So much for optimization

[Sunner]

Actually when I apt-got gnome I got part 2.4 and part 2.2.

 

[Spyro]

Originally posted by: Sunner
Actually when I apt-got gnome I got part 2.4 and part 2.2.

Same here, which is why I look carefully before I do any major upgrades. (Is gnome still half and half, my one debian box is running Stable so I haven't checked this recently).

 

[Kadarin]

Originally posted by: BingBongWongFooey
I have a very strong distaste for gentoo's forums. They are basically the epitome of what I don't like about the bulk of gentoo's users.

If I may ask, what is it that you dislike about the bulk of gentoo's users? I have had a couple of problems that I was able to easily solve by searching through their forums, and I solved a problem, posted my solution, and that solution has helped someone else. I've had all good experiences so far, though admittedly I'm still a gentoo n00b.

 

[Sunner]

Originally posted by: Astaroth33

Originally posted by: BingBongWongFooey
I have a very strong distaste for gentoo's forums. They are basically the epitome of what I don't like about the bulk of gentoo's users.

If I may ask, what is it that you dislike about the bulk of gentoo's users? I have had a couple of problems that I was able to easily solve by searching through their forums, and I solved a problem, posted my solution, and that solution has helped someone else. I've had all good experiences so far, though admittedly I'm still a gentoo n00b.

I agree.

I guess BBWF is talking about the whole "Woot, it took me one week to compile, but damn it's fast!!!!" thing.
But I don't see a whole lot of that over there, and people are just generally nice and helpful, which is a nice change, much welcome after reading some Debian mailing list archives or spending some time on a Solaris channel on IRC.

 

[Barnaby W. Füi]

Originally posted by: Sunner

Originally posted by: Astaroth33

Originally posted by: BingBongWongFooey
I have a very strong distaste for gentoo's forums. They are basically the epitome of what I don't like about the bulk of gentoo's users.

If I may ask, what is it that you dislike about the bulk of gentoo's users? I have had a couple of problems that I was able to easily solve by searching through their forums, and I solved a problem, posted my solution, and that solution has helped someone else. I've had all good experiences so far, though admittedly I'm still a gentoo n00b.

I agree.

I guess BBWF is talking about the whole "Woot, it took me one week to compile, but damn it's fast!!!!" thing.

Yeah, that's basically it. I'm sure most of them are OK people and there I know plenty of people that use gentoo that are as cool as anyone else (like you two 😛), but there are a decent number of these ultra-newbie, uninformed.... dorks. The best analogy I could ever make is that they're the ricers of linux. 😛 They're new to it but all they care about is stuff that I find stupid like eye candy and raw speed. And that's all it boils down to - they're just interested in different things than I am. But I still reserve the right to get annoyed with them and lash out every so often. 😉

But I don't see a whole lot of that over there, and people are just generally nice and helpful, which is a nice change, much welcome after reading some Debian mailing list archives or spending some time on a Solaris channel on IRC.

My ideal environment would be somewhere in-between. A lot of places (like the ones you mention) can be unnecessarily harsh.

 

[Sunner]

They're new to it but all they care about is stuff that I find stupid like eye candy and raw speed.

Hey!
I like eyecandy!

 

[chsh1ca]

Originally posted by: BingBongWongFooey
Yeah, that's basically it. I'm sure most of them are OK people and there I know plenty of people that use gentoo that are as cool as anyone else (like you two 😛), but there are a decent number of these ultra-newbie, uninformed.... dorks. The best analogy I could ever make is that they're the ricers of linux. 😛 They're new to it but all they care about is stuff that I find stupid like eye candy and raw speed. And that's all it boils down to - they're just interested in different things than I am. But I still reserve the right to get annoyed with them and lash out every so often. 😉

I can't say my experience with gentoo users has been good at all. I personally think that it's just the need to feel like part of that higher-knowledge group of folks who prefer source packages and don't mind hacking things up a little to get them working (cough*damnbrokenmakefilesinsistingthatcpandtarareborked*cough*whentheyworkjustfineaftercheckingisdisabledandIrunthescript*cough). These are guys who wouldn't necessarily have a clue as to how to do it on their own, so the scripted installer saves them.
By the by, does anyone else find it ironic that most of the people who think that compiling everything from scratch in a 'highly optimized way' seem to run on much older, slower systems, and sometimes make wild claims about how much better everything runs (ie: I tried RedHat on my friend's 2.4C OCed to 3.1w/HT, and it's SLOW compared to Gentoo running on my T-Bird 1GHz)? Aside from that I have to laugh at how long some of these people wait for installs. The particular Gentooite I'm thinking of took about a week to get a system up and running on his Athlon-800.
Whatever floats your boat, but personally I like to actually use my computer the day I install.

