xp with i3 processor

[Jat]

is xp compatiable with i3 processor

[allen2]

Yes.

[Tripredacus]

The CPU itself isn't really where you want to focus on whether or not it is XP compatible. You want to see if the motherboard you have, or want to use, is compatible with XP, ie there are XP drivers available for it on the manufacturer's website. As an example. XP is supported on all Intel Desktop and Workstation boards currently available that support an i3 processor.

[allen2]

The CPU itself isn't really where you want to focus on whether or not it is XP compatible. You want to see if the motherboard you have, or want to use, is compatible with XP, ie there are XP drivers available for it on the manufacturer's website. As an example. XP is supported on all Intel Desktop and Workstation boards currently available that support an i3 processor.

I don't completly agree with you: the itanium isn't supported by XP x86.

Also if you install win 3.11 on dual core or better you won't get all the benefits of the dual core architecture.

[Tripredacus]

The CPU itself isn't really where you want to focus on whether or not it is XP compatible. You want to see if the motherboard you have, or want to use, is compatible with XP, ie there are XP drivers available for it on the manufacturer's website. As an example. XP is supported on all Intel Desktop and Workstation boards currently available that support an i3 processor.

I don't completly agree with you: the itanium isn't supported by XP x86.

Also if you install win 3.11 on dual core or better you won't get all the benefits of the dual core architecture.

Ok then, find me a motherboard that uses an Itanium CPU that has XP support...

[cluberti]

I don't completly agree with you: the itanium isn't supported by XP x86.

If you have the ungodly amounts of money required for an Itanium system, you probably aren't running Windows XP on it.

[allen2]

The CPU itself isn't really where you want to focus on whether or not it is XP compatible. You want to see if the motherboard you have, or want to use, is compatible with XP, ie there are XP drivers available for it on the manufacturer's website. As an example. XP is supported on all Intel Desktop and Workstation boards currently available that support an i3 processor.

I don't completly agree with you: the itanium isn't supported by XP x86.

Also if you install win 3.11 on dual core or better you won't get all the benefits of the dual core architecture.

Ok then, find me a motherboard that uses an Itanium CPU that has XP support...

Here it is : a HP Zx2000 http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareIndex.jsp?lang=en&cc=fr&prodNameId=82076&prodTypeId=12454&prodSeriesId=82074&swEnvOID=1026&taskId=135&swLang=13

And as for the price you can find one there : http://cgi.ebay.fr/HP-ZX2000-Itanium-2-900MHz-4GB-36GB-DVD-A7844-84002-/300333964940?cmd=ViewItem&pt=COMP_EN_Workstations&hash=item45ed4c9e8c

It isn't exactly cheap but some designers or modelers might need this kind of power (instead of buying silicon graphic workstations).

[CoffeeFiend]

And as for the price you can find one there : http://cgi.ebay.fr/HP-ZX2000-Itanium-2-900MHz-4GB-36GB-DVD-A7844-84002-/300333964940?cmd=ViewItem&pt=COMP_EN_Workstations&hash=item45ed4c9e8c

It isn't exactly cheap but some designers or modelers might need this kind of power (instead of buying silicon graphic workstations).

That *is* dirt cheap. Because it sucks, it's darn old (from 2002), it's slow as molasses (reviews say things like "1.1 times the integer performance of a 1.67GHz AMD Athlon XP 2100+ processor" about the faster 1GHz part -- ouch), and it's 2nd hand (the kind of people that used to need this would buy it new).

That's totally unlike the kind of system cluberti was making reference to. Buying that back when it was current or a similar system was several thousands of dollars. If you went with a 4-way with plenty of RAM, lots of fast drives and all then it was in the 5 figures.

[cluberti]

And they're still 10s of thousands for an Itanium system nowadays as well. Itanium's market is really anything that needs ridiculous floating point perf (like database apps/servers), and supercomputing (and even then for the cost, a host of regular x86-64 (amd64) processor servers can be had for the same $ and perform better in almost everything - except floating point operations).

Using the Itanic for anything other than databases, things like physics modeling, render farms (and it'd have to be really large), or something that needs 64bit arithmetic precision is really a waste of money - kernel-mode only has access to the first 32 FP registers, whereas a user-mode app has access to all 128 (which is partially why it's so darned good at FP operations). Also, they have memory bandwidth and FSB limitations for large-memory applications, so anything not tuned specifically to run at high perf on an Itanic will run poorly. Also, Itanic is faster than comparable PA-RISC processors (and most true RISC processors in general), but IBM's POWER can keep up at most things. Given Intel's EPIC architecture for the Itanic (rather than true RISC) means the code written for it must be tuned very carefully to get good perf - poor coding or a poor compiler will result in poor performance - while the same is true on x86/x86-64, those architectures are much more forgiving. Since we're talking about a Windows workstation for Itanic, this is a good point - there isn't much that'll run on an Itanic in the Windows world that isn't emulated (entirely) as x86, and the ia64 apps are very few. It's also worth noting that Windows Server 2008 R2 is the last Itanic OS from Microsoft, due to the fact that people running Itanium pretty much only run Unix - HP-UX, BSD, and Solaris.

