December 28, 2008
"Andrei Alexandrescu" <SeeWebsiteForEmail@erdani.org> wrote in message news:gj7591$2tec$1@digitalmars.com...
> Nick Sabalausky wrote:
>> "Andrei Alexandrescu" <SeeWebsiteForEmail@erdani.org> wrote in message news:gj6mds$28iv$1@digitalmars.com...
>>> Nick Sabalausky wrote:
>>>> Ah ha, there's that usual "if you go and buy a PC" catch. Which begs the question, why would I? My existing system does everything I need it to do perfectly fine. And since I'm not petty enough to allow anyone to shame me into buying a new system just by calling my *current* system "legacy", that leaves no real reason for me to buy a new one.
>>> I agree that often there is little incentive to upgrade. In particular incentive can be negative when it comes to Vista vs. XP.
>>>
>>
>> I'm incredibly jealous of how Vista only highlights the filename (minus suffix) when you go to rename a file. I *really* want that. But yea, that alone isn't enough to balance out the reasons against upgrading.
>>
>>> [snip]
>>>>> so supporting 64bit is just supporting the current technology. it's
>>>>> not about fancy servers or anything like that, just supporting the
>>>>> current standards. that's a minimun that should be expected from any
>>>>> compiler implementation nowadays.
>>>>> b) even though for now there is a compatability mode in most OSes, why
>>>>> would I want to limit the performance and abilities of my PC to old
>>>>> technology which is being faded away?
>>>>>
>>>> Even in 32-bit "legacy" mode, 64-bit systems are absurdly fast anyway.
>>> Talk about adding insult to injury. This is a rather random statement to make. Really, browsing the Web, writing documents, or writing emails is all you want from a computer? I'd say, until computers are not at least potentially capable of doing most intellectual tasks that people do, we're not in the position to say that computers are fast enough. When seen from that perspective, computers are absurdly slow and scarce in resources. The human brain's capacity bypasses our largest systems by a few orders of magnitude, and if we want to claim doing anything close, we should at least have that capacity. But even way, way before that, any NLP or speech recognition system that does anything interesting needs days, weeks, or months to train on computer clusters, when it all should run in real time. Please understand that from that perspective the claim that computers are plenty fast and memory is plenty large is rather shortsighted.
>>>
>>
>> When a reasonably-priced computer comes around that can actually do those sorts of things, I may very well be finally enticed to upgrade. But like you said, as it stands right now, even the high-end stuff can't do it. So it's really a non-issue for now.
>
> I don't understand. This is like a reply to another thread. This anyone would agree with. I agree that for your current computing work and perceived needs you don't feel about upgrading your hardware. I mean, what's really there to disagree. But that has nothing to do with the generalizations aired before a la "64-bit systems are absurdly fast anyway" or that there's no need for 64-bit. To write software that tackles hard problems one really needs the fastest hardware one's budget can buy. I can't understand what you say except in the frame that you indiscriminately assume that everybody else has your wants and needs from a computer (and consequently is a snob for getting a relatively fast one). Really that's a rather... unsophisticated world view to go by. I'm even amazed I need to spell this out.
>

You didn't need to spell it out, you just needed to pay more attention to what I've said, as you appear to have misunderstood much of it.  I've flat out said a number of times by now that, yes, there are legitimate uses for 64-bit. Heck even my original post regarding 64-bit indicated as much ("What are you writing, video editors and 3D modeling apps?"). What I *have* been saying is that #1 **I** am not currently interested in 64-bit, and #2 I feel there are too many people out there that only *think* they need it, and even worse, expect that everyone else should also be jumping head-first into 64-bit just because it's there. (Note again, that in that previous sentence, I did *not* indicate that "no one" has a need for 64-bit).

Nowhere have I ever said that 64-bit is and forever will be useless for everyone. Please stop coloring my comments in that light.

