Windows Vista Rules.

Page 18 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
Originally posted by: idiotekniQues
Originally posted by: Sunner
Originally posted by: idiotekniQues
i do

That's why I said most

of course, but substitute a lot of photographs for something else.

also, considering the ridiculous megapixel races out there, even little compact cameras are 10mp and any user coming back from a weekend getaway will easily have 100 shots to go through or what not.

and if not that, there are a lot of people that multitask with video/graphics/photo's etc that would fall into a similar category.

more testing is needed, but anandtechs preliminary statement says that 2gb is not enough, when i bet a lot of us power users thought that our 2gb investment was going to be vista proof.

i stand by my statement that while it may not affect everybody, it is much more than a niche market

The vast majority of users are like my parents, surfing, email, office apps, etc, the usual.
In my parents case, they get by just fine with an old Coppermine based Celeron 1.3 GHz with 448 MB of RAM(yeah, I took what was at hand, hence the odd amount )

The statement would have been more valuable if they had actually checked memory usage with the OS's.
 

idiotekniQues

Platinum Member
Jan 4, 2007
2,572
0
71
well with vista memory usage is usually more 100% because of superfetch, which xp does not have - so comparing memory usage is tricky.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: nweaver
lol....before you make some of these statements, you might ask around....Nothinman is a respected member of AT forums. His post count isn't from Neffing in OT all day, it's from technical posts, and valid opinions. You don't have to agree with him, but to call him a n00b who searches google is going to get you flamed.

I'm going to humor you and assume this is not Nothinman posting under a different id...but yes, a lot of his posts and comments are n00bish and if that's what's gaining respect around here it doesn't say much about this community in general.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: Nothinman
There was obvious immediate benefits when going from 16-bit to 32-bit software, the same can't be said of 32-bit to 64-bit. The only software that's been running into the 32-bit VM limitations is extreme niche software like high end databases, 3D renderers, etc and those already have ports to other 64-bit arches. For Linux users the only reason not to install a 64-bit distro is lack of flash and IMO that's a good thing anyway. But on Windows most of your software is closed and can only be recompiled by the developer so you'll still be using virtually all 32-bit software anyway so you should be looking for compelling reasons to switch, not vice versa.

Wow, obvious and immediate? That's not how I remember it...in fact the same lame arguments you are posing now were commonplace back in the Win 3.1 to Win95 transition. You forgot, or is this before your time? And there is quite a bit of open source software available for Windows...maybe you need to expand your searches-for-smartness to include sourceforge.net =)

Unless all of your data is kept on the stack and you don't use any strings, arrays, etc then pointers will be a requirement even if it's hidden behind a language interface so that you don't see it.

If you have to start the sentence with "unless" then it just means you're going to fabricate a condition to make it seem like you could have been right originally. You don't need pointers, however not using them is bad coding practice.

And I take if you've done a ton of multithreaded programming so it just comes as second nature to you, right?

No I haven't, but I'm not going to sit here and pretend I do. Your comment along the lines of: "SMP has been around for decades, but developers still haven't figured out how to use it" pretty much solidifies you as clueless.

The point is that CPUs are getting exponentially more capable, and 64-bit computing is an improvement and not a handicap.

No, 64-bit computing isn't a handicap but 64-bit Windows is.

On obsolete or underpowered hardware, perhaps...but in general, I don't don't see many startups eagerly rushing in to compete with MS on the OS front so it's not like your choices are plentiful.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: idiotekniQues
according to anandtech, vista needs over 2gb to be truly efficient:

by the time i get vista ill have 2 more 1gb sticks but oi vay!

i use photoshop extensively and multitask when i use it, this type of performance will not cut it unless it can be fixed via software updates due to code not working well in vista with ps

How do you know that Anand didn't get a payoff to help stimulate more ram sales? For most people it is possible to run Vista quite comfortably with 2GB.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Wow, obvious and immediate? That's not how I remember it...in fact the same lame arguments you are posing now were commonplace back in the Win 3.1 to Win95 transition. You forgot, or is this before your time? And there is quite a bit of open source software available for Windows...maybe you need to expand your searches-for-smartness to include sourceforge.net =)

Yes, going from max memory of 16M to 4G was definitely obvious and immediate, as was the addition of protected mode in the 386. And I do know that a lot of OSS software compiles and runs on Windows, I used to use a good bit of it at work.

If you have to start the sentence with "unless" then it just means you're going to fabricate a condition to make it seem like you could have been right originally. You don't need pointers, however not using them is bad coding practice.

Starting a sentence with "unless" merely indicates that I do know that there are situtations where the opposite can be true but that I find it unlikely. The only time you'll have code without pointers is if you're using a language that abstacts that functionality away from you, but they'll still be used by the low level libraries.

No I haven't, but I'm not going to sit here and pretend I do. Your comment along the lines of: "SMP has been around for decades, but developers still haven't figured out how to use it" pretty much solidifies you as clueless.

I take it that you didn't see Carmack's recent comments on how the PS3 is a huge PITA to develop for because it pretty much requires multithreaded code to get good performance? And how he also dislikes the fact that Intel and AMD have stopped the Mhz race in favor of jacking up the number of cores in the chips for the same reason? My last 3 machines have all been SMP an I can count the number of multithreaded programs I've used on all of them on two hands. If multithreaded programming is as easy as you seem to think it is, why hasn't be become more popular?

On obsolete or underpowered hardware, perhaps...but in general, I don't don't see many startups eagerly rushing in to compete with MS on the OS front so it's not like your choices are plentiful.

