Toggle sidebar Toggle sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Is the 2.26GHz 8-Core really THAT slow??

WardC

WardC

macrumors 68030
Original poster
I have read reports on here about the 2.26GHz 8-core Mac Pro being reportedly slow...but is it really THAT slow?? I mean, I am in the market for a Mac Pro right now and I don't really know if I will wait until the 6-core machines come out. How slow does the 2.26GHz feel in the Mac OS when navigating, opening apps? Is there a big difference between the 2.66GHz and the 2.93GHz? Also comparably how is the single core 3.33GHz vs the 8-core 2.26GHz? Any help would be appreciated. Thanks!!
 
I can't speak to a lot comparative experience with the other models, but, no my 2.26GHz 8-core Mac Pro does not feel slow at all (as an aside, an X-25M works wonders on perceived speed).

I am hankering for a 5870, though. 😉
 
Well.. 2.26 vs 2.93... 670mhz difference... It was only just around a decade ago when people were working on Pentium III machines with a ~600mhz clock rate.

But I digress....

When it comes to opening apps and navigating menus, you're really not going to tell the difference. A faster processor will certainly make these things faster, but it's fractions of a second difference. And seriously, who measures performance based on opening an app or a menu. You want to look at hard number crunching instead.

Lets take it from a simple empirical view in a "perfect" scenario...
2.26 billion clock cycles vs 2.93 billion clock cycles.... 670 million clock cycle difference.
Lets say you're trying to ray-trace an image (easiest scenario i can think of right now) which means you're going from pixel to object, rather than object to pixel. Let's say it's fairly large image at 1680x1024 image - it has 1720320 pixels. Now lets say rendering each pixel on average involves 300 operations in the program - it's not too complex. Each operation on average takes 10 clock cycles.

10x300x1720320 = 5160960000 clock cycles = ~5.16billion clock cycles.

What does this mean? Well, simply, on a 2.26Ghz machine, it would take 2.28 seconds to render, and 1.76 seconds to render on a 2.93Ghz machine.

This is really a simplistic way of looking at it in a "perfect" scenario - i.e. no memory bottlenecks and interference from other apps - also running in a non-threaded mode. But the idea applies to anything else really.

And while the example is small, and increase in the clock cycles in the example will show an (unsurprisingly) linear increase in the difference between the two processors.

The comparison also stands when comparing 3.33Ghz and 2.26Ghz in non-threaded applications. But its a different story with threaded stuff. An 8-core 2.26Ghz machine will be faster than a Quad-core 3.33Ghz machine if all cores are being utilised either by the program you're using or the sheer amount of individual applications you're using concurrently (this of which would require some serious multitasking 😉).
 
Vylen said:
Well.. 2.26 vs 2.93... 670mhz difference... It was only just around a decade ago when proplr were working on Pentium III machines with a ~600mhz clock rate.

When it comes to opening apps and navigating menus, you're really not going to tell the difference. A faster processor will certainly make these things faster, but it's fractions of a second difference. And seriously, who measures performance based on opening an app or a menu. You want to look at hard number crunching instead.

Lets take it from a simple empirical view in a "perfect" scenario...
2.26 billion clock cycles vs 2.93 billion clock cycles.... 670 million clock cycle difference.
Lets say you're trying to ray-trace an image (easiest scenario i can think of right now) which means you're going from pixel to object, rather than object to pixel. Let's say it's fairly image at 1680x1024 image - it has 1720320 pixels. Now lets say rendering each pixel on average involves 300 operations in the program - it's not too complex. Each operation on average takes 10 clock cycles.

10x300x1720320 = 5160960000 clock cycles = ~5.16billion clock cycles.

What does this mean? Well, simply, on a 2.26Ghz machine, it would take 2.28 seconds to render, and 1.76 seconds to render on a 2.93Ghz machine.

This is really a simplistic way of looking at it in a "perfect" scenario - i.e. no memory bottlenecks and interference from other apps - also running in a non-threaded mode. But the idea applies to anything else really.

And while the example is small, and increase in the clock cycles in the example will show an (unsurprisingly) linear increase in the difference between the two processors.

The comparison also stands when comparing 3.33Ghz and 2.26Ghz in non-threaded applications. But its a different story with threaded stuff. An 8-core 2.26Ghz machine will be faster than a Quad-core 3.33Ghz machine if all cores are being utilised either by the program you're using or the sheer amount of individual applications you're using concurrently (this of which would require some serious multitasking 😉).
Click to expand...

I don;t think you should compare stuff like that (Pentium III vs a Core ix). Different architectures and cores. What we need is Cow2Beef.exe picture to show you that 2.26GHz is plenty fast enough, specially with 8 cores.
 
I wasn't comparing... I was just "saying" ... cause you know.. these days we take it for granted that processors these days are in the 2Ghz range...