(I do, however, stand by my comment that *right now* trying to maximize performance on 64-bit machines is usually a misspent effort. For instance, certain game developers, like Epic and Crytek, have been focusing their target systems and optimization efforts on high-end stuff. That's just stupid as it artificially shrinks their target market. They'd be better off putting their optimization effort on lower-ends so that they can *increase* their market instead. But, yes, obviously there are going to be fringe-case exceptions even with this, such as researchers writing custom DNA-processing code that's only ever going to run on their super-duper-cluster.)


December 28, 2008
On Sat, 27 Dec 2008 17:57:39 -0500, Nick Sabalausky wrote:

> "Derek Parnell" <derek@psych.ward> wrote in message news:nkr1wyvyj3vv$.qr1gd1h779fx.dlg@40tude.net...
>> On Sat, 27 Dec 2008 15:45:57 -0500, Nick Sabalausky wrote:
>>
>>> ... judging by number of people here asking for
>>> 64-bit, I find it highly unlikely that most of them have plans to work on
>>> such things either.
>>
>> My interest in 64-bit hardware support is based on the belief that before too long, buying a new 32-bit platform might be a difficult thing to do. Five years from now, I don't want to be forced into finding a good second-hand machine just so I can work with D.
>>
> 
> I don't want to be forced into buying a new 64-bit machine just because a whole bunch of "gotta have the faciest stuff out there" people have deemed 32-bit insufficient for all computing needs.

And neither do I, but the momentum is far too strong for me to affect, so I assume that it won't be long before I will be forced to get a 64-bit machine - even if I don't want to.

> Besides, can't 64-bit machines run 32-bit code?

Sure, the same way the 32-bit machines run 16-bit software, and that's not giving anyone troubles ;-)

-- 
Derek Parnell
Melbourne, Australia
skype: derek.j.parnell
December 28, 2008
Chad J пишет:
> Walter Bright wrote:
>> What platforms for dmd would you be most interested in using?
>>
>> .net
>> jvm
>> mac osx 32 bit intel
>> mac osx 64 bit intel
>> linux 64 bit
>> windows 64 bit
>> freebsd 32 bit
>> netbsd 32 bit
>>
>> other?
> 
> Mac OSX 32 and 64 intel
> Windows 32 and 64
> Linux 32 and 64
> WinCE on ARM (yeah, I know it's not on the list, but it matters)
> Linux on ARM (ditto)
> 
> Or, better yet:
> Cross-platform C code.
> Get me that and I have a lot less reason to even care about the others.

And it seems to me, probably this will help to write programs on D for
GPU (for CUDA-like engines)
December 28, 2008
On Sat, 27 Dec 2008 19:45:48 -0800, Walter Bright wrote:

> BCS wrote:
>> Agree. Compilers seem to need to be written for the corner cases. "No one will ever need to do that" is never a valid answer.
> 
> The other thing to consider is that the type of programmer interested in D likely would also be one wanting to work with the latest and greatest machines, pushing their limits, and that means 64 bits.

I'm already saving up for the 128-bits CPU's. ;-)

-- 
Derek Parnell
Melbourne, Australia
skype: derek.j.parnell
December 28, 2008
On 2008-12-27 09:02:57 -0500, "Nick Sabalausky" <a@a.a> said:

> Ordinarily, I would agree with the need for PowerPC Mac support. But when
> one spends time in the Apple world, they really need to accept the fact that
> their systems will become abandoned at break-neck speed. That's just the way
> the Apple world works these days, and that will continue to be standard
> procedure for at least as long as Jobs in in charge.

Apple doesn't care much about backward compatibility for most of its own apps, but still provide means for developers to work with older versions of the Mac OS. Using Xcode 2.0, you can develop applications targeted at Mac OS X 10.3 (Panther, a PowerPC-only OS), even on an Intel Mac. And using Xcode 1.5, which was updated to work properly on Mac OS X 10.5 Leopard, you can support as far as Mac OS X 10.2 (Jaguar).

I don't expect Apple to remove PowerPC support from its devlopment tools until a few more years. After all, Apple relies on these development tools to support their own apps that need to work on older OSs, like iTunes which still works on Panther.