Maybe your choices are limited, but mine aren't and I'm perfectly happy running 64-bit Linux.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: Nothinman
Yes, going from max memory of 16M to 4G was definitely obvious and immediate, as was the addition of protected mode in the 386. And I do know that a lot of OSS software compiles and runs on Windows, I used to use a good bit of it at work.

Oh, so back when Win95 was introduced you (and presumably MOST people) upgraded to 4 GB of ram? Yes, that would mean people had memory that was 3-4 times more than the biggest hard drives back then. Seriously, WTF? Go ahead and try to boot Win95 with 4GB or ram in your system...........................let us know how that works out for you.

Protected mode - guess what - uses more memory...back in a time when most people thought 16MB was "more than enough for anything". What does that mean? Slow performance, generally unresponsive OS compared to Win3.1 which was pretty "snappy". You also seem to forget that it wasn't until XP that 32-bit Windows was able to do most of the things we're used to. Windows 95 the 98SE were pretty much still GUIs for DOS.

Speaking of DOS, the DOS4GW 32-bit protected mode extender was around quite some time before Win95 hit the market.

Starting a sentence with "unless" merely indicates that I do know that there are situtations where the opposite can be true but that I find it unlikely. The only time you'll have code without pointers is if you're using a language that abstacts that functionality away from you, but they'll still be used by the low level libraries.

Most modern languages are high level and you really don't know what's going on under the hood...but until you show me some programs you've written, you should really stop talking about something you have no clue about.

I take it that you didn't see Carmack's recent comments on how the PS3 is a huge PITA to develop for because it pretty much requires multithreaded code to get good performance? And how he also dislikes the fact that Intel and AMD have stopped the Mhz race in favor of jacking up the number of cores in the chips for the same reason? My last 3 machines have all been SMP an I can count the number of multithreaded programs I've used on all of them on two hands. If multithreaded programming is as easy as you seem to think it is, why hasn't be become more popular?

You just don't give up do you? Carmack complains so you automatically assume he speaks for everyone? First of all, a lot the games his company developed are multithreaded. The PS3 is not the PC, and you are taking his comments way out of context. The PS3 archictecture is quite different from PC, and developing for it is DIFFERENT. That means he and his developers will need to learn something new, and learning something new in business generally translates into spending (investing?) money, not making money.

I'm glad you had 3 SMP machines...but I doubt you represent any major demographic of power users, enterprise user or anyone who would require or be able to utilize the benefits of SMP. I never said it was easy, you're just trying to pass it off as something only 2 people on the planet can do. Keep in mind that multi-threading isn't really beneficial on programs that are not doing more than one thing at a time...which is why it's doubtful you'd see a multi-threaded version of Notepad.

Maybe your choices are limited, but mine aren't and I'm perfectly happy running 64-bit Linux.

I can't think of any reason, even a bad reason, that you would be running Linux...nevermind that it is 64-bit linux. Make sure you defrag it regularly with your 64-bit multi-threaded defrag program.
 

nweaver

Diamond Member
Jan 21, 2001
6,813
1
0
Originally posted by: EricMartello
Originally posted by: Nothinman
Yes, going from max memory of 16M to 4G was definitely obvious and immediate, as was the addition of protected mode in the 386. And I do know that a lot of OSS software compiles and runs on Windows, I used to use a good bit of it at work.

Oh, so back when Win95 was introduced you (and presumably MOST people) upgraded to 4 GB of ram? Yes, that would mean people had memory that was 3-4 times more than the biggest hard drives back then. Seriously, WTF? Go ahead and try to boot Win95 with 4GB or ram in your system...........................let us know how that works out for you.

Protected mode - guess what - uses more memory...back in a time when most people thought 16MB was "more than enough for anything". What does that mean? Slow performance, generally unresponsive OS compared to Win3.1 which was pretty "snappy". You also seem to forget that it wasn't until XP that 32-bit Windows was able to do most of the things we're used to. Windows 95 the 98SE were pretty much still GUIs for DOS.

Speaking of DOS, the DOS4GW 32-bit protected mode extender was around quite some time before Win95 hit the market.

Starting a sentence with "unless" merely indicates that I do know that there are situtations where the opposite can be true but that I find it unlikely. The only time you'll have code without pointers is if you're using a language that abstacts that functionality away from you, but they'll still be used by the low level libraries.

Most modern languages are high level and you really don't know what's going on under the hood...but until you show me some programs you've written, you should really stop talking about something you have no clue about.

I take it that you didn't see Carmack's recent comments on how the PS3 is a huge PITA to develop for because it pretty much requires multithreaded code to get good performance? And how he also dislikes the fact that Intel and AMD have stopped the Mhz race in favor of jacking up the number of cores in the chips for the same reason? My last 3 machines have all been SMP an I can count the number of multithreaded programs I've used on all of them on two hands. If multithreaded programming is as easy as you seem to think it is, why hasn't be become more popular?

You just don't give up do you? Carmack complains so you automatically assume he speaks for everyone? First of all, a lot the games his company developed are multithreaded. The PS3 is not the PC, and you are taking his comments way out of context. The PS3 archictecture is quite different from PC, and developing for it is DIFFERENT. That means he and his developers will need to learn something new, and learning something new in business generally translates into spending (investing?) money, not making money.

