Guess who's gone all 64bit!!!!!

[bloodline]

MEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEee!!!!

It's so cool!! Everything is... just like 32 bit... :-?

[bloodline]

Speelgoedmannetje wrote:
in fact, I see very little progress in PC world
yeah, the video processing and the 3d stuff is a lot faster nowadays, but that's about it.

Hehehe, ok, it's not that bad.. WindowsXP 64bit is weird though... it's sort of two opertaing systems in one, I can't believe it's taken then over a year to get it to this point :-/

[bloodline]

mdma wrote:

bloodline wrote:

Speelgoedmannetje wrote:
in fact, I see very little progress in PC world
yeah, the video processing and the 3d stuff is a lot faster nowadays, but that's about it.

Hehehe, ok, it's not that bad.. WindowsXP 64bit is weird though... it's sort of two opertaing systems in one, I can't believe it's taken then over a year to get it to this point :-/

Back handers from Intel is no doubt the reason they delayed releasing Windows x64.

I expect so, but it's pretty lame... even for Windows.

[bloodline]

KennyR wrote:
I think most people subconciously think 64 bit will offer a lot more performance than 32. Actually it's usually the opposite, and even when a 64 bit CPU is actually used optimally, it's still not a lightyear ahead of current CPUs.

Well this Athlon54 at 2Ghz is certainly a good margin faster than my 3,06Ghz P4... but that's because the P4 sux :-D

[bloodline]

mdma wrote:

bloodline wrote:

KennyR wrote:
I think most people subconciously think 64 bit will offer a lot more performance than 32. Actually it's usually the opposite, and even when a 64 bit CPU is actually used optimally, it's still not a lightyear ahead of current CPUs.

Well this Athlon54 at 2Ghz is certainly a good margin faster than my 3,06Ghz P4... but that's because the P4 sux :-D

Sh!t loads of extra registers is what gives it it's speed. Oh, and the 1MB of cache on-die. :-)

64bit Linux is a hell of a lot faster on this Athlon64 3200+ of mine than 32bit Linux is.

Snap! I've got a 3200 (939 version) too :-)

Actaulyl now I've got all the drivers installed in the Windows64, it's starting to really get fast :-o It's not a little faster than my P4... it's a LOT faster (in 64bit mode anyway)!!!

[bloodline]

mdma wrote:

bloodline wrote:

mdma wrote:

bloodline wrote:

KennyR wrote:
I think most people subconciously think 64 bit will offer a lot more performance than 32. Actually it's usually the opposite, and even when a 64 bit CPU is actually used optimally, it's still not a lightyear ahead of current CPUs.

Well this Athlon54 at 2Ghz is certainly a good margin faster than my 3,06Ghz P4... but that's because the P4 sux :-D

Sh!t loads of extra registers is what gives it it's speed. Oh, and the 1MB of cache on-die. :-)

64bit Linux is a hell of a lot faster on this Athlon64 3200+ of mine than 32bit Linux is.

Snap! I've got a 3200 (939 version) too :-)

Actaulyl now I've got all the drivers installed in the Windows64, it's starting to really get fast :-o It's not a little faster than my P4... it's a LOT faster (in 64bit mode anyway)!!!

Mine's in my laptop it gets hot as fek!! :-)

Have you tried AROS on it it yet? It works fine for me on this laptop with Athlon64, VIA chipset and ATI 9600.

If only we had an ATI driver for AROS it'd be even quicker than it is in VESA mode.

Have you installed the AMD processor driver from the AMD website yet?

I've run AROS on this machine... I nearly pooped my pants :-D Though I've yet to get the AROS Nvidia driver to recognise my PCIExpress gfx card :-) Still it's sooperdooper fast in VESA mode!

AMD Processor driver? I'll have a look now :-D

[bloodline]

T_Bone wrote:
I want a **TRUE** 64 bit CPU!!!

64 bit ASSEMBLY OPCODES baby!! YEA!
(with that many opcodes, assembly would pretty much be a high level language)

:nervous: Don't let T-Bone near any CPU design lab! :-D

[bloodline]

Karlos wrote:

T_Bone wrote:
I want a **TRUE** 64 bit CPU!!!

64 bit ASSEMBLY OPCODES baby!! YEA!
(with that many opcodes, assembly would pretty much be a high level language)

Well, suppose you made a super RISC load/store architecture where you had, eg 65536 integer and 65536 floating point registers, you could easily have 64-bit opcodes. 16-bits for the instruction and up to 16-bits each for each register definition :lol:

Hmmmm, you would probably just have a 64K register file...

[bloodline]

Karlos wrote:

bloodline wrote:

Hmmmm, you would probably just have a 64K register file...