P.S I should really proof read my posts before posting, instead of editing after the fact... People who quote my old typo'd stuff... is... well, old and typo'd 😛
 
Vylen said:
I wasn't comparing... I was just "saying" ... cause you know.. these days we take it for granted that processors these days are in the 2Ghz range...
Click to expand...

True ture. However, we still need Cow2Beef.exe to explain why his slower 2.26GHz cores won't matter much.

Edit - Cow2Beef.exe is here...

Click to enlarge

1199200826519.png


Click to enlarge
 
Damn pictures. Makes explaining stuff easier than using words.

Thanks for the post 😛 Saves me the stupid effort of thinking of a multithreaded empirical example using ray-tracing 😉
 
Vylen said:
Damn pictures. Makes explaining stuff easier than using words.

Thanks for the post 😛 Saves me the stupid effort of thinking of a multithreaded empirical example using ray-tracing 😉
Click to expand...

I know, and the same thing can be said for nVidia's CUDA or ATI's offering.
 
Depends on the program.

I have a 2.26 at home and in apps that don't really use all the cores, yeah, it can be annoying. For example, I do lots of H264 HD encoding in MPEGStreamclip and it uses about 1/3rd the power. There also seems to be a bug where if I have it try to encode multiple videos at once, it crashes after 10-15 seconds. So in this instance, yeah, a quad 3.33 would probably be better.

But it's not like I ever feel my computer is slow or can't handle what I'm doing.
 
I am curious about this as well. I have been on the fence for months at my company as to when and how to upgrade our PowerMac G5s. We have Adobe CS5 shelfed, just waiting for an Intel processor...I have it at home on my iMac and CAN'T WAIT to unbox it here at work.

I've been going back and forth between an 8-core or quad core for my machine. I will be doing work in Premiere, After Effects, Photoshop, InDesign, Dreamweaver and Illustrator, with some DVD authoring and music on the side. I know CS5 has some potential for all 8 cores, but I do not know if the extra $$$ is worth it in terms of speed. One day I think 2.26 8-core, 12 GB RAM. Then the next, 2.93/3.33 quad with 6GB/8GB RAM...oh, just wait one more week for upgrades.... I usually like to run with several Adobe CS3 apps, mail, three browsers, Excel, Word, and Navicat open at the same time, and this machine is fast enough for daily work, but it feels and performs slower even than my iMac at home.

The other machine I will be purchasing will not be used for any rendering or motion graphics, so I am not as concerned about it (base quad, base RAM, 512MB ATI). But mine needs to be a beast in comparison.

Will Adobe CS5 apps make the 8-core model worth it? Any one have any real experience or knowledge of the two compared using the CS5 apps?
 
ahenson said:
I am curious about this as well. I have been on the fence for months at my company as to when and how to upgrade our PowerMac G5s. We have Adobe CS5 shelfed, just waiting for an Intel processor...I have it at home on my iMac and CAN'T WAIT to unbox it here at work.

I've been going back and forth between an 8-core or quad core for my machine. I will be doing work in Premiere, After Effects, Photoshop, InDesign, Dreamweaver and Illustrator, with some DVD authoring and music on the side. I know CS5 has some potential for all 8 cores, but I do not know if the extra $$$ is worth it in terms of speed. One day I think 2.26 8-core, 12 GB RAM. Then the next, 2.93/3.33 quad with 6GB/8GB RAM...oh, just wait one more week for upgrades.... I usually like to run with several Adobe CS3 apps, mail, three browsers, Excel, Word, and Navicat open at the same time, and this machine is fast enough for daily work, but it feels and performs slower even than my iMac at home.

The other machine I will be purchasing will not be used for any rendering or motion graphics, so I am not as concerned about it (base quad, base RAM, 512MB ATI). But mine needs to be a beast in comparison.

Will Adobe CS5 apps make the 8-core model worth it? Any one have any real experience or knowledge of the two compared using the CS5 apps?
Click to expand...

See Cow2Beef.exe above. Pretty much explains what the difference will be v
 
This is ridiculous!

I have a 2008 mac pro, 2.8 octo with 4TB and 10gig of ram. Im not going to lie even my 2 year old machines screams. Decoding a 1.5gb 720 mkv of lost takes 15 mins! I am a professional photographer and graphic designer and saying any mac pro cannot run OSX quickly is ridiculous! my pro benches at nearly 11,000 the newest macbook pro 13" benches at 3500 and runs osx perfectly! yes there is faster technology! but these machines are made for the long haul! that is why they are a pro machine, no pro upgrades there main machine every upgrade! they are simply too expensive and there is no need to! they are work horses and will be used to destruction.
I think you people need something better to think about! yes they havent been upgraded in a stupid time! but they will be! just wait a little longer, and to say the new mac pros are going to be stupidly faster than the current ones is wrong! 15-20% max and how noticeable will that be in osx, there will be none. Buy some ssd's then youl see some performance, wait for developers to finaly make 64 bit versions of there software for you to see a difference. CS5 is the first 64 bit CS for the mac and it will run well on all intel 64 bit pro workstation.
Final note get a grip make some money from your machines and stop annoying everyone with this.. OMG the 8 core 2.26 mac pro the best machine apple makes and is slow in osx, its ********.
 