I'm glad you had 3 SMP machines...but I doubt you represent any major demographic of power users, enterprise user or anyone who would require or be able to utilize the benefits of SMP. I never said it was easy, you're just trying to pass it off as something only 2 people on the planet can do. Keep in mind that multi-threading isn't really beneficial on programs that are not doing more than one thing at a time...which is why it's doubtful you'd see a multi-threaded version of Notepad.

Maybe your choices are limited, but mine aren't and I'm perfectly happy running 64-bit Linux.

I can't think of any reason, even a bad reason, that you would be running Linux...nevermind that it is 64-bit linux. Make sure you defrag it regularly with your 64-bit multi-threaded defrag program.

Wow, craptastic nested quoting ftw


So, Nothinman has a rep around here. you are the noob, so I think if anyone needs to throw down what they have worked on, it's you. What oh so wonderfull app have you worked on..lets see some source if you are all high and mighty.

Oh, and bringing up "OMG!!1! the PS3 is multithreaded!!" is pointless when talking about PC's. Name 3 games that are SMP aware (only one I know if is Q3, irrc). On top of that, name 2 that got better performance.


Oh, and why run linux64...because it works...it works well. 64 bit doesn't suck, just Windows 64.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: nweaver
Wow, craptastic nested quoting ftw


So, Nothinman has a rep around here. you are the noob, so I think if anyone needs to throw down what they have worked on, it's you. What oh so wonderfull app have you worked on..lets see some source if you are all high and mighty.

Oh, and bringing up "OMG!!1! the PS3 is multithreaded!!" is pointless when talking about PC's. Name 3 games that are SMP aware (only one I know if is Q3, irrc). On top of that, name 2 that got better performance.


Oh, and why run linux64...because it works...it works well. 64 bit doesn't suck, just Windows 64.

The nesting isn't up to your high standards huh? Well if you don't like it, tell resident genius Anand to upgrade to Vbulletin like every other respectable forum out there, instead of using this Coldfusion based junk that his friend made and nobody else uses. =)

Anyway, I'll dumb it down for you because I Don't want you to have an aneurysm:

- I did not bring up the PS3, my friend Nothinman did that...in essence, I agree with you on that point.

- Why would I bother to show source to someone like you...I'd have better luck trying to convince a chimp not to fling his crap all over the place.

- Quake 3, Quake 4, Doom 3, Return to Wolfenstein, Star Trek Elite Force, IL-2 Sturmovik...and if you dig around, I'm sure you can find more - and they all experienced better performance. But you know what? There really has been no incentive to make games multithreaded when 99% of the users had a single processor. Now with CPUs that have 2 or more cores becoming the standard, THERE IS A REASON TO MULTITHREAD. It's simply supply and demand, not a question of difficulty.

- V1st4 5uX d00d!!!!!!!!!! u R a n00b!!!!!!!!!!!!!!!!!!! 128-b1tZ i5 d4 sh1t!!!!!!!!!!!! Seriously, you're clearly not l33t if you're gonna sit there and judge people based on their post counts, I mean you have way more than me and you don't have a clue, you can barely read what is written without messing it up. Maybe you should stop posting and do some learning.
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
The Quake3 multithreaded version was scrapped a long time ago. Id released several updates that didn't include any smp support at all. Killed it dead, they did.

If you saw a performance increase in other games it was probably because a lot of the overhead involving sound drivers, file system drivers, as well as other OS stuff was offloaded to another proccessor. Dollar for Dollar you usually would see better gaming benchmark results by spending in on a faster single core CPU rather then 2 slightly slower CPUs.

The thing is, yes, multithreading is much more difficult way to code. However tools are increasing in sophistication. A lot of the initial programming difficulty people were expecting with PS3 gaming was gone by the fact that you'd program against familar-ish APIs like OpenGL which had their libs speciifcly optimized by Sony for the PS3.

If your a programmer (and thus obviously know much more about this then me) you can check out http://www.cellperformance.com/articles/ for code snippets and such from developers for the Cell proccessor. It's using things like GCC and other standard Linux tools, some of the people on the site are actual professional console game programmers. (it's certaintly different from programming for a regular single core x86 system... but it's not _that_ different)

Part of the reason why people were worried is because with PS1 and PS2 all the code had to be very custom to get the best performance out of it. Did you know that the PS1 and PS2 ran with completely software driven 3d performance?
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
The SMP support in Quake 3 was experimental, but with the Doom 3 engine SMP gains were a bit better. Multithreading applications does add a level of complexity to a program, and that translates to added cost. Game developers really had no financial incentive to support SMP because most of their customers did not use multi-processor computers. That is now changing with the rapid influx of multi-core processors...and as you mentioned, the programming tools will most likely be refined to improve the efficiency of coding multithreaded apps.

I'm not sure but I belive the PS1 and PS2 used a form of "accelerated emulation". The hardware in both systems was geared toward processing the data, however the "3D Driver" contained the features that could be accessed by the API. This was mostly good because it allowed skilled developers to push the limits of the consoles...you can really see the improvements when you compare a game like FFVII to FFIX on the PS1, and on the PS2 it didn't have bilinear filtering on a lot of the early titles, giving them a really jaggy appearance. This was addressed later on.
 

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
EricMartello, I'm a mere sysadmin, so I know very little about programming(I did make a bouncing ball screen saver in C when I was 11 or so, but that was a long time ago), but you're not making a very good case for yourself by insulting and belittling others.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: Sunner
EricMartello, I'm a mere sysadmin, so I know very little about programming(I did make a bouncing ball screen saver in C when I was 11 or so, but that was a long time ago), but you're not making a very good case for yourself by insulting and belittling others.