Buying a workstation today means getting an x86-64 machine and running an x86-64 64bit OS on it, whether that be Windows, Linux, Unix, or BSD. Itanic is a niche server product only, and while rumors of it's death are a little exaggerated, it's not a platform that has any real forward momentum into any other markets (nor any public plans by Intel to do so either).

[allen2]

I don't agree with you at all cluberti: any serious application developer must optimize his code taking a full advantage of the platform the application will run on.

The general purpose of Intel and Microsoft is to make you think that only a better hardware will perform faster but that's totally wrong. For example multi-core and 64 bits are out for years and most windows applications aren't optimized for them.

Also with all pre-made "solution", Microsoft kill developers creativity: why creating a new algorithm when Microsoft already made one and if it isn't optimized, the application user has to buy a faster hardware.

Sometimes, it is like the developper tried to solve a 2nd degree equation without the formula but trying every possible roots.

[CoffeeFiend]

I don't agree with you at all cluberti: any serious application developer must optimize his code taking a full advantage of the platform the application will run on

Yes, but the Itanium is still significantly different. The CPU has a fundamentally different design, not only the registers and instruction set and such, but also the way the compiler has to optimize for the code (it has to be optimized very differently -- the usual out of order stuff won't run so great; again, because of the EPIC architecture).

For example multi-core and 64 bits are out for years and most windows applications aren't optimized for them

Most apps that require that kind of power to get decent performance are, and that's the main thing. I couldn't care less if notepad could use up 64 cores but I'm happy that my video codecs, Photoshop, SolidWorks and others can.

I'm not real sure what the Intel/MS bashing was about...

[allen2]

Most benchmarks (for example this one )out there show that x86 OS are not behind x64 OS in all tests.

[CoffeeFiend]

Most benchmarks (for example this one )out there show that x86 OS are not behind x64 OS in all tests.

Not so sure what your point was, but x86 is behind on every single test you linked to (20-something% in winrar, 6.x% in * mark vantage, 5.x% in Cinebench), save for the GPU-bound one (which might not be very much optimized either) where it obviously ties.

There are multiple benefits to x64, from twice as many registers (they're larger too) both the general purpose and the SIMD kind, to better calling convention (far less stack operations), to more available memory. So it's hardly surprising it runs faster (usually around 10% but sometimes it's very significant like Photoshop where it's sometimes more than twice as fast). Also there's more and more x64-only apps.

[cluberti]

I don't agree with you at all cluberti: any serious application developer must optimize his code taking a full advantage of the platform the application will run on.

And? There aren't really that many serious developers out there that do this, otherwise we wouldn't (for example) have the abominations that are Adobe Flash or Acrobat, or even Internet Explorer. Optimizations aren't always for speed, also - sometimes you optimize for supportability, rather than performance. Just because an app is slow doesn't mean it isn't optimized.

The general purpose of Intel and Microsoft is to make you think that only a better hardware will perform faster but that's totally wrong.

I would argue that's your opinion - they are out to make you think you need newer versions of their hardware or software, respectively, which may or may not be true. Ultimately, they do this because otherwise they wouldn't make much money - "buy this again!" doesn't sell as well as "this <new product> has this new <must-have feature>! buy now!".

For example multi-core and 64 bits are out for years and most windows applications aren't optimized for them.

It's a chicken and egg situation - without pushing the hardware, software developers just won't bother writing for 64bit (that goes for both applications and drivers) - for example, Windows has been x86-64 64bit since 2003, yet it took until Windows 7 in 2010 to really push driver and software developers to start writing *good* 64bit drivers and applications. Is it necessary? Probably not, but it's not without benefit either.

Also with all pre-made "solution", Microsoft kill developers creativity: why creating a new algorithm when Microsoft already made one and if it isn't optimized, the application user has to buy a faster hardware.

Sometimes, it is like the developper tried to solve a 2nd degree equation without the formula but trying every possible roots.

Seems like you really just want to bash Microsoft with your posts - perfectly fine opinion (maybe not incorrect in all scenarios either), but my opinion is that it's clouding your judgement to at least some of the facts.

[allen2]

@ Coffeefiend:

I linked to this benchmark on purpose:

- As you can see with 2GB, x86 is first in most test (except winrar and cinebench which are probably heavily x64 optimized). So the 2GB aren't dealt correctly by the x64 OS as it is slowing down applications.

- With more ram, better hardware, x64 works better. That's a hardware selling argument. Also, i am not sure if Win 7 x86 would handle the 4GB ram like XP was: reducing it to 3.7Gb in the best case.

My point was that software must adapt to hardware and not the other way.

@cluberti:

I agree on Adobe, theese kitties are cpu/memory hungry like tigers. As for supportability and performance, it is possible to optimize for both by recompiling your code before runing it like Linux from scratch method. Of course with piracy, developers on MS platforms can't go this way.

For the chicken/egg, it is strange to think that open source software were x64 optimized faster than the others.

Perhaps you're right saying it's clouding my judgement ; i simply can't tell either way.