"We Never Use Assembly Language"
in [ASM]
|
From: nessuno on 12 Mar 2006 14:57 randyhyde(a)earthlink.net wrote: > o//annabee wrote: >> randyhyde(a)earthlink.net wrote: >> [cut] > [cut] > > And the > > time they initially invested would be payed back handsomly by orders of > > magnitude. Assembly, done properly, with a propper tool, Like RosAsm is > > simply at least 100 times, as effective a tool as any HLL. > Prove this statement. > Words are easy to pass through your lips. Writing a game, as Tim Sweeny > has done, is a bit more work. People will take Tim Sweeny's opinion > over your's every time as you don't have anything to back up your > claims. Tim does. Tim may *not* be right, mind you, but at least he's > done the work and knows the problems. You haven't, and you don't. > [cut] >> [cut] > > > I think you confuse the concept of "hate assembly" with "hate all all > > > other languages but assembly." The two ideas are not the same. The fact > > > that someone recognizes that assembly is not the tool for every job in > > > no way implies that they hate assembly. > > > > Yes it DOES. > > Whatever you say. You only demonstrate your bigotry and ignorance with > such a statement. > [cut] >> [cut] After a long time I haven't been reading from this NG it happens that today I am here watching a few threads and finding that not much has changed with respect to Wannabee way of thinking about programming, especially about efficiency of programming with Assembly. These days I have started to hack some Linux kernel code in the scheduler subsystem because I need some information to be delivered to user space about processes when migrating from one CPU to another in SMP machines. The code I have to read, study and hack in order to get those statistics is 95% C code (actually GNU/C code). The most of it is located in two files that are, relative to the location of kernel code, ../include/kernel/sched.h ./kernel/sched.c. They are about 12,000 lines of C code that I have to read and study. I's say "wow". It means that I simply will modify two files in order to have my hacks working in a lot of different machines without the need to understand each different CPU's architecture and without the need to be able to write about 12 different Assembly sintaxes. Can you imagine I had to study Assembly for Compaq Alpha AXP, Sun SPARC and UltraSPARC, Motorola 68000, PowerPC, PowerPC64, ARM, Hitachi SuperH, IBM S/390, MIPS, HP PA-RISC, Intel IA-64, DEC VAX, AMD x86-64, AXIS CRIS, and Renesas M32R architectures? It means that if my work will some day be appreciated by the community of Linux kernel developers I will not be forced to "translate" my code in all those Assemblies in order to have it merged in the official tree. Ehi, Wannabee? Has it all above mentioned anything to to with productivity of HLL compared to Assembly? Just one more thought. Did I understand well that a RosAsm executable must be produced from a single file? Did I understand well that you can't link together different object files to produce a single executable? If it were applied to Linux kernel code we couldn't have had that beautiful mechanism that is the module loading/unloading that permits to link pieces of code (like devices drivers) at run-time into a working kernel. Do you understand what does it mean in terms of memory savings? The kernel is able to link module's code at will, e.g when a new piece of hardware is attached at a running machine, and to unlink the module's code when it doesn't need it anymore. Can you do it with RosAsm, seen that it can't even link different pre-compiled objects together in a single executable? Please forgive me and explain if I got wrong informations about Rosasm. Some months ago, when I heard about RosAsm in this NG, I really wanted to give it a try but later I discovered that it can't work on my operating system and I don't want to purchase a MS-Win licence only to run Ros Asm. I hope one day to see it running on a free and open source operating system, no matter which. fabio de francesco
From: nessuno on 12 Mar 2006 14:57 randyhyde(a)earthlink.net wrote: > o//annabee wrote: >> randyhyde(a)earthlink.net wrote: >> [cut] > [cut] > > And the > > time they initially invested would be payed back handsomly by orders of > > magnitude. Assembly, done properly, with a propper tool, Like RosAsm is > > simply at least 100 times, as effective a tool as any HLL. > Prove this statement. > Words are easy to pass through your lips. Writing a game, as Tim Sweeny > has done, is a bit more work. People will take Tim Sweeny's opinion > over your's every time as you don't have anything to back up your > claims. Tim does. Tim may *not* be right, mind you, but at least he's > done the work and knows the problems. You haven't, and you don't. > [cut] >> [cut] > > > I think you confuse the concept of "hate assembly" with "hate all all > > > other languages but assembly." The two ideas are not the same. The fact > > > that someone recognizes that assembly is not the tool for every job in > > > no way implies that they hate assembly. > > > > Yes it DOES. > > Whatever you say. You only demonstrate your bigotry and ignorance with > such a statement. > [cut] >> [cut] After a long time I haven't been reading from this NG it happens that today I am here watching a few threads and finding that not much has changed with respect to Wannabee way of thinking about programming, especially about efficiency of programming with Assembly. These days I have started to hack some Linux kernel code in the scheduler subsystem because I need some information to be delivered to user space about processes when migrating from one CPU to another in SMP machines. The code I have to read, study and hack in order to get those statistics is 95% C code (actually GNU/C code). The most of it is located in two files that are, relative to the location of kernel code, ../include/kernel/sched.h ./kernel/sched.c. They are about 12,000 lines of C code that I have to read and study. I's say "wow". It means that I simply will modify two files in order to have my hacks working in a lot of different machines without the need to understand each different CPU's architecture and without the need to be able to write about 12 different Assembly sintaxes. Can you imagine I had to study Assembly for Compaq Alpha AXP, Sun SPARC and UltraSPARC, Motorola 68000, PowerPC, PowerPC64, ARM, Hitachi SuperH, IBM S/390, MIPS, HP PA-RISC, Intel IA-64, DEC VAX, AMD x86-64, AXIS CRIS, and Renesas M32R architectures? It means that if my work will some day be appreciated by the community of Linux kernel developers I will not be forced to "translate" my code in all those Assemblies in order to have it merged in the official tree. Ehi, Wannabee? Has it all above mentioned anything to to with productivity of HLL compared to Assembly? Just one more thought. Did I understand well that a RosAsm executable must be produced from a single file? Did I understand well that you can't link together different object files to produce a single executable? If it were applied to Linux kernel code we couldn't have had that beautiful mechanism that is the module loading/unloading that permits to link pieces of code (like devices drivers) at run-time into a working kernel. Do you understand what does it mean in terms of memory savings? The kernel is able to link module's code at will, e.g when a new piece of hardware is attached at a running machine, and to unlink the module's code when it doesn't need it anymore. Can you do it with RosAsm, seen that it can't even link different pre-compiled objects together in a single executable? Please forgive me and explain if I got wrong informations about Rosasm. Some months ago, when I heard about RosAsm in this NG, I really wanted to give it a try but later I discovered that it can't work on my operating system and I don't want to purchase a MS-Win licence only to run Ros Asm. I hope one day to see it running on a free and open source operating system, no matter which. fabio de francesco
From: o//annabee on 12 Mar 2006 16:18 P? Sun, 12 Mar 2006 19:22:12 +0100, skrev sevagK <kahlinor(a)yahoo.com>: > > Spam Killer wrote: >> On 11 Mar 2006 20:57:36 -0800, "sevagK" wrote: >> > >> >That's the benchmark we have around here of testing how fast Betov >> >looses an argument. >> > >> Hi Troll! >> >> There's a much better indicator of what's going on around here. >> >> If the "Evil Mr. Hyde" and his "HLA-Trolls" viciously attack somone, >> you can be sure he knows much about Assembly Language. >> >> They team up against such people, to drive them away, because they can >> tell newbies how much they deceive them! > > You have deceived yourself, dude. The vicious attacks are done by > Rene, sometimes supported by his goons. Hyde responds to his > insanities, sometimes joined by others and not just people who use HLA > mind you, but by people who can see reason. > > See the indicater yourself if you have not yet poked out both your > eyes. Everytime Randall posts something, even if it has nothing to do > with Betov or Rosasm, you'll see a nutty Rene post insanities like the > one above showing everyone the bare fact: Betov has no competence to > respond to Randall's knowledge level. If Ren? has no competanse, then who wrote RosAsm ? :)) If it was not Ren?, then I would like to know who is responsible for the most efficient programming tool I ever had the pleasure of working with. If Randall Hyde has any competanse, then who wrote HLA? If it was not Randall I like to know who is responsible for this catastrophy. :) > -sevag.k > www.geocities.com/kahlinor > --
From: Charles A. Crayne on 12 Mar 2006 17:05 On 12 Mar 2006 11:57:02 -0800 "nessuno" <fmdf66(a)yahoo.com> wrote: :I really wanted :to give it a try but later I discovered that it can't work on my :operating system and I don't want to purchase a MS-Win licence only to :run Ros Asm. I ran it under WINE, and although I did not test it rigorously, the only function I found which did not work was debug -- and even that could probably be fixed without too much effort. -- Chuck
From: sevagK on 12 Mar 2006 17:13
Spam Killer wrote: > On 12 Mar 2006 10:22:12 -0800, "sevagK" wrote: > > >[...] > > Funny, a Troll who pretends he doesn't know what a Troll is. :-) Really now. The one pretending is you. A troll is someone who spews out useless posts to draw attention to him or herself. There is no bigger troll on this forum than Betov. Just count the number of posts by this clown that amount to zip. It's at the top of the list. -sevag.k www.geocities.com/kahlinor |