That's great little boy!!! Now that you're 12, why don't you start working on your next screen saver - a blank screen...
 

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
Originally posted by: EricMartello
Originally posted by: Sunner
EricMartello, I'm a mere sysadmin, so I know very little about programming(I did make a bouncing ball screen saver in C when I was 11 or so, but that was a long time ago), but you're not making a very good case for yourself by insulting and belittling others.

That's great little boy!!! Now that you're 12, why don't you start working on your next screen saver - a blank screen...

Ok then.
A troll, I actually thought you were serious for a while.
Now shoo, back to OT and P&N you go.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Oh, so back when Win95 was introduced you (and presumably MOST people) upgraded to 4 GB of ram? Yes, that would mean people had memory that was 3-4 times more than the biggest hard drives back then. Seriously, WTF? Go ahead and try to boot Win95 with 4GB or ram in your system...........................let us know how that works out for you.

Yea sure, just like you went out and bought 1TB of memory for your 64-bit system as soon as you bought it, right?

Protected mode - guess what - uses more memory...back in a time when most people thought 16MB was "more than enough for anything". What does that mean? Slow performance, generally unresponsive OS compared to Win3.1 which was pretty "snappy". You also seem to forget that it wasn't until XP that 32-bit Windows was able to do most of the things we're used to. Windows 95 the 98SE were pretty much still GUIs for DOS.

Protected mode doesn't use more memory it just changes the way that memory is addressed so that VM, paging, context switching, etc worked. And NT did "most of the things we're used to" just fine back when I had it installed.

Speaking of DOS, the DOS4GW 32-bit protected mode extender was around quite some time before Win95 hit the market.

And? You're the one that brought up Win95.

Most modern languages are high level and you really don't know what's going on under the hood...but until you show me some programs you've written, you should really stop talking about something you have no clue about.

Well I would consider C pretty popular and high level and yet you still have to deal with pointers yourself. But even that's beside the point, you keep insisting that if you use a language that hides them from you (C#, perl, Java, etc) that pointers magically don't exist in your program which is so untrue it's not even funny.

You just don't give up do you? Carmack complains so you automatically assume he speaks for everyone?

No and generally he's considered one of the best game programmers of our time so I figured that you might actually listen to him since you're blowing me off.

First of all, a lot the games his company developed are multithreaded.

But only at a very small level, doing things like audio decoding, network I/O, etc in another thread. Splitting up the rendering or AI into multiple threads is much, much more difficult. And Q3's SMP mode was disabled in the last patch because it was so unstable, I wouldn't exactly call that a good example.

The PS3 is not the PC, and you are taking his comments way out of context. The PS3 archictecture is quite different from PC, and developing for it is DIFFERENT.

But his comments are still appropriate because he was specifically complaining about multiple cores which is what's happening on the desktop too, just right now we've only got 2 or 4 cores usually.

I'm glad you had 3 SMP machines...but I doubt you represent any major demographic of power users, enterprise user or anyone who would require or be able to utilize the benefits of SMP.

I never said I wasn't a niche user. But I do find it funny that you think I'm obviously not able to use the 2 cores in my machine even though you want "regular users" to have 64-bit machines with 2, 4, 8 cores so that they can browse the web and play Blast Billiards.

I never said it was easy, you're just trying to pass it off as something only 2 people on the planet can do. Keep in mind that multi-threading isn't really beneficial on programs that are not doing more than one thing at a time...which is why it's doubtful you'd see a multi-threaded version of Notepad.

No you never did say that, but you implied it because you called me an idiot for saying that it's difficult. But threading is extremely difficult and while I'm sure more than 2 people can successfully do it, I would guess that the real number is a pretty low percentage of developers. Especially if we include Windows "developers" that have only ever used crap like VB in that list since they've probably never even heard words like spinlock or mutex before.

Oh, and bringing up "OMG!!1! the PS3 is multithreaded!!" is pointless when talking about PC's. Name 3 games that are SMP aware (only one I know if is Q3, irrc). On top of that, name 2 that got better performance.

The only one I can think of is the UT series and AFAIK all they did was put the ogg audio decoding in another thread but it did help a bit since ogg decoding is fairly expensive.

- Quake 3, Quake 4, Doom 3, Return to Wolfenstein, Star Trek Elite Force, IL-2 Sturmovik...and if you dig around, I'm sure you can find more - and they all experienced better performance. But you know what? There really has been no incentive to make games multithreaded when 99% of the users had a single processor. Now with CPUs that have 2 or more cores becoming the standard, THERE IS A REASON TO MULTITHREAD. It's simply supply and demand, not a question of difficulty.

Q3's SMP code sucked and was broken with the last patch and if Doom3 is multithreaded it's not doing it very well because I couldn't get it to go over 100% of 1 CPU here, same thing with ET which is based on RTCW. I even ran a doom3 timedemo with 1 CPU online and with 2 CPUs online and the outcome was almost exactly the same in both cases, not exactly what I would call a shining example of a multithreaded game.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: Nothinman
Yea sure, just like you went out and bought 1TB of memory for your 64-bit system as soon as you bought it, right?

Why would I buy 1 TB? Are you somehow suggesting that memory access is the only benefit of a 64-bit system, which you have already admitted to be using? No, that was a response to your comment about IMMEDIATE benefits of going from Win3.1 to Win95 - to which I said the benefits were there, but just as it is now with 64-bit, it's not IMMEDIATE.

