AMD and SuSE Porting Linux to Sledgehammer

-|Oblom|- writes "AMD has partnered with SUSE to port Linux to its upcoming 64-bit Sledgehammer chip. The story is on CNET and the projects site is here www.x86-64.org Well... I have been waiting for a while for this announcment. Hopefully by the end of next year I'll be running dual-core 1.5Ghz(at least) Sledgehammer with Linux on it"

Comments

More bits != More speed

[Chris Burke]

The reason to switch to 64 bits is not performance. Extra bits don't give you more speed. It only increases what you can do (which might possibly give more speed, but generally wider datapaths are slower, not faster).

The most important reason to want 64 bits is for server applications which want an address space larger than 4GB. 64 bits of virtual address space is the main attraction of these chips, and only for servers. Which is why Sledgehammer is being pushed as a server-only proc to compete with Merced.

You might get other benefits, like 64 bit integer ops being faster (but not necessarily... adequate bypass networks in a 32-bit proc might make this a wash). Which is only a benefit if your app uses lots of 64 bit integer ops.

There are also penalties -- for example, the page table hierarchy has 4 levels, which means more memory accesses on a TLB miss.

16->32 was different, because it also gave you all kinds of benefits like protected mode, virtual memory, and other stuff i'm too lazy to remember.
And 16 bits was never really enough.

Anyway, the point is that there is no real reason to worry about current apps moving to 64 bits when Sledge hits. Those server apps that will benefit will switch, and those that won't have no reason to (which is the true beauty of this processor).

Re:More bits != More speed

[Chris Burke]

IA32 already supports 80-bit FP, and has 80-bit FP registers. nobody uses that, but the upshot is double precision is already there and cooking. doubleint conversions will probably be easier with 64bit int registers, but those really aren't the critical path.

Though getting away from the stack-based FP architecture _will_ be an improvement. ^_^

Re:More bits != More speed

[Inoshiro]

"Extra bits don't give you more speed. It only increases what you can do (which might possibly give more speed, but generally wider datapaths are slower, not faster)."

No true at all. Wider data paths are faster because they can move more data per tick. Compare: NARROW FAST SCSI -> 10Mbps, WIDE FAST SCSI -> 20mbps.

And if you increase the number of clock ticks by 2, you get ULTRA WIDE SCSI -> 40Mbps!

It's a choice of pumping more bits per tick, or having more ticks with which to pump bits. An AMD Athlon may have 1 billion ticks with which to pump bits, but the bus it's connected to is incredibly slow compared to it. Widen the path, and you get double the performance with today's technology.

I think your flaw is in thinking based on what you know. As most first year chem students, and they'll say that wet air is heavier than dry air because a wet towel is heavier than a dry one. This is flawed logic.
---

TLB misses.

[Christopher Thomas]

There are also penalties -- for example, the page table hierarchy has 4 levels, which means more memory accesses on a TLB miss.

While this is true, its impact can be reduced by having a multi-level TLB (cache the physical addresses for lower level pages of the page table as well as the physical addresses for the destination page). Several architectures already do this.

This would allow me to get a new page's physical address with a single lookup, as long as it's within the same block of pages as another page I've already accessed.

Re:SuSE

[rbrito]

Indeed. I rarely post here to Slashdot, but I think that people should give more credit where credit is Due.

I'm an avid Debian GNU/Linux user (and I do intend to be a Debian Developer if I can in the near future), but I can't help but recognize all the good things that SuSE Linux has been paying kernel hackers for.

They seem to be incredibly commited to the Free Software movement, yet they get very little credit.

Indeed, people wouldn't have support for many high-end devices and methods if it were not for the support that SuSE is putting into Linux. I won't mention all them, but there are some of the things that I remember:

• Funding for development of drivers for X;
  
• Funding for development of ReiserFS;
  
• Funding for the Alsa Project;
  
• Funding for IDE/ATA Drivers.
  

