GonzoDaGr8
Platinum Member
I'm just glad Apple is done with Motoscrota processors. :beer:
-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
Official Apple Keynote Thread - webcast online at about 3 pm est.
Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
LoL - I was reading a big arse debate in OSnews how Apple misled folks with their GCC benchmarks. Apparently they used a Dell machine running RedHat 9.0 which used i386 optimizations to go up against there new G5's. Had they used a Dell machine with Gentoo or XP they would of not come head. So yeah the new G5's are faster running against a Dell pc running RedHat 9.0 but that is not very honest approach at benchmarking if you ask me. As always folks I would what untill you have a real non-bais, and non-Apple benchmarks to see if what Steve Jobs ( a known lair ) is shoveling is true.
Willoughbyva
Diamond Member
Can't you use a monitor besides apples' if you buy on of their machines? If I had the cash I would consider buying one, but I guess I will continue to use Linux. I would like to try out OSX though.
GonzoDaGr8
Platinum Member
For now.. I'm sure that they will phase out the G4's soon enough.
GonzoDaGr8
Platinum Member
Yeah, Prolly won't see it in any powerbooks anytime soon..Look how long it took them just to get the G4 into the powerbook. I was actually meaning them getting the new G5(970) into iMacs and eMacs.
If you are concerned about compiling, why not get a dual Opteron? Pretty sure if you compiled the 64bit Linux kernel it would dominate the G5, or at least be a more fair comparison. The Opteron would be cheaper, but arguable less useful for general desktop use. Have to wait for the Athlon64. I still think the only good things about Mac are OSX (they would have died without it), the iPod and how beautiful the hardware looks (my friend has a Ti Book with a 23 inch Cinema display, it's amazing). Besides the custom software/interface in OSX they really don't have much going for them (you are paying them 30% profit margin on their high end desktop systems).
Also not sure why they are even talking about Quake III, isn't there a UT2003 for Mac yet? Anyway HL2 and Doom 3 are all that will matter (doubt HL2 will be released for OSX but who knows).
Also not sure why they are even talking about Quake III, isn't there a UT2003 for Mac yet? Anyway HL2 and Doom 3 are all that will matter (doubt HL2 will be released for OSX but who knows).
Argh, ok as a RedHat user (I used RedHat 95% of the time for almost a year), I have to chime in.
GCC offers several CPU specific optimizations. The flags involved are -mcpu and -march. Note that -march implies -mcpu. The best way to explain how these work is to use an example. I had a Pentium 4. If I compiled using -march=pentium4, GCC would use Pentium 4 specific instructions (aside from SSE and SSE2) in addition to scheduling these instructions in a manner optimized for the Pentium 4. Because P4 specific instructions aren't guaranteed to work on other CPUs, we can safely say that this is the highest optimization setting. But, if I were to copy this executable to an Athlon, it might not run. If it did, it might run poorly as the instructions were scheduled to take advantage of the Pentium 4's architecture.
Now, there is also the -mcpu option. What this tells the compiler is to generate code that includes scheduling optimizations but which does not use any special instructions available to that CPU that may not also be available to other processors. For instance, compiling with -mcpu=pentium4 will optimize the instruction scheduling for the Pentium 4. But I am guaranteed to be able to run the executable on an Athlon or a Pentium 3. Any i386 processor for that matter. It may not run as well as the instruction order has been tuned to the P4, but it will run nonetheless.
RedHat is not compiled for i386 in the general sense that most people first think. Only i386 instructions are used, but programs included in the distribution are often compiled with -mcpu=i686 to enhance performance for current-class CPUs while retaining compatibility. That way, RedHat can guarantee you can install the OS on a Pentium which is at the i586 level, or a VIA C3 or some other oddball x86-compatible CPU. For the most part, applications rarely benefit from these GCC optimizations. People have this deluded impression that GCC will spit out code that is 20% faster when in actuality you probably are lucky to get a 3 or 4% gain.
There are some components that RedHat does optimize because they are so fundamental and frequently-used that optimizations make sense. RedHat releases kernels that are compiled with -march optimizations. They also release libraries like glibc compiled with -march=i686. But in no way is RedHat daft enough, or masochistic-enough, to release software that is somehow tuned to i386 and nothing else.
I often wish the Gentoo people would realize they are often wasting their time compiling if their intention is to tweak for performance. I wish they'd realize that any performance difference between Gentoo and RH often has to do with the fact RH by default ships with a lot of running services, and a fairly large kernel. But for some reason, they've permeated this myth across the net that by compiling all your software from source for 3 days, you're bound to get a 20% improvement in speed for your text editor.
So what you have hear is the Gentoo myth applied to Apple saying they somehow used RH9 on the Dell to be destructive. I'm sorry, I just don't buy it. They probably used RH9 on the Dell because RedHat is a recognizable Linux distribution to people curious about switching away from Windows.
GCC offers several CPU specific optimizations. The flags involved are -mcpu and -march. Note that -march implies -mcpu. The best way to explain how these work is to use an example. I had a Pentium 4. If I compiled using -march=pentium4, GCC would use Pentium 4 specific instructions (aside from SSE and SSE2) in addition to scheduling these instructions in a manner optimized for the Pentium 4. Because P4 specific instructions aren't guaranteed to work on other CPUs, we can safely say that this is the highest optimization setting. But, if I were to copy this executable to an Athlon, it might not run. If it did, it might run poorly as the instructions were scheduled to take advantage of the Pentium 4's architecture.
Now, there is also the -mcpu option. What this tells the compiler is to generate code that includes scheduling optimizations but which does not use any special instructions available to that CPU that may not also be available to other processors. For instance, compiling with -mcpu=pentium4 will optimize the instruction scheduling for the Pentium 4. But I am guaranteed to be able to run the executable on an Athlon or a Pentium 3. Any i386 processor for that matter. It may not run as well as the instruction order has been tuned to the P4, but it will run nonetheless.
RedHat is not compiled for i386 in the general sense that most people first think. Only i386 instructions are used, but programs included in the distribution are often compiled with -mcpu=i686 to enhance performance for current-class CPUs while retaining compatibility. That way, RedHat can guarantee you can install the OS on a Pentium which is at the i586 level, or a VIA C3 or some other oddball x86-compatible CPU. For the most part, applications rarely benefit from these GCC optimizations. People have this deluded impression that GCC will spit out code that is 20% faster when in actuality you probably are lucky to get a 3 or 4% gain.
There are some components that RedHat does optimize because they are so fundamental and frequently-used that optimizations make sense. RedHat releases kernels that are compiled with -march optimizations. They also release libraries like glibc compiled with -march=i686. But in no way is RedHat daft enough, or masochistic-enough, to release software that is somehow tuned to i386 and nothing else.
I often wish the Gentoo people would realize they are often wasting their time compiling if their intention is to tweak for performance. I wish they'd realize that any performance difference between Gentoo and RH often has to do with the fact RH by default ships with a lot of running services, and a fairly large kernel. But for some reason, they've permeated this myth across the net that by compiling all your software from source for 3 days, you're bound to get a 20% improvement in speed for your text editor.
So what you have hear is the Gentoo myth applied to Apple saying they somehow used RH9 on the Dell to be destructive. I'm sorry, I just don't buy it. They probably used RH9 on the Dell because RedHat is a recognizable Linux distribution to people curious about switching away from Windows.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 25K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 24K
-
-