Protected mode doesn't use more memory it just changes the way that memory is addressed so that VM, paging, context switching, etc worked. And NT did "most of the things we're used to" just fine back when I had it installed.

Let's think about this...protected mode came into existence to allow programs to address more memory, and it also enabled the use of virtual memory. BUT GUESS WHAT? The memory segments are now PROTECTED - read that as isolated from each other. This means that each task gets its own uniquely addressable space, which cannot be accessed by other programs. In other words you may need multiple copies of the same thing in memory for each instance of the program that needs said access to said data.

Speaking of DOS, the DOS4GW 32-bit protected mode extender was around quite some time before Win95 hit the market.

And? You're the one that brought up Win95.

Yeah but you must know that DOS4GW is not an operating system, it merely allows DOS programs which are 16-bit to have 32-bit protected mode access to memory. The reason I mentioned this? Because there has always been a need to move forward, even when everybody was comfortable with their 16-bit software running on essentially a 32-bit processor...the software was behind the times, just like it is now. We have 64-bit processors but most people are stuck on 32-bit, and that's going to change soon.

Well I would consider C pretty popular and high level and yet you still have to deal with pointers yourself. But even that's beside the point, you keep insisting that if you use a language that hides them from you (C#, perl, Java, etc) that pointers magically don't exist in your program which is so untrue it's not even funny.

Are we defining the parameters of a program here, or just arbitrarily taking about languages. C is high level, it's not really OO. C++ is the best choice for all-around coding, in my opinion.

The necessity of pointers really depends on WHAT you are coding. Up to a certain point you can write really inefficient code that avoids using them, resulting in excess memory usage and most likely a slow-running app that is unstable. Your blanket statement that ALL programs in C++ (just so we're at least on the same page as far as language) need pointers is absolutely incorrect.

No and generally he's considered one of the best game programmers of our time so I figured that you might actually listen to him since you're blowing me off.

Didn't carmack also tout Nvidia and downplay the Radeon for quite some time? I respect him for what he did, and what he can do, but like the rest of us he's still human and if he wants to vent and rant, let him. Maybe you want to take his rants as gospel, but he's not the only talented game programmer currently in existence, and you can't blame him for wanting to let off some steam.

But only at a very small level, doing things like audio decoding, network I/O, etc in another thread. Splitting up the rendering or AI into multiple threads is much, much more difficult. And Q3's SMP mode was disabled in the last patch because it was so unstable, I wouldn't exactly call that a good example.

Yes yes, it's quite obvious that Carmack was wise enough not to divert the bulk of his resources to make Q3 a fully-efficient SMP program. He must have had access to inside info - that most consumers at that time had single CPU, single CORE processors. But Carmack didn't just do that for the hell of it, he and most other forward-thinking programmers realized that SMP is, in the near future, going to be a mainstream thing, and that it's best to get an early jump in order to stay ahead of the competition.

The PS3 is not the PC, and you are taking his comments way out of context. The PS3 archictecture is quite different from PC, and developing for it is DIFFERENT.

But his comments are still appropriate because he was specifically complaining about multiple cores which is what's happening on the desktop too, just right now we've only got 2 or 4 cores usually.
[/quote]

You say complaining, I say venting. It's unlikely that he will toss in his hat as a programmer because multi-core CPUs are becoming the norm. Once he gets used to it, we should see some awesome new game from id.

I never said I wasn't a niche user. But I do find it funny that you think I'm obviously not able to use the 2 cores in my machine even though you want "regular users" to have 64-bit machines with 2, 4, 8 cores so that they can browse the web and play Blast Billiards.

Well this is a thread about Vista, and the original start of our lively chat was you stating that Vista is slower, to which I agreed, assuming we are talking about running it on inadequate hardware. The point here is that Vista is designed for tomorrow's hardware...not that budget crap a lot of people on here keep slopping together, excited that they saved a few hundred bucks by overclocking instead of just buying the faster CPU.

No you never did say that, but you implied it because you called me an idiot for saying that it's difficult. But threading is extremely difficult and while I'm sure more than 2 people can successfully do it, I would guess that the real number is a pretty low percentage of developers. Especially if we include Windows "developers" that have only ever used crap like VB in that list since they've probably never even heard words like spinlock or mutex before.

I never called you an idiot, but I guess you realized that just by reading my response, at least in the context of your statement. Windows developers who use VB aren't really coders, they're the people who come out of DeVry and get all happy about their $35K a year position spilling burnt coffee on keyboards and posting on forums.

Your original statement was playing up SMP like some sort of elite form of programming, and granted you need real skill to do it properly, it is not as rare or exclusive as you seem to think it is. Globally, there are many skilled and talented programmers who can produce quality multi-threaded software.

Oh, and bringing up "OMG!!1! the PS3 is multithreaded!!" is pointless when talking about PC's. Name 3 games that are SMP aware (only one I know if is Q3, irrc). On top of that, name 2 that got better performance.

The only one I can think of is the UT series and AFAIK all they did was put the ogg audio decoding in another thread but it did help a bit since ogg decoding is fairly expensive.Q]

That wasn't my statement, and I already listed several SMP-enabled games, so don't wreck your brain thinking.

Q3's SMP code sucked and was broken with the last patch and if Doom3 is multithreaded it's not doing it very well because I couldn't get it to go over 100% of 1 CPU here, same thing with ET which is based on RTCW. I even ran a doom3 timedemo with 1 CPU online and with 2 CPUs online and the outcome was almost exactly the same in both cases, not exactly what I would call a shining example of a multithreaded game.