Well, you could still have L2 cache, but you'd have less need of L1 cache :lol:

I dread to think how difficult to code such a system would be implement, let alone code for. It would be difficult to manage such a huge register file effectively in a multitasking OS.

Still, whilst we're being silly :-D...

Mind you, if you had a default 'offset' register that allows your definition of 'register 0' to be offset to any register in the file, you could implement a sort of "fixed register size per task" kernel, where each task would have it' own block of X registers. Once the register file fills, youd page out the ones for those tasks that arent doing anything.

Other than that, each task would just see eg 32 registers belonging to it that the OS almost never has to save/restore as part of task switching.

I wonder what the upper limit is for a sensible 'directly accessible' (that is not including shadowning and forwarding) register count?

I think PPC's 32 is more than enough, really...

If you read how the register rename feature on an Althon64 works, it's a bit like what you've described :-)

[bloodline]

mdma wrote:
I am now triple booting as opposed to dual booting.

I have Windows XP Pro, SuSE 9.2 Pro 64bit, and now Windows XP Pro x64 Edition.

WinUAE was fast as hell already on 32bit Windows, but if I reboot into 64bit Windows it's even faster. Even though WinUAE is still a 32bit application.

Nice. :-D

I wonder if Tony will compile a 64bit version of UAE? It could really make good use of those extra registers.  Would that break the JIT as it's for x86 only?  I know the 64bit version of PearPC can't do JIT for the same reason.

I'll Email him. The JIT would have to be redesigned for the A64, it's a totally ditterent CPU design... I wonder, though, if the JIT could run in a 32bit thread while the rest of WinUAE could run in 64bit mode... I have no idea how Win64 handles this sort of stuff :-?

Might be an idea to Email Bernie and see if he wants to make an Athlon64 JIT :-D

[bloodline]

mdma wrote:

bloodline wrote:

mdma wrote:
I am now triple booting as opposed to dual booting.

I have Windows XP Pro, SuSE 9.2 Pro 64bit, and now Windows XP Pro x64 Edition.

WinUAE was fast as hell already on 32bit Windows, but if I reboot into 64bit Windows it's even faster. Even though WinUAE is still a 32bit application.

Nice. :-D

I wonder if Tony will compile a 64bit version of UAE? It could really make good use of those extra registers.  Would that break the JIT as it's for x86 only?  I know the 64bit version of PearPC can't do JIT for the same reason.

I'll Email him. The JIT would have to be redesigned for the A64, it's a totally ditterent CPU design... I wonder, though, if the JIT could run in a 32bit thread while the rest of WinUAE could run in 64bit mode... I have no idea how Win64 handles this sort of stuff :-?

Might be an idea to Email Bernie and see if he wants to make an Athlon64 JIT :-D

And a PPC JIT while he's at it too. :-)

I've been thinking about this one... A PPC->Athlon64 JIT...

Acording to the PearPC people the Lack of registers in the x86 do not cause much of a speed hit, it is emulating the PPC MMU that is the speed killer...

[bloodline]

mdma wrote:

the Lack of registers in the x86 do not cause much of a speed hit

So imagine the speed increase that doubling the amount of resiters available will have.

Err... not a lot.. :-?

[bloodline]

I think the speed up from 32bit to 64bit that mdma and myself are experiencing is simply due to more elegant optimisation allowed for by 16 registers... also the ISA of the x86-64 long mode has been "cleaned up", probably allowing instrution sequencing to allow better branch prediction and better caching, not to mention it probably looks a lot more like the RISC core, which means less overhead too.

-Edit-

<----Look at the size of that Cache!!! :-o

-/Edit-

Note, the Athlon64's register size is now 64bit long, Not sure if there is much 64bit integer processing in Windows but that will be speeded up. Maybe CPU's are so fast now that Microsoft needs counters that can go above 4billion for their "Your Brand new CPU is too slow" delay loops ;-)

[bloodline]

Karlos wrote:

bloodline wrote:
I think the speed up from 32bit to 64bit that mdma and myself are experiencing is simply due to more elegant optimisation allowed for by 16 registers...

Which is basically what I was trying to say, but you managed a lot better job of it :lol:

I think you need to cut down on the registers, mate... 32 seem to have made your head go funny... Stick to 16, you KNOW it makes sense :-D

[bloodline]

Karlos wrote:
Surely you need 64 registers to complete the 64ness of it all? :-P

:lol: I learned asm with 16 regs, I don't need any more :-)


-Off Topic- I was just curious about what this 2Ghz 64bit chip could do... and decided to set the Front Side Bus to 230Mhz (up from 200Mhz). I now have the thing running at 2.3Ghz, I didn't need to increase the Vcore... the Temp has gone up from 23C to 26C... and it's totally stable... VERY fast too... I'm glad I paid extra for some branded Memory now :-D