Stop Microsoft
Operating Systems => macOS => Topic started by: Pissed_Macman on 29 June 2003, 08:32
-
http://overclockers.com/tips00408/ (http://overclockers.com/tips00408/)
-
we will have to see for real when we get a chance to actually use the g5's. i would hate for Apple to be lying to us. huhm. or it means that windows pc users are actually scared and will say anything. or it could be that the code was written better for one system and not the other. or fuck. it means the x86 can do single tasks faster.
[ June 29, 2003: Message edited by: ecsyle:951 ]
-
This article is utter hogwash. They are trying to discredit Apple's tests for using a slower compiler, when it is the common compiler for both platforms. Of course, the PC will score higher if it uses a compiler by Intel or Microsoft, just like the Mac will score higher with a compiler from IBM. I don't know where the author gets that Apple uses a different compiler, as gcc was used for all the machines tested.
It shows how ignorant the author is when he dismisses the gcc compiler as almost unheard of, when it is the defacto standard in free UN*X operating systems. He even goes as far as claiming that Windows is the "native environment" of the x86! What a bigot!
[ June 29, 2003: Message edited by: Laukev7 ]
-
I looked at the tables in AMDzone (the site where the author apparently takes his figures), and they are very unprofessional. If you take a look, you'll notice that many processors are mislabled. A ridiculous mistake I found was "Apple 3GHz P4". Go figure. :rolleyes:
[ June 29, 2003: Message edited by: Laukev7 ]
-
Not to mention the fact that he uses the official Spec Scores that rated the Xeon. The spec tests to compare are diffrent, they try and level the playing field for both systems involved considering that you are comparing cross platform performance.
Bullshit.
-
For a second I thought they had something legit. This goes to prove that charts are shit when they're made by someone unprofessional and biased.
-
it was funny when he dismissed GCC as an unheard of compiler, I'm not even aware of the names of other compilers that are present in visual C++, etc.
It shows how stupid this guy is when he puts on there that using Linux was not cool, implying that the machines would run slower with linux. pft.
However, its not like Apple didn't blow up these results.. of course they did, just not to the extent that he implies. For instance, I highly doubt that Photoshop runs 2x as fast. Just remember when Jobs said the single 867G4 ran Photoshop 80% faster than the 1.8GHz P4... and of course he was totally full of shit...
One thing that bothered me bout the Photoshop G5 to PC comparisons is that ONE step on the PC took forever, if that one step wasn't there the PC would be neck and neck. This just shows how Apple picked the filter that was hardest on the PC and not so hard on the Mac just as a crowd pleaser. Intel could have very easily picked a filter that is really hard on the Mac and easy on the PC just to slow the Mac down, then the PC would be 2x as fast.
So.... I believe that the G5 will be faster than most stuff, however, not as fast as Jobs says.
P.S. My dad has a dual Xeon system. He is very unhappy with it. Before the Xeons he had a dual PIII. The Xeon Machine is faster, but not that much faster. The really funny thing about it is at any given time it only uses 50% of each CPU. Hows that for multi processor management in Windows. I keep telling him to install Linux so it can take advantage of the Xeon's and the dual CPU's. But so far he just has XP on it.
Just my 2Cents, however this post is a little longer than 2Cents worth, so I'll say its my 6Cents.
-
quote:
Originally posted by jeffberg: Mac Capitalist:
it was funny when he dismissed GCC as an unheard of compiler, I'm not even aware of the names of other compilers that are present in visual C++, etc.
It shows how stupid this guy is when he puts on there that using Linux was not cool, implying that the machines would run slower with linux. pft.
However, its not like Apple didn't blow up these results.. of course they did, just not to the extent that he implies. For instance, I highly doubt that Photoshop runs 2x as fast. Just remember when Jobs said the single 867G4 ran Photoshop 80% faster than the 1.8GHz P4... and of course he was totally full of shit...
One thing that bothered me bout the Photoshop G5 to PC comparisons is that ONE step on the PC took forever, if that one step wasn't there the PC would be neck and neck. This just shows how Apple picked the filter that was hardest on the PC and not so hard on the Mac just as a crowd pleaser. Intel could have very easily picked a filter that is really hard on the Mac and easy on the PC just to slow the Mac down, then the PC would be 2x as fast.
So.... I believe that the G5 will be faster than most stuff, however, not as fast as Jobs says.