Many people need those things (which shows the relevance of the support) and I'm sure that there are many other projects with which SuSE may be involved. Congratulations!

Roger...

Re:Sledgehammer as a high end server platform?

[Chris Burke]

Sledgehammer is aimed at the same market space as the Itanic (sorry, reg, but I love this name). Only big servers really care about 64 bits at this time, but 'this time' seems to be the forseeable future.

As far as ports... according to the docs, it should run any 32 bit OS that runs on current x86 hw. They won't take advantage of the 64 bit features, but if you don't need em...

This is a very smooth move by AMD.

Re:This shouldn't be hard

[Fizgig]

Well, it'll work with Linux right now, but you can't do anything 64-bit at all with it without a port. Sledgehammer has 3 modes. The first is legacy mode. You use that to run a 32-bit operating system and 32-bit applications. Linux will run in this way out-of-the-box. You can't run 64-bit applications in this mode and have to reboot to change. Then they have "long mode " which is split into two other modes, whose names I don't remember. Long mode runs a 64-bit operating system. The first of the long modes (compatibility mode or something) runs 32-bit applications on the 64-bit operating system. The second mode runs 64-bit applications. The two long modes can be changed by a context switch, so you can be running 64-bit and 32-bit applications on the same 64-bit OS.

SuSE

[MicroBerto]

It seems that SuSE is so involved in many projects out there, and doesn't get much credit. And they don't have a very large market share on the distribution level either. My favorite thing that they help a lot in is The Alsa Project (great sound drivers). SuSE certainly deserves more credit for helping keep Linux on the bleeding edge, so I just thought I'd toss that in.

Mike Roberto
- GAIM: MicroBerto

This shouldn't be hard

[cvd6262]

One of the comments (it might have been from Linus), when AMD announced their 64-bit chip was that this should be the easiest Linux port ever.

Is this because Linux can be so easily manipulated for it's host environment, or because it's just powerful enough to run already on a 64-bit machine?

AMD and Intel

[LordDartan]

Since AMD and Intel are each making their own instruction set for 64bit, does that mean that software will have to be written for both?? Or is there going to be some compatability layer that allows AMD chips to run Intel code and vice versa??

Re:AMD and Intel

[Joe Groff]

GCJ is working on this sort of thing; it compiles Java source to native code, and can also compile Java .class bytecode into native code as well. It also has some cool features, like the ability for native compiled code to call on Java bytecode, etc. It's a cool project, check it out.

- Joe

This is great.

[Th3 D0t]

This should accelerate the mainstream acceptance of the 64-bit architecture greatly. It will easy to port applications from current x86 linux to this architecture. Now if they'd make a 64-bit version of Kylix, we'd be all set! MS should have fun playing catch up with the Windows monstrosity.
---

Shouldn't be too long

[_Mustang]

after that the rest of the distro's get support for it as well. Based on the article, I would question just how much effort is going to be put into the actual applications. Sure the OS will support this, but since this new X64 will run X32 apps just as well where is the incentive for writing enhanced applications..?

Re:Shouldn't be too long

[arivanov]

ust as well where is the incentive for writing enhanced applications..?

In the binary only commercial world - none. Sun is already locked up in this paradigm. Have a look at solaris. 64 bit kernel, most of the apps are 32 bit in order to be compatible with older sparcs.

If the app is distributed as source and the toolchain is 64 bit the app will be promoted automagically.

Sledgehammer as a high end server platform?

[cheezus]

It's nice to see that the linux community is getting help from a corporation in supporting new hardware. It's a nice change from hardware vendors having to make sure their product is MS compatable, instead of vice versa like it should be

So what is AMD's plan? Is Sledgehammer going to be used in highend servers? If this is the case, I think they are definelty taking the right course in not only helping out linux, but also protecting their interests. It would be hard for other chip manufactures to compete with a more powerful platform that had multiOS support. Linux and (i'm gonna assume here) NT/2000 are a good start. Has there been any news from the BSD camp on a port? I mean, "of course it runs NetBSD", doesn't it?