Why are you telling me like it's something I made? Tell your friend Mr. Carmack that his SMP implementations suck. You may also want to note that the games here were not originally coded to be SMP, so making them "SMP Enabled" requires extensive re-engineering. They will only "play" with the code so far, after a certain point it just takes up too much time that they need to be spending on new products. If the game was, from the start, designed with SMP-support in mind, rest assured there would be significant gains in performance.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
Why would I buy 1 TB? Are you somehow suggesting that memory access is the only benefit of a 64-bit system, which you have already admitted to be using? No, that was a response to your comment about IMMEDIATE benefits of going from Win3.1 to Win95 - to which I said the benefits were there, but just as it is now with 64-bit, it's not IMMEDIATE.

Because you said "Oh, so back when Win95 was introduced you (and presumably MOST people) upgraded to 4 GB of ram?" since 4G was the new limit when 32-bit CPUs were introduced. And yes, it's the only tangible one. The extra GPRs in long mode will help a little bit, but they're seperate from the CPU's 64-bit-ness.

Let's think about this...protected mode came into existence to allow programs to address more memory, and it also enabled the use of virtual memory. BUT GUESS WHAT? The memory segments are now PROTECTED - read that as isolated from each other. This means that each task gets its own uniquely addressable space, which cannot be accessed by other programs. In other words you may need multiple copies of the same thing in memory for each instance of the program that needs said access to said data.

So you really think that there's a hundred copies of MFC.dll in memory on your machine right now? If you really think that's true I suggest you get a copy of Inside Windows or Understanding the Linux kernel and read the memory management chapters. Physical addresses can be mapped into multiple process's address space and they're handled by COW if one process decides to modify them.

Are we defining the parameters of a program here, or just arbitrarily taking about languages. C is high level, it's not really OO. C++ is the best choice for all-around coding, in my opinion.

I never that C was OO and being OO has nothing to do with the ability of a language to use pointers or not.

The necessity of pointers really depends on WHAT you are coding.

Unless your code just consists of something so simple as "x = 2 +2; return 0;" there'll almost certainly be pointers uses even they're hidden from you by some library functions.

Didn't carmack also tout Nvidia and downplay the Radeon for quite some time?

If he did, maybe that's what kicked ATI in the ass and made them get better? Frankly I still hate ATI because their Linux drivers suck so I have no idea how their hardware compares anymore.

Maybe you want to take his rants as gospel, but he's not the only talented game programmer currently in existence, and you can't blame him for wanting to let off some steam.

But you can't deny that he is a hugely talented programmer so if he says something is difficult to program it almost certainly is so. If he came out in an interview and said that orange juice was bad for you of course I wouldn't just run out and throw away all of my orange juice because he has no medical background but in this case he is talking about something he knows.

You say complaining, I say venting. It's unlikely that he will toss in his hat as a programmer because multi-core CPUs are becoming the norm. Once he gets used to it, we should see some awesome new game from id.

I'm not saying that they won't be able to do it, but it'll cause developement time to increase a lot and most likely introduce a lot more bugs since debugging things like thread locking and catching race conditions between threads is a huge PITA.

I never called you an idiot, but I guess you realized that just by reading my response, at least in the context of your statement.

That's funny, you can read more into my text and say I implied something that I didn't but when I do the same thing it's just a realization on my part.

Your original statement was playing up SMP like some sort of elite form of programming, and granted you need real skill to do it properly, it is not as rare or exclusive as you seem to think it is. Globally, there are many skilled and talented programmers who can produce quality multi-threaded software.

Then where is all of it? Why is it all relegated to niche things like databases, A/V encoders, etc?

Tell your friend Mr. Carmack that his SMP implementations suck.

He already knows that since he disabled it in Q3 and as far as I can tell the Doom3 SMP code doen't actually do anything performance-wise.

You may also want to note that the games here were not originally coded to be SMP, so making them "SMP Enabled" requires extensive re-engineering.

Doom3 wasn't designed to be multithreaded from the beginning?
 

bsobel

Moderator Emeritus<br>Elite Member
Dec 9, 2001
13,346
0
0
Ok, need to drop a couple comments...

If multithreaded programming is as easy as you seem to think it is, why hasn't be become more popular?

Simple, we are on about year 2 of the real SMP rollout. You'll laugh, but it's been hyperthreading which has accelerated SMP more than any other technology. You and I have had SMP desktop for years, and yes few desktop applications took advantage of it. Why? Because you and I had SMP desktops, not our parents, our friends, and our coworkers. We've had SMP in the datacenter for years, and name me a data center applications (web tier, database tier, etc) which isn't accelerated by SMP.

On the desktop side few developers would code for SMP because so little of their market had boxes that would use it. The value trade off wasn't there. I remember the first NDA discussions with Intel on HT, they were pushing this aspect and that aspect, but my takeaway was 'over the next 12 months we will fix more SMP bugs in consumer software than have been fixed in total over the prior decade'. As an industry we did (processor affinity rarely is an issue on consumer software today).

But, we are only 2 years in and most development houses code for the most common configs. HT and now multi-core is driving SMP mainstream, mainstream enough that people will be building and expecting it not 'just supporting it'.

The amount of cores is orthognal to whether you run a 64-bit OS or not, you can run your quad core CPU in 32-bit mode and still get all 4 cores working for you.