P.S. My dad has a dual Xeon system. He is very unhappy with it. Before the Xeons he had a dual PIII. The Xeon Machine is faster, but not that much faster. The really funny thing about it is at any given time it only uses 50% of each CPU. Hows that for multi processor management in Windows. I keep telling him to install Linux so it can take advantage of the Xeon's and the dual CPU's. But so far he just has XP on it.
Just my 2Cents, however this post is a little longer than 2Cents worth, so I'll say its my 6Cents.
One step? Dude I saw the comparison, ALL the steps took a significantly longer time, watch it again. That PC was slow as fuck... I don't care what the numbers say, they don't mean shit.
I don't care if it processes 1.85645446Million microziples and has a half terabite BIOS overune memry jackoff cumshoot accelerator cartrige xtreeem. It fucking kicked the shit out of a dual Xeon... case closed.
-
quote:
Originally posted by psyjax: plain 'ol psyjax:
One step? Dude I saw the comparison, ALL the steps took a significantly longer time, watch it again. That PC was slow as fuck... I don't care what the numbers say, they don't mean shit.
I don't care if it processes 1.85645446Million microziples and has a half terabite BIOS overune memry jackoff cumshoot accelerator cartrige xtreeem. It fucking kicked the shit out of a dual Xeon... case closed.
yes, they were all slower, but only slightly... that one step was an extreme apple used to twist the the numbers.
Also, I bet the PC wasn't using more than 75% of both its CPU's, the Mac was using 100%. Sure, thats windows fault. What happens when u get Linux on that thing?
[ June 29, 2003: Message edited by: jeffberg: Mac Capitalist ]
-
Why is it that you only take the Photoshop tests in consideration? What about the other tests? The G5 beat the PC hands down at the audio software test. The G5 was reportedly 3-5 times faster than the PC at the scientific/genoma/DNA software test, confirmed by the CEO's testimonial. It even beats the PC at Quake III. How many tests do you need to see that the G5 is much faster than today's Pentium or Xeon?
-
quote:
Originally posted by Laukev7:
Why is it that you only take the Photoshop tests in consideration? What about the other tests? The G5 beat the PC hands down at the audio software test. The G5 was reportedly 3-5 times faster than the PC at the scientific/genoma/DNA software test, confirmed by the CEO's testimonial. It even beats the PC at Quake III. How many tests do you need to see that the G5 is much faster than today's Pentium or Xeon?
well to be able to see if one system is significantly faster than another system. I need to use them both. I know Apple is, not full of shit, but stretching the truth... just as any other company would.
And again, I recall the 867 G4 vs 1.8P4. I have a 733 and a 1.7Athlon. Jobs said the 867 is 80% faster at Photoshop. I would say the PC is probably 80% faster than my Mac, no matter what I do, no matter how I reformat it, etc. Because of that blatant lie from Apple I do not trust these results and will never trust any of their results or any other company's results. I do agree that the G5 is one fast Puppy. But I doubt it is as fast as they say. Again, I will just have to wait until I can use one in an Apple Store.
-
quote:
Originally posted by jeffberg: Mac Capitalist:
And again, I recall the 867 G4 vs 1.8P4. I have a 733 and a 1.7Athlon.
Athlons have shorter piplelines then a Pentium 4, by far. This would definitally affect the outcome. Doesn't everyone and their cat know that mghz does not equal power?
[ June 29, 2003: Message edited by: bossesjoe: too lazy to post much ]
-
Apple claimed to be 80% faster than the Pentium IV, not the Athlon. As Bossesjoe said, there is quite a difference between a 1.8 GHz P4 and a 1.7 Athlon (assuming 1.7 is the actual speed of the processor, not AMD's approximate P4 equivalent, but even then it's still an approximation).
-
quote:
Originally posted by bossesjoe: too lazy to post much:
Athlons have shorter piplelines then a Pentium 4, by far. This would definitally affect the outcome. Doesn't everyone and their cat know that mghz does not equal power?
[ June 29, 2003: Message edited by: bossesjoe: too lazy to post much ]
yes, but AMD rates it a 2100+ as fast a 2.1P4. That would not equal an increase from 80% slower to 80% faster. I just don't understand why you guys believe every word that comes out of corporation mouths? All they want to do is sell stuff, make money. Whether the product is good or not is company specific. But the bottom line is that they will do anything to sell you their product. Lying is in that list of anything. I don't trust any corporate comparisons, they will always favor themselves. Just like I wouldn't call (that one auto insurance company with the Ford Explorers on their commericals that say they will give u the rates of their competitors, can't fucking remember their name) I wouldn't call them and ask for others rates and expect to find any one elses rates that are better than theirs. Why would you expect Apple not to stretch the truth. Why wouldn't the CEO's lie along with Apple, I'm sure most of the products they sell are for Macintosh. If they can get people to think the Mac is better they will sell more Software.