---

Re:Sledgehammer as a high end server platform?

[Oestergaard]

Don't expect NT/2K to run in ``real 64-bit mode'' on the sledgehammer anytime soon. NT never ran 64-bit on the Alpha either, user-space processes were always 32-bit.

Porting NT to a 64-bit CPU may be doable, and it seems MS is doing just that. With Linux that took some effort too, but luckily that was done while Linux was still young.

However, there's a lot more to 64-bit computing than just the kernel. You want your user-space to go 64-bit as well, and this is where it gets innteresting... Why didn't NT on the Alpha run 64-bit processes ? Well, does Win_32_ ring a bell ?
In POSIX-like systems such as GNU/Linux, you use data-types such as int and char* for passing integers and pointers to data. In Win32 you use DWORD and LPxxx etc. The Win32 DWORD is _exactly_ 32 bits, and every program written for Win32 will tend to depend on this. Even worse. the LPxxx pointers are _also_ assumed to fit in the same space as a DWORD, namely 32 bits. A program using char* to reference a block of data, will be equally valid on 16, 32 and 64 bit architectures, because the datatype states it's just a pointer. In Win32 your LPxxx types are 32 bits, meaning they make no sense in a 64-bit environment. Tough.

What I'm getting at (slowly I know) is, that while Microsoft is currently trying to design a Win64 API that everyone must now port their programs to in order to take advantage of the 64-bit environment, the GNU/Linux community will hit the ``make; make install'' combination, and the vast majority of applications will only need minor fixes if they have not already been ported to a 64-bit architecture. But chances are, of course, that the vast majority of GNU/Linux apps have been working just fine on 64-bit architectures for years.

64-bit windows will not happen in the next few years. Even if Microsoft should ship a native 64-bit NT, _and_ actually finish up Win64. It's the applications that matter, and that is one thing that Windows just do not have (despite popular (trash-media) belief). Go easy :)

AMD SMP

[Rayban]

Will we be able to finally get SMP with these AMD chips? Right now we have these amazing Athlons capable of symmetric multi-processing, but no mobo hardware support for them.

IMHO, AMD has gone the right way with x86-64, rather than a whole new instruction set. At this point in the game, I don't think they have enough market pull to convince people of once standard vs. another. It's a bit of a shame Intel and AMD couldn't have cooperated on a comment 64-bit spec, but I know exactly what sort of chance that would have (it involves a snowball and a very warm place...).

Re:AMD SMP

[snort]

This fall we'll have SMP Athlons. We're waiting for AMD's 760MP chipset... Supposed to do SMP and DDR SDRAM.

I just wish they'd kick me down a few engineering test samples.

Re:Why Sledgehammer?

[jallen02]

Its simple, its functional.

The name implies strength

You dont forget about it

Those are most of the major tenets of a good trade mark / advertisement.

I like it ;)

Jeremy


If you think education is expensive, try ignornace

It's a "for profit" company?!?!

[karmma]

Obviously, you haven't seen VA Linux stock prices. Wow.

---------------------
This space available. Reasonable rates.

Re:Why Sledgehammer?

[kwsNI]

The name implies strength

To me, it implies that the system will be so loud, you need gas filled ear protectors to use it ;)

kwsNI

Could this also mean 3DNow! support in gcc?

[ncc74656]

The index page for x86-64.org indicated that one of the earlier goals for the project will be to add x86-64 support to gcc (gotta have gcc before you can do anything else). Since x86-64 is a superset of IA-32 (what we're all using already) and 3DNow!, what do you suppose are the possibilities of 3DNow! optimizers finding their way into gcc? For those of us running K6-*s, Athlons, and Durons, this would be a Good Thing (TM), and I suspect it's something would benefit Sledgehammer once it's out.