Your statement is logically correct but discounts other factors which will drive multi-core development. For example, there will become a minimum amount of memory people want to have available per core (yes, its 'all' available, but in a very simple fashion having 2 cores with 4gig is one thing, but having 32 cores with 4 gig is quite another).

My 'on order' workstation (ugh, I wish Dell would hurry up, Supreme Commander is almost here ) is a dual-quad core xeon with 32gigs. Those 8 cores 'overrun' (in my mind) a 4 gig maximum. While I tend to be bleeding edge, those types of machines will be mainstream in 18-24 months and that is exactly what will drive 64bit adoption.

Bill
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
On the desktop side few developers would code for SMP because so little of their market had boxes that would use it. The value trade off wasn't there.

And for most developers the tradeoff value still won't be there in 5 years because unless your app is CPU bound spliting the work up into threads won't buy you anything. Do you really think Adobe's doing to multithread acroread just because multicore systems are becoming popular?

While I tend to be bleeding edge, those types of machines will be mainstream in 18-24 months and that is exactly what will drive 64bit adoption.

I highly doubt that 32G will be anywhere near mainstream in 24 months. We're still at the point where 1G is average and 4G is extreme, those numbers aren't going to go up 8 fold in 2 years unless memory gets a helluva lot cheaper and Vista64 eclipses Vista32's adoption.

And that spurs another question, what is Vista64's memory limit? I believe Vista32's still limited to 4G for some reason but I know XP64 is higher so it makes sense that Vista64 would have the same limit as XP64.
 

bsobel

Moderator Emeritus<br>Elite Member
Dec 9, 2001
13,346
0
0
I highly doubt that 32G will be anywhere near mainstream in 24 months. We're still at the point where 1G is average and 4G is extreme, those numbers aren't going to go up 8 fold in 2 years unless memory gets a helluva lot cheaper and Vista64 eclipses Vista32's adoption.

My point was more around the push in cores will start pushing us clearly past 4gig (I tend to think 2gig per core is where things will average out for higher end consumer machines). So I'd expect to see around 8 gig on the quads at that time (16 on the machine I'm getting, but native oct will outpace dual-quad I think).

And that spurs another question, what is Vista64's memory limit? I believe Vista32's still limited to 4G for some reason but I know XP64 is higher so it makes sense that Vista64 would have the same limit as XP64.

8gig as I recall for home basic, 16gig for home premimum and 128+ for the higher end versions.

Bill

 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
My point was more around the push in cores will start pushing us clearly past 4gig (I tend to think 2gig per core is where things will average out for higher end consumer machines). So I'd expect to see around 8 gig on the quads at that time (16 on the machine I'm getting, but native oct will outpace dual-quad I think).

That still seems a little ambitious since 2G chips are still in the $500-$600 range. 2G and 4G aren't uncommon because you can get 1G chips for like a $100 a piece, but I doubt many people will be willing to spend $2K for 8G of memory when there's no real reason yet. And I've already seen several people say they're just going to stick with 1G or 2G and get a cheap flash drive for ReadyBoost since it cuts down on the paging penalty.
 

Doom Machine

Senior member
Oct 23, 2005
346
0
0
i have 2 gigs in vista, it didnt use pagefile so i turned it off and it still doesnt use readyboost even after playing fear and obliviion, so with 2gigs i wouldnt even bother with a usb flash just yet, its memory management is excellent and makes room for higher priority things like gaming
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: Nothinman
Because you said "Oh, so back when Win95 was introduced you (and presumably MOST people) upgraded to 4 GB of ram?" since 4G was the new limit when 32-bit CPUs were introduced. And yes, it's the only tangible one. The extra GPRs in long mode will help a little bit, but they're seperate from the CPU's 64-bit-ness.

New limit? Did you realize that since the 386 processor, CPUs were essentially 32-bit and capable of addressing larger amounts of memory. The issue here never was the hardware, it was the software. DOS was 16-bit, Win3.1 wasn't even a real OS...Windows95 was 32-bit but you know what, it wouldn't boot if you put 1GB of memory in your system. Keep it in the context here, the transition from DOS/Win3.1 to WIndows 95 (the first mainstream 32-bit OS), was shaky and not "immediate", as it is now with the 64-bit transition.

So you really think that there's a hundred copies of MFC.dll in memory on your machine right now? If you really think that's true I suggest you get a copy of Inside Windows or Understanding the Linux kernel and read the memory management chapters. Physical addresses can be mapped into multiple process's address space and they're handled by COW if one process decides to modify them.

Do you know what a DLL is? It's a SHARED LIBRARY. That means that programs can load them as required, or the DLL itself can be loaded into memory and made available to any program which needs it. Protected mode memory means that the data being stored at a given memory address for a given app or process cannot be accessed by other processes or apps. And yes, you can program your app in such a way that it poaches another instance's protected memory, and this is how you end up with app instability.

I never that C was OO and being OO has nothing to do with the ability of a language to use pointers or not.

Unless your code just consists of something so simple as "x = 2 +2; return 0;" there'll almost certainly be pointers uses even they're hidden from you by some library functions.

Like I said, show me some programs you've written since you're so into talking about it like your Carmack's apprentice or something.

If he did, maybe that's what kicked ATI in the ass and made them get better? Frankly I still hate ATI because their Linux drivers suck so I have no idea how their hardware compares anymore.

No, I think ATI was already better at the time Carmack was saying Nvidia is the sh1t. The GeForce hasn't been better than the Radeon since the Radeon was released. We'll see if the 8800 will be eclipsed by ATI's new GPU. Linux drivers...don't care much because you really only need 2D in linux. Gaming is best saved for Windows, and the Radeon drivers for Vista are currently a lot better than Nvidia's.

