If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#21
|
|||
|
|||
ISA does not matter
In article q3Zco.13329$Sr7.6946@hurricane,
Owen Shepherd wrote: Brett Davis wrote: If PowerPC was just another RISC chip this could not have happened. Being late to the market with a me-too product would not have worked. PowerPC dominates because of AltiVec and the bitfield extract and other cool useful instructions that give good performance and good inner loop code density. Programmers LIKE PowerPC, whereas MIPS and ARM are tolerated. (This is a major reason PowerPC dominates, pity the fool manager that picks MIPS and cant find good programmers to work for him.) ISA does matter, just not the way you think it does. Brett - Actually working on ARM code right now. My general observation of talking to PowerPC developers has been: 1. IBM couldn't have made a more impenetrable assembly language. Seriously guys, never heard of register prefixes? Apple's variation was so much nicer. I write in C with a style that converts to clean assembly. (Which can be kinda odd looking to others to read.) I will do an AND 0x4f00, and the IBM compiler will issue some crazy indecipherable gibberish opcode I have to look up in the book. Then I have to read that page several times and squint, uhhh, Ok... This kinda sucks, would like to be able to right click on that disassembly and get the opcode translated to simple form. Often though that magic bit field extract opcode saved an instruction in my inner loop. So I only whine a little about the inconvenience. Then I look at the MIPS version for the Playstation2 port, and barf. Could that compiler possibly be more stupid with regard to constants in code, hell no. Steaming pile of dodo for all the compiled code. Violates the golden rule "its the software stupid". Instead of $2000 compilers, you should spend millions on your compiler, and give it away. Free razor handles, so you can sell the blades. This is not AVR32 where you can excuse a compiler for being new, MIPS has been around 25 years! And now the architecture is obsolete, overdue to be replaced by a MIPS copy of Thumb2. Basically anyone that uses MIPS today is an idiot, PowerPC and ARM give far better code size and performance. Any manager that picked MIPS because it was cheaper, does not know basic math. (Yes, there may be a niche or two left for MIPS, mostly installed base.) 2. Why are the bits numbered back to front? Way to confuse the hell out of people. For a monochrome bitmapped display IBM is ordering the bits the correct way. Just because the way you were taught, and the way everyone else orders the bits is the other way, does not make that right. I prefer IBM bit order, to each his own. On the other hand, ARM development has been a rather pleasant experience; to each his own - Owen (Not to say that any ISA is perfect; god knows ARM has its foibles, just like any other. They're just not as pervasive.) |
#22
|
|||
|
|||
ISA does not matter
In article q3Zco.13329$Sr7.6946@hurricane,
Owen Shepherd wrote: Brett Davis wrote: If PowerPC was just another RISC chip this could not have happened. Being late to the market with a me-too product would not have worked. PowerPC dominates because of AltiVec and the bitfield extract and other cool useful instructions that give good performance and good inner loop code density. Programmers LIKE PowerPC, whereas MIPS and ARM are tolerated. (This is a major reason PowerPC dominates, pity the fool manager that picks MIPS and cant find good programmers to work for him.) ISA does matter, just not the way you think it does. Brett - Actually working on ARM code right now. My general observation of talking to PowerPC developers has been: 1. IBM couldn't have made a more impenetrable assembly language. Seriously guys, never heard of register prefixes? Apple's variation was so much nicer. Most people program in C, so you only see this when debugging and you pop open the disassembly window, which I always have open. Seriously gibberish opcodes, have to look them up in the book and growl. Generally its a constant that the compiler tuned into a bit field extract. Which saved an instruction, or saved a register, making me happy. Then I look at the MIPS port, man that MIPS compiler is a stinking pile of do do. Way more instructions issued, and issued in a brain damaged way that hurts performance. A 15 year old compiler that was never good in its heyday, and has not been improved. The whole idea of RISC is simpler faster more efficient compilers. The MIPS crowd has dropped the ball (financial reasons) and as a result have crippled their chances in the market to sell MIPS chips. Maybe there is some $2000 compiler for MIPS that does not suck, we used whatever we were given, which was whatever was cheap. This is a classic give away the razor handle to sell the blades situation. Letting people use crap compiler, resulting in people thinking your product sucks, is just a bad idea. 2. Why are the bits numbered back to front? Way to confuse the hell out of people. For operations on a monochrome bitmapped display IBM is ordering the bits in the correct order. Just because this is opposite of everyone else and what you were taught does not mean it wrong. On the other hand, ARM development has been a rather pleasant experience; to each his own |
#23
|
|||
|
|||
ISA does not matter
In article q3Zco.13329$Sr7.6946@hurricane,
Owen Shepherd wrote: Brett Davis wrote: If PowerPC was just another RISC chip this could not have happened. Being late to the market with a me-too product would not have worked. PowerPC dominates because of AltiVec and the bitfield extract and other cool useful instructions that give good performance and good inner loop code density. Programmers LIKE PowerPC, whereas MIPS and ARM are tolerated. (This is a major reason PowerPC dominates, pity the fool manager that picks MIPS and cant find good programmers to work for him.) ISA does matter, just not the way you think it does. Brett - Actually working on ARM code right now. My general observation of talking to PowerPC developers has been: 1. IBM couldn't have made a more impenetrable assembly language. Seriously guys, never heard of register prefixes? Apple's variation was so much nicer. Most people program in C, so you only see this when debugging and you pop open the disassembly window, which I always have open. Seriously gibberish opcodes, have to look them up in the book and growl. Generally its a constant that the compiler tuned into a bit field extract. Which saved an instruction, or saved a register, making me happy. Then I look at the MIPS port, man that MIPS compiler I used was a steaming pile of do do. Way more instructions issued, and issued in a poor way that hurts performance. A 15 year old compiler that was never good in its heyday, and has not been improved. The whole idea of RISC is simpler faster more efficient compilers. The MIPS crowd has dropped the ball (financial reasons) and as a result have crippled their chances in the market to sell MIPS chips. Maybe there is some $2000 compiler for MIPS that was not poor, we used whatever we were given, which was whatever was available. This is a classic give away the razor handle to sell the blades situation. Letting people use a poor compiler, resulting in software engineers thinking poorly of your CPU, is just a bad idea. 2. Why are the bits numbered back to front? Way to confuse the hell out of people. For operations on a monochrome bitmapped display IBM is ordering the bits in the correct order. Just because this is opposite of everyone else and what you were taught does not mean it wrong. On the other hand, ARM development has been a rather pleasant experience; to each his own |
#24
|
|||
|
|||
ISA does not matter
In article q3Zco.13329$Sr7.6946@hurricane,
Owen Shepherd wrote: Brett Davis wrote: If PowerPC was just another RISC chip this could not have happened. Being late to the market with a me-too product would not have worked. PowerPC dominates because of AltiVec and the bitfield extract and other cool useful instructions that give good performance and good inner loop code density. Programmers LIKE PowerPC, whereas MIPS and ARM are tolerated. (This is a major reason PowerPC dominates, pity the fool manager that picks MIPS and cant find good programmers to work for him.) ISA does matter, just not the way you think it does. Brett - Actually working on ARM code right now. My general observation of talking to PowerPC developers has been: 1. IBM couldn't have made a more impenetrable assembly language. Seriously guys, never heard of register prefixes? Apple's variation was so much nicer. Most people program in C, so you only see this when debugging and you pop open the disassembly window, which I always have open. Seriously gibberish opcodes, have to look them up in the book and growl. Generally its a constant that the compiler tuned into a bit field extract. Which saved an instruction, or saved a register, making me happy. Then I look at the MIPS port, man that MIPS compiler I used was a steaming pile of do do. Way more instructions issued, and issued in a poor way that hurts performance. A 15 year old compiler that was never good in its heyday, and has not been improved. The whole idea of RISC is simpler faster more efficient compilers. The MIPS crowd has dropped the ball (financial reasons) and as a result have crippled their chances in the market to sell MIPS chips. Maybe there is some $2000 compiler for MIPS that was not poor, we used whatever we were given, which was whatever was available. This is a classic give away the razor handle to sell the blades situation. Letting people use a poor compiler, resulting in software engineers thinking poorly of your CPU, is just a bad idea. 2. Why are the bits numbered back to front? Way to confuse the hell out of people. For operations on a monochrome bitmapped display IBM is ordering the bits in the correct order. Just because this is opposite of everyone else and what you were taught does not mean it wrong. On the other hand, ARM development has been a rather pleasant experience; to each his own |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
fsb speed - why does it matter? | James Hanley | Overclocking | 52 | November 7th 04 01:04 PM |
6800 GT, does it matter which brand? | husker | Nvidia Videocards | 6 | October 13th 04 03:41 PM |
Does a high FSB really matter?? | Gareth Jones | Asus Motherboards | 10 | May 29th 04 03:31 AM |
Cannot fit ISA card into ISA slot | Peter C | Dell Computers | 0 | July 25th 03 08:48 PM |
Does this matter? | jimrx4 | Intel | 4 | July 13th 03 05:51 PM |