which requires a certain (albeit basic) level of GPU capability to function). Personally I don't need a lot of that stuff (I turn Aero off on my Win7 bootcamp) but many people love those effects so what can you do?

JAVA and Flash both take advantage of the GPU for certain tasks. All of these things are still important to look at when considering the total capability of a system, so while IO Speed, RAM speed, CPU speed, etc. are all still core of a system, GPU does play a significant part in a modern day operating system and should not be overlooked while producing comparative numbers for systems.

That said, I also feel that looking at one number for the whole system is misleading and numbers for separate categories should be considered when evaluating (seeing a system ranked at 9500 means nothing if the score is all CPU and you need GPU or all GPU when your tasks require more CPU, etc.)

In regards to something like Opera - I couldn't see how it would improve much to have GPU acceleration EXCEPT for when decoding something like HTML5 video - general use I'd imagine would have no effect unless they are doing some sort of bizarre 3D rendering within the application.
 

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
I have no problem believing that the GPU can do heavy lifting. If I remember correctly, there are suggestions that GPU speeds, in certain cases, can be magnitudes faster than a CPU and in those instances, it's worth the effort. However, that's not my problem. My issue is with the "acceleration of everything." Certain features truly are best meant for certain tasks.

I forgot to do that test with the acceleration on and off. I'll try that now.
 
Joined
Feb 25, 2009
Messages
2,112
Reaction score
71
Points
48
Your Mac's Specs
Late 2013 rMBP, i7, 750m gpu, OSX versions 10.9.3, 10.10
To look at the numbers and debate on quality you need to compare them to other numbers - look on Novabench for similar system specs and compare all of the numbers. If you were to look at - for example - my MBA:

novabenchair.png


you could see that our RAM ranks very close to each other, your CPU is much more powerful, your GPU is a little more powerful (same model, might be a fluke but since the HD3000 is tied to the CPU yours may be clocked higher, I don't know how Apple handles that) but my hardware benchmark is much higher than yours and the overall comparison is not far off between the two machines (about 50 points) but where those points are might be important to you, and also seeing a higher hardware test may cause you to think about replacing your hard drive with an SSD. You never know.

Personally, I usually just use base benchmark tests to make sure my machine is operating within the region of other machines of similar speeds - if I find that mine is significantly below comparably spec'd machines (or significantly higher) I might wonder why and investigate what I might either improve or if something is possibly not operating right (I had a GPU in one system rank really low which was odd compared to other systems with the same GPU - after more experimentation I found the GPU was actually faulty and had to be replaced)

In the long run, real world performance is your final judgement. Benchmarks give you a snapshot of what the system is based upon certain criteria that can judge for general use for comparison against similar machines.

Im not real sure what these numbers mean but is this decent?


ScreenShot2012-07-15at81642AM.png
 

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
Okay, I just did Peacekeeper (because I knew it did video and graphics tests) in Opera with GPU acceleration on and off. With it on, it achieved a score of 2691. With it off, it had a score of 2698. I'll see if I can find a large hi-def video stream and test that as well.
 
Joined
Feb 25, 2009
Messages
2,112
Reaction score
71
Points
48
Your Mac's Specs
Late 2013 rMBP, i7, 750m gpu, OSX versions 10.9.3, 10.10
I have no problem believing that the GPU can do heavy lifting. If I remember correctly, there are suggestions that GPU speeds, in certain cases, can be magnitudes faster than a CPU and in those instances, it's worth the effort. However, that's not my problem. My issue is with the "acceleration of everything." Certain features truly are best meant for certain tasks.

I forgot to do that test with the acceleration on and off. I'll try that now.

I do agree with you there that acceleration just for sake of claiming something has been accelerated by some type of hardware is a waste unless the benefit truly outweighs the costs and increased requirements of the acceleration.

If accelerating my desktop gains the average person an equivalent of 1fps in terms of smoothness of GUI operation then it's a complete waste of development hours and a worthless feature that has only been put in for proof it can be done and to be able to sell a new product even tho it won't really help. If on the other hand it changes a choppy poor experience into a smooth enjoyable experience I'm all for it (like improving a desktop rendering to go from the equivalent of 20fps to a solid 60fps) especially when they do it without requiring new higher cost equipment.

On those levels, I completely agree with you Vansmith
 
Joined
Feb 25, 2009
Messages
2,112
Reaction score
71
Points
48
Your Mac's Specs
Late 2013 rMBP, i7, 750m gpu, OSX versions 10.9.3, 10.10
Okay, I just did Peacekeeper (because I knew it did video and graphics tests) in Opera with GPU acceleration on and off. With it on, it achieved a score of 2691. With it off, it had a score of 2698. I'll see if I can find a large hi-def video stream and test that as well.