This is ridiculous!

BTW im sorry to the pro's out there who know exactly what i mean, there is too many threads from too many people talking about mac pros being slow! they arnt apple just hasnt upgraded them and yes they are expensive at the moment. Im waiting for all the mac pro lovers turned haters because they are over a year old suddenly loving and being more disappointed because they will more expensive. The reason for this > pros pay for them because they earn money from them, amateurs complain because the cant afford them because they would be an ultimate luxury item compared to an iMac £1000 cheaper. Also for the continuous post when they finaly arrive on people asking what spec to buy etc! This to me spells out amateur if you cant work out what machine you need and are a pro then there is something wrong.
 
tomscott1988 said:
BTW im sorry to the pro's out there who know exactly what i mean, there is too many threads from too many people talking about mac pros being slow! they arnt apple just hasnt upgraded them and yes they are expensive at the moment. Im waiting for all the mac pro lovers turned haters because they are over a year old suddenly loving and being more disappointed because they will more expensive. The reason for this > pros pay for them because they earn money from them, amateurs complain because the cant afford them because they would be an ultimate luxury item compared to an iMac £1000 cheaper. Also for the continuous post when they finaly arrive on people asking what spec to buy etc! This to me spells out amateur if you cant work out what machine you need and are a pro then there is something wrong.
Click to expand...

What's with the double post?
 
it all depends on the applications used...
my 2.26 octad will outperform a 3.2 octad harpertown when rendering in maxwell render (using all cores [physical and logical, so 16 cores] 100% usage. BUT in single thread performance, it loses. If you are using multithreaded as has been said it can be quite fast. I'm not saying it is a slouch by any means in single thread either, it is a fast machine all around 😀
add more ram and a ssd and it'll scream
 
WardC said:
I have read reports on here about the 2.26GHz 8-core Mac Pro being reportedly slow...but is it really THAT slow?? I mean, I am in the market for a Mac Pro right now and I don't really know if I will wait until the 6-core machines come out. How slow does the 2.26GHz feel in the Mac OS when navigating, opening apps? Is there a big difference between the 2.66GHz and the 2.93GHz? Also comparably how is the single core 3.33GHz vs the 8-core 2.26GHz? Any help would be appreciated. Thanks!!
Click to expand...

For navigating / opening apps, your Hard Drive is responsible for 90%+ of the speed. RAM can be a factor (as a lot of RAM can help caching). An ASUS EEE-PC with an Intel Atom can be reasonably snappy, as it uses an SD so seek times are low, until it's pitiful RAM gets full and it just hangs ...

Open Activity Monitor (or Task Manager if you are on Windows) and look at CPU use. If it's flat-out on one core, then a 3GHz CPU will make it much faster. If it's flat-out on a couple of cores, then more cores might help. If it's 50% on two cores, it probably means that one process is switching between cores ... it's really just 100% on one core (so a 3HGz CPU is good).

If you just want to open iPhoto faster, then get an SSD.
 
quick reaaaaaaaaaaaaaally n00b question 😱 I didn't undestand the pic could someone explain it to me please 🙂 😕? thanks 😱
 
bender o said:
quick reaaaaaaaaaaaaaally n00b question 😱 I didn't undestand the pic could someone explain it to me please 🙂 😕? thanks 😱
Click to expand...

Think of it this way. The left side of the picture is a single core processor (in the pic it says quad core, but ignore that, it was a mistake). A single processor has a task to complete, which is to convert 4 cows into beef for burgers. So, it takes the signle core CPU 8 seconds to process each cow. Hence, 8 seconds times 4 cows is 32 seconds time to complete the whole program.

Now, look at the right side, that's a quad core CPU. It can process stuff much faster with each core. Now think of the previous program we ran on the single core. The program is to feed the CPU with cows to convert into beef. No problem, the Quad core will take 8 seconds total, because each core (all 4 of them) takes up a cow to process. Taking 24 seconds less to process all 4 cows than the single core.
 
jav6454 said:
Think of it this way. The left side of the picture is a single core processor (in the pic it says quad core, but ignore that, it was a mistake). A single processor has a task to complete, which is to convert 4 cows into beef for burgers. So, it takes the signle core CPU 8 seconds to process each cow. Hence, 8 seconds times 4 cows is 32 seconds time to complete the whole program.

Now, look at the right side, that's a quad core CPU. It can process stuff much faster with each core. Now think of the previous program we ran on the single core. The program is to feed the CPU with cows to convert into beef. No problem, the Quad core will take 8 seconds total, because each core (all 4 of them) takes up a cow to process. Taking 24 seconds less to process all 4 cows than the single core.
Click to expand...

Great thanks 🙂 I thought that at first but the name on the first pic through me off lol
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back