I just don't see any reason to believe Apple... so I don't
-
In any case, most of the articles that attempt to debunk the G5 so far have little credibility, and most of them fall in the same pitfalls as the ones I pointed out for this one. Whether the G5 is faster or not (which I believe it is), no one can deny that Apple is in a far better position, and their Powermac G5 is very competitive.
-
quote:
Originally posted by Laukev7:
In any case, most of the articles that attempt to debunk the G5 so far have little credibility, and most of them fall in the same pitfalls as the ones I pointed out for this one. Whether the G5 is faster or not (which I believe it is), no one can deny that Apple is in a far better position, and their Powermac G5 is very competitive.
I can agree with this statement... it just kind of scares me when people trust these big companies and don't even second guess them, thats all.
-
ok, let apple use an ultra tweaked ppc compiler (ibm designed ppc) agiasn't an ultra tweaked ix86 compiler (intel compiler) then compare the tests agianst gcc. that way apple will have conducted a slightly fairer review, and we will have an idea of where gcc needs to be improved.
-
but atleast this run, the g5 is a piece of ibm's work, a little power4. however, apples incrementalism bothers me. they should have used some form of liquid cooling, and pushed this thing to ridulus speeds. if it is quicker then a 3ghz p4 xeon at 2ghz, imagine what the g5 could do at 3 ghz or beyound. or if they went numa and use 4 of the them.... osx could run on numa hardware. and by now your probably thinkin "numa? holyshit, he's talking about sgi supercomputers". well, apple is great at bringing technolgy not usually associated with the desktop to the desktop, like 64bit machines, UNIX, and many others. so why not inexpensive numa
-
Pardon my ignorance, but what is 'numa'?
-
i dont know what it is but i want one.
lets see that would be #5,764,932 on my wish list.
-
quote:
Pardon my ignorance, but what is 'numa'?
NUMA - Non-uniform memory access
quote:
NUMA stands for non-uniform memory access. It's a design that can be used to build scalable systems. Basically you take zero or more processors, a memory and bus controller, and some RAM and bundle them together into a unit. (Typically this unit is on its own circuit board.) Attach two or more of these units together with a high-speed interconnect and you've got a NUMA system. Processor A on board 1 has local access to memory bank I, but has to go across the interconnect to reach memory bank II or board 2. Hence, non-uniform.
The Power Mac G5 is not a NUMA system. All memory access goes through the Apple-designed System Controller. The System Controller connects the processor(s) to memory and graphics, and to the HyperTransport fabric that provides access to the internal storage and expansion slots.
So if processor 1 wants to get at any piece of memory in the system, it goes through the system controller and into the RAM to get it. Processor 2 does exactly the same thing. Therefore the Power Mac G5 is not a NUMA system, because all memory accesses are uniform.
Now, at least in theory it would be possible for Apple to design a NUMA system around the Power Mac G5 architecture. They could design a System Controller to HyperTransport to System Controller link that basically connects two Power Mac G5-style systems together and allows them to function as a single computer. (Note that we're not talking about clustering here; we're talking about one computer with more than two processors.) But the Power Mac G5 is not it.
-
Thank you for this information, Boris. If I understand correctly, each processor gets its own memory. But what is the point?
-
Just one question: If the G5 powermac's aren't supposed to be shipping until august how could they get a hold of a G5 machine to do these tests ?
-
Maybe they based this on Apple's tests.
-
Speaking of memory, does the new G5 have L3 cache? It doesn't look like it's present.
-
The memory bus is fast enough that it doesn't need a L3 cache. The memory bus on the G4 was only 167MHz, on the G5 it is 1GHz.
-
ERrrrmmmmmm, GCC is has has been said already the de-facto compiler...... Un-educated idiots these Winblows / PCEEEEEEEEEEEEEEEE users.
What really makes me laugh though is that in a real life demo, say a very large Photoshop actions test (as seen on the Apple Keynote), they will still look at their stupid test results and deny all knowledge of what their eyes are actually seeing!!!!!
who cares what they think