It's true that most Apple apps and many third-party ones are only compatible with the two or three latest major operating system versions, but consider this: three major operating system versions would take us to Mac OS X 10.8 (10.6, Snow Leopard, being the first removing support for PowerPC), which can't really happen before 2012 or 2013. In the meanwhile, being able to compile apps only for Intel processors will be a disadventage for application developers choosing to work in D.

That said, I concede that it'd be a little silly to do a big investment in a PowerPC D compiler for Mac OS X. But it's sad that I won't be able to use an eventual DMD for Mac OS X unless I buy a new computer.

-- 
Michel Fortin
michel.fortin@michelf.com
http://michelf.com/

December 28, 2008
On 2008-12-27 15:45:58 -0500, Walter Bright <newshound1@digitalmars.com> said:

> Michel Fortin wrote:
>> (Not that I expect to see DMD generate PowerPC code in a near future.)
> 
> There was a ppc code generator at one time, but it got lost.

What do you mean, lost?

Any chance of it being resurected? Even if it generates suboptimal code, it'd be great to have something that works.

-- 
Michel Fortin
michel.fortin@michelf.com
http://michelf.com/

December 28, 2008
Order of preference:
  Mac 32bit
  * 64bit
  .net
  --- below only theoretically interesting
  freebsd
  netbsd
  jvm


For my personal project I'm currently fine with only 32-bit x86 support.  Mac support would definitely be the most important step here, especially since many people I know which would like to use D are on Macs and are discouraged by the bad support for that platform. Also it will be an important platform for the time when the project goes public.

At work I would like to eventually start introducing D when D2 has settled a bit (the planned threading model will be very interesting there). In that environment all of Win/32/64 and Mac/32/64 would have to be supported. 64-bit support can be very useful there, as virtual address space is a major optimization target (image processing and database operations). In addition, 64-bit support is required by marketing needs alone (people with their shiny new Vista-64 PC asking why the software does not use their hardware properly). Currently we also still have to support PPC. I'm not sure how long that requirement holds, probably depends on Apple's own deprecation policy.

.NET support would be very welcome if it could act as a transparent bridge between D and .NET (same would go for D and Cocoa) - This would be the part where Managed C++ (or Objective-C++) is currently used. If .net should be supported before OS X, this could also be used as a nice platform replacement.
December 28, 2008
Walter Bright wrote:
> John Reimer wrote:
>> Wired is not necessarily backwards. :)  Despite the tangly lines, it's just easier to keep secure.
> 
> There are more reasons to like wireds. Plug the wire in, and it works, you're up on the LAN. With wireless, there's usually 3-5 minutes of fiddling to get connected. And if it won't connect, you have no idea why, so you go cold boot the machine, cold boot the wireless access point, hold the antenna up, etc. Phui.

Interference can be a huge problem as well.  Basically all in-home devices use the same spectrum, and it's not uncommon for cordless phones to knock a WiFi laptop offline, etc.  And then there's house construction.  My parents used to live in an older house with plaster walls, and plaster walls are laid on steel mesh to hold the plaster in place.  They may as well have been living in a giant Faraday cage as far as WiFi was concerned.  Wireless is a great option, but if I *can* run a wire easily from the router to wherever then it's definitely preferred.


Sean
December 28, 2008
Walter Bright wrote:
> Nick Sabalausky wrote:
>> - Graphics card that's pixel shader v1 (was a pre-pixel-shader GeForceMX 2 for a long time, only upgraded because I found this one for about $40 and wanted to play around with pixel shaders).
> 
> I don't have a graphics card because of heat/fan/noise.

It's possible to get a GeForce 6800 with a passive cooler (heatsink) instead of a fan.  I bet you can get a newer one as well, though you may have to opt for the "mobile" version.  Either way, I've got an 8800 with a fan right now but it's really very quiet--it pays to read reviews that discuss noise levels of various brands.


Sean
December 28, 2008
Reply to Walter,

> What platforms for dmd would you be most interested in using?
> 
> .net
> jvm
> mac osx 32 bit intel
> mac osx 64 bit intel
> linux 64 bit
> windows 64 bit
> freebsd 32 bit
> netbsd 32 bit
> other?
> 

whatever you decide to support, can it be fully open source?