 

[Sunner]

I personally think that it's just the need to feel like part of that higher-knowledge group of folks who prefer source packages and don't mind hacking things up a little to get them working)

I could say the same about ~95% of all Slack users 😉

I've tried a crapload of distros, and I used to always end up with Debian in the end, on the rig Im using now, I started out with Debian, but as I mentioned above, installing Debian on an nForce2 mobo is a major PITA, while on this box, the live-install CD just picked up both the NIC's, sound, etc etc, it worked perfectly out of the box.

 

[sciencewhiz]

Originally posted by: Sunner
Actually when I apt-got gnome I got part 2.4 and part 2.2.

Yes, that happened when it first came out, but within a few days it was all 2.4

 

[chsh1ca]

Originally posted by: Sunner
I could say the same about ~95% of all Slack users 😉

The difference of course being that Slack ACTUALLY makes you work for it, it doesn't have scripts to make you 1337, you have to be 1337. 😛
Actually, Slack is really for anyone who is tired of RPM dependancy issues and waiting on the maintainers of apt to get software running, doesn't mind tarballs and/or Doing It Themselves(TM). To bastardize an old phrase, "Once you are capable of going Slack, you'll never go back." 😀
EDIT: Just to be clear, I'm saying that the difference between real slack users and real gentoo users is that Slack users HAVE to know what they're doing -- at least partially, while Gentoo users WANT to know what they're doing. HUGE 😀
(I am joking for those of you who are humour impaired)

I've tried a crapload of distros, and I used to always end up with Debian in the end, on the rig Im using now, I started out with Debian, but as I mentioned above, installing Debian on an nForce2 mobo is a major PITA, while on this box, the live-install CD just picked up both the NIC's, sound, etc etc, it worked perfectly out of the box.

Nothing wrong with being a masochist. 😛

For those of you willing to flame away, I'm just kidding, hence the smiley.

 

[n0cmonkey]

Doesn't Ultra SPARC and maybe Alpha have bigger registers? If this is (atleast partially) due to register issues, would -O3 work better on archs with bigger registers?

Must stay out of distro debate... 😛

 

[Barnaby W. Füi]

Originally posted by: n0cmonkey
Doesn't Ultra SPARC and maybe Alpha have bigger registers? If this is (atleast partially) due to register issues, would -O3 work better on archs with bigger registers?

Must stay out of distro debate... 😛

You mean cache? I thought a register was just a byte. Or a word, or the size of a pointer, or something.

 

[n0cmonkey]

Originally posted by: BingBongWongFooey

Originally posted by: n0cmonkey
Doesn't Ultra SPARC and maybe Alpha have bigger registers? If this is (atleast partially) due to register issues, would -O3 work better on archs with bigger registers?

Must stay out of distro debate... 😛

You mean cache? I thought a register was just a byte. Or a word, or the size of a pointer, or something.

Maybe. It's the end of the day for me. I keep trying to not post when I'm tired...

 

[n0cmonkey]

No, I think I mean registers. DISCLAIMER: I am tired, sore, and grumpy; so I could easily be reading this incorrectly (and using a semicolon incorrectly). Corrections are appreciated.

After two years on the road Intel introduced Pentium II in 1997 with 32 bit Main CPU register size, 7M number of transistors, and 64GB addressable memory. Recently, in 1999 Intel released Pentium III with 32 to 128 Main CPU register sizes, 500MHz-clock speed, and 8.2 million numbers of transistors.

I take that to mean the registers on the P2 are 32 bits. I guess the p3 has bigger registers, but that doesn't sit right with me. I was thinking ):Q) that the Ultra SPARC and other higher end processors had larger, and more registers (not just cache). I could definitely be wrong though.

EDIT:

// Alpha is a load/store RISC architecture with all operations done between
// registers. Alpha has 32 integer registers and 32 floating registers, each
// 64 bits. Integer register R31 and floating register F31 are always zero.
// Longword (32-bit) and quadword (64-bit) integers are supported. Four
// floating datatypes are supported: VAX F-float, VAX G-float, IEEE single
// (32-bit), and IEEE double (64-bit). Memory is accessed via 64-bit virtual
// little-endian byte addresses.

Alphas have 64bit registers? Taken from here. I think the previous quote about the pentium cpus is also from this site, but I'm not 100% positive.

