What is the maximum RAM for the T60p?

[dxben]

Can it hold up to 4GB? And if so, what is the cheapest way to get there? I imagine without buying chips from Lenevo?

[astro]

There are some other threads about this, if you have a browse around.

Plenty of places to buy from, including Newegg.

Supposedly there may be a 3GB restriction in Windows because it is 32-bit and some memory space is reserved for other things. Don't know if this applies if you're not using Windows.

[lithium726]

3GB.

the BIOS will only recognize 3GB, it is a limitation of the i945 chipset.

[dxben]

The Lanevo site is saying 4GB maximum on this model:

http://www-131.ibm.com/webapp/wcs/store ... er=200783U

[Hamid]

Go to the system properties, You will see that Windows has Physical Address Extension (PAE) enabled. For more info go to:

http://www.microsoft.com/whdc/system/pl ... ae_os.mspx

I am note sure about the i945 chipset.

HTH,
Hamid

[dxben]

PAE is for beyond 4GB, not 4GB. 4GB is useable in 32bit since the OS can use 2GB and any one process can have up to 2GB of address space.

I am simply asking how much memory can be physically fit into the T60p and accepted by the bios?

My question simply is, if I put two 2GB chips in a T60p, would the bios recognize it and thus would it be available to my code to manage?

[lithium726]

The Lanevo site is saying 4GB maximum on this model:

http://www-131.ibm.com/webapp/wcs/store ... er=200783U

Yes, but when the machine was laucned a bunch of us called to confim the 4GB. While you can put two 2GB DIMMs in the machine, it will only recognize 3GB.

[dxben]

I see. Well they should change that misleading info on their site.

Thanks for pointing it out. So it is indeed a BIOS limitation. Wth is with that?

These little nitty things detract from overall value IMO...

[lithium726]

I believe it is inherent of Intel's 32bit architechture and method of addressing memory. Has nothing to do with IBM/Lenovo

I am not 100% sure on this as I am not an Intel engineer, but I do recall reading this.

[donking!]

I believe it is inherent of Intel's 32bit architechture and method of addressing memory. Has nothing to do with IBM/Lenovo

I am not 100% sure on this as I am not an Intel engineer, but I do recall reading this.

No, no, no. Sigh. This has been discussed so many times in other threads in this forum. In fact there is currently another thread on exactly this topic running at the same time.

Come on people! No only can't you use the search function, but you don't even look at the current top ten threads?

Please see the post I just put in another thread on this topic:

http://forum.thinkpads.com/viewtopic.php?t=23736

And see this article:

http://www.interact-sw.co.uk/iangblog/2 ... s3gbenough

There is no inherent 3Gb limit to the 32bit architecture.

There is no inherent 3Gb limit to Windows.

The 945 chipset is not limited to 3Gb (if you look at Intel's web site for the specs it specifically says it can handle 4Gb).

*

However, Lenovo has been wishy washy on this point and reported different things in different places, so we're not going to know for sure until someone tries it. But there is no reason in principle that 4Gb shouldn't work in a 32bit system with the 945 chipset.

On the other hand, if you put two 2Gb modules in, they should be recognized. It's just that, depending on the motherboard, all 4Gb or RAM may not be addressable (only 3Gb).

*

But again, we just don't know with the T60 until someone tries!

[lithium726]

I believe it is inherent of Intel's 32bit architechture and method of addressing memory. Has nothing to do with IBM/Lenovo

I am not 100% sure on this as I am not an Intel engineer, but I do recall reading this.

There is no inherent 3Gb limit to the 32bit architecture.

That is not what i said. I said INTEL'S 32bit, as i remember reading this somewhere. It is not only limited to the 945 either, ive have heard reports of it not working on other intel desktop sets as well.

Update: I just read your link. It seems that this IS a limitation of the ENTIRE 32-bit architechture. Yes, the processor can see it. but due to I/0 reservations only 3GB is going to use it. How is this not a limitation?

There is no inherent 3Gb limit to Windows.

never said there was.

The 945 chipset is not limited to 3Gb (if you look at Intel's web site for the specs it specifically says it can handle 4Gb).

and if youll look at hte link you posted, youll find that all the chipsets claiming 4GB support can see 4GB but not actually use it. This is not supporting IMO.
[/b]

[donking!]

lithium726, didn't mean to attribute the claim about Windows and 3Gb to you. I was just bringing it up, because that point often comes up in this discussion. So I was trying to cover the basics. Sorry if it appeared I was attributing this claim to you.

