View mode: basic / threaded / horizontal-split · Log in · Help
December 28, 2008
Re: dmd platform support - poll
"Andrei Alexandrescu" <> wrote in message 
> Nick Sabalausky wrote:
>> "Andrei Alexandrescu" <> wrote in message 
>> news:gj6mds$28iv$
>>> 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
Re: dmd platform support - poll
On Sat, 27 Dec 2008 17:57:39 -0500, Nick Sabalausky wrote:

> "Derek Parnell" <derek@psych.ward> wrote in message 
> news:nkr1wyvyj3vv$
>> 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
Re: dmd platform support - poll
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
Re: dmd platform support - poll
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
Re: dmd platform support - poll
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
December 28, 2008
Re: dmd platform support - poll
On 2008-12-27 15:45:58 -0500, Walter Bright <> 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
December 28, 2008
Re: dmd platform support - poll
Order of preference:
  Mac 32bit
  * 64bit
  --- below only theoretically interesting

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
Re: dmd platform support - poll
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.

December 28, 2008
Re: dmd platform support - poll
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.

December 28, 2008
Re: dmd platform support - poll
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?
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Top | Discussion index | About this forum | D home