ROFL, obviously one of those times that GPU acceleration was a complete waste of resources ;)
 

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
If accelerating my desktop gains the average person an equivalent of 1fps in terms of smoothness of GUI operation then it's a complete waste of development hours and a worthless feature that has only been put in for proof it can be done and to be able to sell a new product even tho it won't really help. If on the other hand it changes a choppy poor experience into a smooth enjoyable experience I'm all for it (like improving a desktop rendering to go from the equivalent of 20fps to a solid 60fps) especially when they do it without requiring new higher cost equipment.
I'm with you on that one. If there are significant performance advantages, then it's obvious that there's a point and I say "go for it." Otherwise, it's utterly useless.

ROFL, obviously one of those times that GPU acceleration was a complete waste of resources ;)
You like that? Haha. I don't have an impressive GPU or anything (HD 4000) but I was expecting something!
 
Joined
Apr 2, 2012
Messages
106
Reaction score
1
Points
18
Location
Ohio
Your Mac's Specs
Macbook Pro 2.8 GHz I7 4GB mem version 10.10 Yosemite // Ipad3 // iPhone 4S
Thank you for that explanation Nethfel!
 
Joined
Jul 18, 2012
Messages
14
Reaction score
0
Points
1
Sorry, I don't know how to capture a screenshot yet. Here is cut/paste:

MacBook Pro (15-inch Mid 2012)
Section Description Score Geekbench Score
Geekbench 2.3.4 Tryout for Mac OS X x86 (32-bit)
Integer Processor integer performance 9735 11877
Floating Point Processor floating point performance 18064
Memory Memory performance 7128
Stream Memory bandwidth performance 7225
Result Information
Upload Date July 22 2012 06:46 PM
Views 1


System Information
MacBook Pro (15-inch Mid 2012)
Operating System Mac OS X 10.7.4 (Build 11E2620)
Model MacBook Pro (15-inch Mid 2012)
Processor Intel Core i7-3720QM @ 2.60 GHz
1 processor, 4 cores, 8 threads
Processor ID GenuineIntel Family 6 Model 58 Stepping 9
L1 Instruction Cache 32 KB x 4
L1 Data Cache 32 KB x 4
L2 Cache 256 KB x 4
L3 Cache 6144 KB
Motherboard Apple Inc. Mac-4B7AC7E43945597E MacBookPro9,1
BIOS Apple Inc. MBP91.88Z.00D3.B02.1203281326
Memory 8192 MB 1600 MHz DDR3
 

vansmith

Senior Member
Joined
Oct 19, 2008
Messages
19,924
Reaction score
559
Points
113
Location
Queensland
Your Mac's Specs
Mini (2014, 2018, 2020), MBA (2020), iPad Pro (2018), iPhone 13 Pro Max, Watch (S6)
Sorry, I don't know how to capture a screenshot yet.
Command-Shift-3 to get the whole screen, Command-Shift-4 to get a selected area or Command-Shift-4 then spacebar to get a specific window.
 
Joined
Jul 23, 2012
Messages
15
Reaction score
0
Points
1
MacBook Pro (15-inch Early 2011)
Section Description Score Geekbench Score
Geekbench 2.3.4 Tryout for Mac OS X x86 (32-bit)
Integer Processor integer performance 8554 10401
Floating Point Processor floating point performance 15932
Memory Memory performance 6059
Stream Memory bandwidth performance 6194
Result Information
Upload Date July 28 2012 03:40 AM
Views 1
System Information
MacBook Pro (15-inch Early 2011)
Operating System Mac OS X 10.7.4 (Build 11E53)
Model MacBook Pro (15-inch Early 2011)
Processor Intel Core i7-2820QM @ 2.30 GHz
1 processor, 4 cores, 8 threads
Processor ID GenuineIntel Family 6 Model 42 Stepping 7
L1 Instruction Cache 32 KB x 4
L1 Data Cache 32 KB x 4
L2 Cache 256 KB x 4
L3 Cache 8192 KB
Motherboard Apple Inc. Mac-94245A3940C91C80 MacBookPro8,2
BIOS Apple Inc. MBP81.88Z.0047.B0E.1104221557
Memory 16384 MB 1333 MHz DDR3
 

Shop Amazon


Shop for your Apple, Mac, iPhone and other computer products on Amazon.
We are a participant in the Amazon Services LLC Associates Program, an affiliate program designed to provide a means for us to earn fees by linking to Amazon and affiliated sites.
Top