But you can't deny that he is a hugely talented programmer so if he says something is difficult to program it almost certainly is so. If he came out in an interview and said that orange juice was bad for you of course I wouldn't just run out and throw away all of my orange juice because he has no medical background but in this case he is talking about something he knows.

I'm not saying that they won't be able to do it, but it'll cause developement time to increase a lot and most likely introduce a lot more bugs since debugging things like thread locking and catching race conditions between threads is a huge PITA.

I did not deny that, did I? Difficult or not, it will be done AND it has been done. Even he has to learn new things every now and then.

Right now development tools are not optimized for SMP programming, however they will be soon, and there will most likely be wrappers and APIs developed that dramatically ease the development of a proper multi-threaded app. Besides, you are not going to be the one doing any programming so why are you so worried?


Then where is all of it? Why is it all relegated to niche things like databases, A/V encoders, etc?

Tell your friend Mr. Carmack that his SMP implementations suck.

Did you notice that the things that use SMP now are mainly software apps that cost hundreds, if not thousands of dollars? What does SQL server go for? $25k? What does a studio-quality AV app go for...like Avid...guessing around $10K. Why do they cost so much? Because they make use of non-mainstream technology to provide substantial performance improvements. The reason SMP enabled games are few, and poorly implement is an ECONOMIC issue, not a technical issue. Now, that is changing...and now, developers who can utilize SMP in their games can potentially create much better games.

He already knows that since he disabled it in Q3 and as far as I can tell the Doom3 SMP code doen't actually do anything performance-wise.

Doom3 wasn't designed to be multithreaded from the beginning?

Games are getting more complex, and there's no doubt that SMP can improve their performance if it is implemented effectively. Even if the game itself is SMP-enabled, the sound and video card drivers, as well as most supporting APIs are not. Hence, the minimal gains.

Do you think the Doom 3 game client was originally designed to be SMP from the ground up? Really?
 

drag

Elite Member
Jul 4, 2002
8,708
0
0
Did you notice that the things that use SMP now are mainly software apps that cost hundreds, if not thousands of dollars? What does SQL server go for? $25k?

There are quite a few SQL 'servers' that go for 0 to several thousand dollars. MySQL, which is the defacto-standard SQL for web services, is multithreaded. Same thing with Apache.


What does a studio-quality AV app go for...like Avid...guessing around $10K.

Cinelerra, Free software non-linear and composition edition is fully multithreaded. As is plenty of other applications, since they need it for their paticular nitch.

The reason SMP enabled games are few, and poorly implement is an ECONOMIC issue, not a technical issue. Now, that is changing...and now, developers who can utilize SMP in their games can potentially create much better games.

Nope it's technical AND economical. It's simply more expensive to do and would reduce performance on the majority of people's computers if it was multithreaded.

BTW multithreaded != faster.

It _can_, but more complex the applications the more difficult it is to get good performance. There are all sorts of locking issues and all that crap people have to deal with threaded apps that they don't have to deal with on single threaded applications. All this added complexity is likely result in utlimately reduced performance and much increased bugginess.

Games are getting more complex, and there's no doubt that SMP can improve their performance if it is implemented effectively. Even if the game itself is SMP-enabled, the sound and video card drivers, as well as most supporting APIs are not. Hence, the minimal gains.

All the more reason for multithreading to be difficult.

To be fair though it's very unusual for game makers nowadays to develop their own rendering system. They generally license it from people like Id, Valve, or Epic games. Once people start adopting the U3 engine, the next engines coming out from Valve and Id, then they will be multihreaded rendered by default.
 

EricMartello

Senior member
Apr 17, 2003
910
0
0
Originally posted by: drag
There are quite a few SQL 'servers' that go for 0 to several thousand dollars. MySQL, which is the defacto-standard SQL for web services, is multithreaded. Same thing with Apache.

Cinelerra, Free software non-linear and composition edition is fully multithreaded. As is plenty of other applications, since they need it for their paticular nitch.

There exists a freeware counterpart to almost every commercial app, but that's not what I'm talking about. You're comparing apples to oranges...do you think that if the freeware apps were being developed by paid programmers, they'd be any cheaper? The answer is no.

Nope it's technical AND economical. It's simply more expensive to do and would reduce performance on the majority of people's computers if it was multithreaded.

BTW multithreaded != faster.

It _can_, but more complex the applications the more difficult it is to get good performance. There are all sorts of locking issues and all that crap people have to deal with threaded apps that they don't have to deal with on single threaded applications. All this added complexity is likely result in utlimately reduced performance and much increased bugginess.

A program that is poorly written will be slow and unstable, regardless of whether it is multi-threaded or not. You really can't say that simply because you need more skill to do something, it is impractical or not worth doing.

The fact of the matter is, there are many apps currently in existence that run reliably on multiple CPUs/Cores, and now that multi-core CPUs are creating a mainstream demand for SMP applications, you will start to see more of them.

All the more reason for multithreading to be difficult.

To be fair though it's very unusual for game makers nowadays to develop their own rendering system. They generally license it from people like Id, Valve, or Epic games. Once people start adopting the U3 engine, the next engines coming out from Valve and Id, then they will be multihreaded rendered by default.

You all like to keep saying SMP is so difficult, and none of you are programmers...but that aside, anything worth doing isn't going to be easy...and that is something you can bank on.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |