Which I believe is the point of the article anyway. Programmers should go back to the basics and produce more efficient and simple code rather than make it more complicated. Any expert in this field please correct me if I'm wrong as I'm not really into this field but thats the conclusion I can come to with my (near zero) knowledge anyway.

well its all about keeping the cash flow coming in though... if bill gates stopped at xp then the companies who do comp softwares and hardware will be stiffled by it. There will be no 8800gts and no dual cores.. im sure there are programmers who can code such systems, but its a big money making business, so all we comsumers have to live by it. and the programmers follow where the money talks

That would be ideal but not in this day and age.

Software companies are producing apps that caters for everybody. They can't be making softwares that only cater for a certain segment of the market.

Take Microsoft word for instance. Frankly speaking, I only use it for typing notes, spelling check, thesaurus, some tables and pictures. That's all. However, there are other functions that I don't use but is used by other people like simple drawing, formulas, review, web page design, etc. And sometimes, there are those rare functions that I need that is available in Word.

Take the generic Paint program in windows as another example. It is so simple and easy to use, I usually use it to edit simple pictures to paste in this forum. However, when I want to do something more advance, the program doesn't have that function. So I have to resort to big-bulky full-of-functions/code apps like photoshop to edit the pictures.

Then there are people who start complaining that a certain app is not good as it does not have this function or that function. Imaging if microsoft word does not have the automatic spell-checker.......

Yes, definitely not memory bandwidth limitation. The smart memory pre-fetch has eliminated a lot of memory transaction on FSB.

Actually, there are just too many branches and serialization in modern software, where the the SMT is completely useless. Other factors like cache and page tables coherency also will slow down the multi threaded performance exponentially when we have more n more cores in the system.


Of course, these days it is impossible to write programs with only assembly or C language. So the optimization won't be as good as oldern days (like the Mac PLUS). Heck, even C programming is a lot slower than ASM programming! (from my observation from internal test suites).

This post has been edited by charge-n-go: Jun 3 2007, 01:48 PM

Prefetching tries to hide latency, it can't hide a lack of memory bandwidth. It increases memory transactions because you don't always prefetch the right data.

Programming wise, it's a matter of where you spend your time optimizing of course. Good programming practices coupled with a good compiler can do very well. Of course for your super critical stuff you will want to go down to ASM, but otherwise, you can probably get better improvements by simply spending your time improving your algorithm rather than hacking at ASM code. Same thing with hardware design. Biggest gains come from improvement in algorithms, the gain typically gets smaller as you move down to a lower level.

Yup. I still remember the days when people talk about code optimization. Assembly (ASM) language was the best!! However, it was also a pain in the neck to programme and virtually impossible to write for programs these days as there are million and million lines of codes......

It is in the software with millions of lines of code (such as an OS) that you typically see a lot of hand tuned ASM code. You don't have to do it all in ASM from scratch, you can compile it and then hand optimize the ASM code for certain critical stuff. Obviously it makes sense to do for a function you call a billion times, or for an important inner loop.

LoL, my mistake. What was I thinking just now

Well said, the algorithm is very important. C / C++ compiler should not have too much problem as Intel is actually providing them to customers, but I dunno how efficient other high level language will be. From my experience in test coding, the most important part will be the algorithm within a FOR loop. If 1 loop is a bit slower, the end result will b farking slow considering there are million of loops (streaming for eg). Actually, some critical functions which access hardware directly (usually APIs) will be written in ASM or C programming.

Wow! That's deep but informative. Not for average users like me!

Thx for the info!

Just curious, do average programmers do ASM these days?

This post has been edited by kmarc: Jun 3 2007, 02:44 PM

Anyone who is serious about performance would not use other less efficient languages

No, you only use ASM to optimize some really critical stuff. You would have to be at least "pretty good" and understand the hardware you are optimizing for before you can do that. Else you may end up doing a worse job than the compiler.

interesting machine language stuff and C ehehe. i rather see some simpler stuff so i can stilll run my old p4 machine with the current world softwares. I dont like change to be so rapid, its an exponential rise, as far as tech is concerned. I hope there will be somewhat a rest so people like me can catch up and quipt myself

i can just say one thing, if anyone thinks that older systems performs better than that of current for their normal tasks..... why dont you just switched to old systems to accomodate your needs.

There are thousands, n perhaps millions of old systems dumped in basement and store waiting, never to be used again. You want the latest technology, and yet you want it to be simple. Let me ask one thing.... can a Ferrari goes from 0 to 100kmph in 3.3seconds if its engine is as simple as a proton's?

u have a newer tech, u need to have new system n new UI n new softwares. If not there is no point for the new technologies to be introduced

One might ask.....do we need Ferraris to go to work or college? But thats not the point of this topic. No 1 here is saying that we do not want to embrace new tech etc nor that old tech is better than new. But the implementation could possibly be improved. I for 1 agree with that.