EDIT 2: Actually, I think this is the site I got the first quote from. Don't hate me for quoting a geocities site 😉

 

[Barnaby W. Füi]

I had a different concept of registers altogether, so you win, even if you're off by a bit. 😛

 

[drag]

You'll find this very interesting then:

from www.arstechnica.com

One of the oldest and longest-running gripes about x86 is that the programming model has only eight GPRs, eight FPRs, and eight SIMD registers. All newer RISC ISAs support many more architectural registers; the PowerPC ISA, for instance, specifies thirty-two of each type of register. Increasing the number of registers allows the processor to cache more data where the execution units can access it immediately; this translates in to a reduced number of LOADs and STOREs, which means less memory subsystem traffic, less waiting for data to load, etc. More registers also give the compiler or programmer more flexibility to schedule instructions so that dependencies are reduced and pipeline bubbles are kept to a minimum. (For more on dependencies and pipeline bubbles, see this article or this article.)

To a programmer on a x86 platform you only SEE 8 registers (of each type of register) for a pentium4

The p4 realy has 128 registers, but do to the limitation of the x87 ISA design it is constrained by compatability to 8, so it shifts them around internally to all 128 using it's internal logic stuff. Probably helps deal with latencies and get those high clock speeds. I guess, stuff is still a bit over my head.

The more I learn about what ISA realy is, the more I realise that it's not going anywere for a LONG LONG time. 🙂

 

[n0cmonkey]

Originally posted by: BingBongWongFooey
I had a different concept of registers altogether, so you win, even if you're off by a bit. 😛

That was horrible. 😛






Yes, I almost laughed out loud. 😉

 

[n0cmonkey]

Originally posted by: drag
You'll find this very interesting then:

from www.arstechnica.com

One of the oldest and longest-running gripes about x86 is that the programming model has only eight GPRs, eight FPRs, and eight SIMD registers. All newer RISC ISAs support many more architectural registers; the PowerPC ISA, for instance, specifies thirty-two of each type of register. Increasing the number of registers allows the processor to cache more data where the execution units can access it immediately; this translates in to a reduced number of LOADs and STOREs, which means less memory subsystem traffic, less waiting for data to load, etc. More registers also give the compiler or programmer more flexibility to schedule instructions so that dependencies are reduced and pipeline bubbles are kept to a minimum. (For more on dependencies and pipeline bubbles, see this article or this article.)

To a programmer on a x86 platform you only SEE 8 registers (of each type of register) for a pentium4

The p4 realy has 128 registers, but do to the limitation of the x87 ISA design it is constrained by compatability to 8, so it shifts them around internally to all 128 using it's internal logic stuff. Probably helps deal with latencies and get those high clock speeds. I guess, stuff is still a bit over my head.

The more I learn about what ISA realy is, the more I realise that it's not going anywere for a LONG LONG time. 🙂

I wonder if there is a performance hit with that... Yes, I know it would be something like 0.00000000000001ms or something crazy like that, but still. Curiosity is a wierd thing when you are exhausted 😛

 

[drag]

The size of the register is what makes them 32 bits or 64 bits or 8 bits or 16 bit proccessors.

It's just how much info they can gulp down in one go. When dealing with 64bit data types a 32 bit proccessor will take 2 steps to proccess it, while 64 bits CPU can get it in one go. 64bits allow faster proccessing of larger data types and you can get much more persision with floating points without a performance penalty.(probably why 64bits is cool for gamers and other high-multimedia stuff)

When we went from 16 to 32 bits Intel just doubled the register size, and that's what AMD did when it created the x86-64. (among other archatectural improvements. )

 

[n0cmonkey]

Originally posted by: drag
The size of the register is what makes them 32 bits or 64 bits or 8 bits or 16 bit proccessors.

It's just how much info they can gulp down in one go. When dealing with 64bit data types a 32 bit proccessor will take 2 steps to proccess it, while 64 bits CPU can get it in one go. 64bits allow faster proccessing of larger data types and you can get much more persision with floating points without a performance penalty.(probably why 64bits is cool for gamers and other high-multimedia stuff)

When we went from 16 to 32 bits Intel just doubled the register size, and that's what AMD did when it created the x86-64. (among other archatectural improvements. )

That definitely makes sense. Thanks!

 

[Sunner]

Originally posted by: drag
The size of the register is what makes them 32 bits or 64 bits or 8 bits or 16 bit proccessors.

It's just how much info they can gulp down in one go. When dealing with 64bit data types a 32 bit proccessor will take 2 steps to proccess it, while 64 bits CPU can get it in one go. 64bits allow faster proccessing of larger data types and you can get much more persision with floating points without a performance penalty.(probably why 64bits is cool for gamers and other high-multimedia stuff)

When we went from 16 to 32 bits Intel just doubled the register size, and that's what AMD did when it created the x86-64. (among other archatectural improvements. )

The bitness tend to be the general purpose registers, x87 still has 64/80 bit FP, while the GPR's are 32 bit.