_/_
/ v \
(IIGS( Scott Alfter (remove Voyager's hull # to send mail)
\_^_/

Re:This is Big

[Oestergaard]

There is one problem though... We need the desktops using this processor if it has to be just remotely affordable, and if we want a decent number of motherboards to choose from

How cheap is the Xeon currently, even though it has very few benefits over the ``desktop'' processors such as Athlon or PII/III ? Not very. Why ? Because it's not sold in mind-boggling quantities. Well, also because Intel prices it in the high end, but that's a chicken-and-egg scenario wrt. the desktop market.

Currently, in this era of the lemming mentailty, we depend on MS windows processor support, if a processor is to be used very widely. Unfortunately this means, the Sledgehammer won't be affordable until MS releases an OS for it. Yes that sucks, but blame society :)

But yes, GNU/Linux support for the Sledgehammer some year or two ahead of Microsoft is going to give us great press. At least for a month or so. But counting in the long-term memory and interest in any kind of history (even just last year's history) of reporters, I doubt that we will benefit much from this once MS finally ships their OS for the Sledgehammer.

Don't get me wrong though. I think this is great, and the Sledgehammer may well prove as an alternative to the high-end and expensive Xeon CPUs from Intel, and they may well be used by those who need it enough to be able to afford it, for things like Oracle, SAP, weather forecasts, nukes, and what gives...
Way to go SuSE ! (and AMD!) :)

Re:AMD and Microsoft

[Oestergaard]

Most likely, I suppose AMD has secretaries like most other companies ;)

But do you think that AMD would be spending time and effort on a Linux port right now, if NT for Sledgehammer was just around the corner, with server applications support etc. etc. ?

My bet is, that AMD tried, and Microsoft were either honest (heh, no let's be serious) or AMD figured out that Oracle/MsSQL/DB2/SAP/whatever on 64-bit NT is much further away than anyone planning to ship a new CPU this decade would like to even think about.

The GNU/Linux system has the easily-adabtable tools, the easily-portable user-space, and an easily portable kernel. While Microsoft has the marketing people, the company with an employment politic that says education doesn't matter, somehow (I wonder why, nah, I don't) doesn't have any of the other.

Re:Why Sledgehammer?

[iceT]

"Why did AMD name their new processor something like that? I mean, normally, you'd associate sledgehammers w/ something you'd want to keep AWAY from your computer."

Obviously, you've never worked with WindowsNT, have you?

Mental note...

[NRLax27]

"Hopefully by the end of next year I'll be running dual-core 1.5Ghz(at least) Sledgehammer with Linux on it"

Mental note: Remember to start very successful geek-news website, and sell to Andover.net in order to have enough money to afford dual-core 1.5Ghz(at least) Sledgehammer machine

Re:Mental note...

[-|Oblom|-]

In fact it's not supposed to be that expencive. The 64bit part of the chip supposed to add around 5% to the core size. So the price will be (according to the common belive that hammer it's dual mustang + 64bit unit) 2x 1.5GhzMustang + 64bit unit + LDT...
I think that when it will be introduced (H2 2001)the price for the cheapest one wil be ~1500$(and then will go down slowly) and it will be hell more cheaper then Itanic system.

PS. I did started now new site, over here (Israel) with intention to sell it later to one of the ISPs :-)

Thanks for the info...

[ackthpt]

Dear Santa,

I've been a very very good boy this year. Please consider the following from my wish list:

AMD Sledgehammer

SuSE Linux

VIA PC 266 chipset (64bit equiv.)

DDR SDRAM

Mobo for all of that

Overclocking tips from Tom's

SCSI controller and 4x45GB 10000RPM drives

A 3D supported LCD letterbox montor

THX surround sound

DVD burner

A DSL provider who actually delivers

100 lbs Kona Espresso beans, 500 lbs mixed Jelly Bellies (no apple, please) & a Thai delivery which stays open past 10 PM

Thanks!

Vote Naked 2000