Update: I just read your link. It seems that this IS a limitation of the ENTIRE 32-bit architechture. Yes, the processor can see it. but due to I/0 reservations only 3GB is going to use it. How is this not a limitation?

Which link? You mean the IanG on Tap article? If you read that entire article, it clearly explains that there is not a 3Gb limit inherent to the 32-bit architecture. Depending on the motherboard design, some machines will be able to address 4Gb, some won't. But it is possible, if the motherboard was designed right.

and if youll look at hte link you posted, youll find that all the chipsets claiming 4GB support can see 4GB but not actually use it. This is not supporting IMO.

That's not right. Again, if you read that whole article, you'll see that some motherboards will see 4Gb but not use it, as you say, but others will be able to address the full 4Gb.

Quoting the article:

"So your memory's all still available, it just happens not to be contiguous. The first 3GB are present where you'd expect to find them, then there's a 1GB hole, reserved for devices, and the 4th GB of memory appears in the 5th GB of address space. Since Pentiums can address 64GB of physical address space, there's no problem with this 32 bit processor addressing stuff that requires 33 bits to reach."

Lastly, here's a link for Intel's data sheet on the 945 chipset. Page 19, under "system memory support" it says: "Maximum Memory supported: up to 4Gb at 400, 533, and 667 MHz."

http://download.intel.com/design/mobile ... 921902.pdf

[lithium726]

Yes, I read the whole thing. However, if programs cannot use the memory, who cares? This is what caught me the most about the extended memory address ranges...

(There is one tiny snag. The top 1GB may now be invisible to PCI devices, making it impossible to DMA into that memory. I say 'may' because I'm not sure if the latest PCs are able to do PCI/system bus address mapping like the old RISC systems could. If they can, that will solve the problem. So long as your device drivers are written properly. Do you feel lucky?)

So, how do we know if the 945 can not only address the memory outside the memory hole, but also DMA into it?

how do programs inside windows address the memory if windows itself does not report it as there?

The reason i said its a "limitation" is this: the Intel processors require 33 pins to address 4GB completly. when 4GB is inserted, the last gig (25%) is reserved for I/O and then remapped after that hole. now then, in the article, the author stated that Intel processors have 36 pins for 64GB of address space, which is entirely correct... however, if one were to put 4GB into a strictly 32bit system with 32 address pins, it would see 3GB and discard the extra gig for I/O. How is this not a limitation? The processor requires a 33bit address space in order to address the 5th GB of address space... more than 32.

The first 3GB are present where you'd expect to find them, then there's a 1GB hole, reserved for devices, and the 4th GB of memory appears in the 5th GB of address space. Since Pentiums can address 64GB of physical address space, there's no problem with this 32 bit processor addressing stuff that requires 33 bits to reach.

If i am misunderstanding, i apologize, but that is waht i got from the article.

[donking!]

Those are all good points (about DMA and the number of address pins on the motherboard in question). But that's why I keep saying, we won't know (with the T60) until someone tries it.

My only point with referring to the article is that it establishes that with the right motherboard a 32-bit processor can properly address 4Gb of memory.

Whether or not the 945 chipset will allow this is uncertain. It seems hopeful, as I indicated, because Intel claims in its datasheet that the 945 can support 4Gb. But there are enough uncertainties that it seems up in the air to me.

Still, all I'm trying to argue is that it is possible for a 32-bit system to address 4Gb of physical address space (with DMA and all). There is no inherent limit to the 32-bit architecture, which imposes a 3Gb limit. Nor are their any inherent limits to Windows doing this.

*

however, if one were to put 4GB into a strictly 32bit system with 32 address pins, it would see 3GB and discard the extra gig for I/O. How is this not a limitation? The processor requires a 33bit address space in order to address the 5th GB of address space... more than 32.

I feel like your point here is a little bit academic. As the article points out, there isn't really any such thing as a "strictly 32bit system with 32 address pins." Intel processors have 36 pins. The question is whether the motherboard uses all the pins or not. The 33bit, as it were, is not just an idea, it's a functionality actually built into the actual processor. So if we're talking about actual 32-bit systems that actually exist, there is no inherent limitation as far as 4Gb physical address space is concerned. If we're talking about theoretical "strickly 32-bit" systems, then yes on a purely abstract mathmatical basis of a idealized 32-bit system, there is a limitation. But I believe the person who started this thread was interested in what an actual system (the T60) can do, not a theoretical "strictly" 32-bit system can do. And I thought, when you first brought up limitations, that you were referring specifically to the 945 chipset, which I do not believe falls under your definition of a "stickly 32bit system."

[lithium726]

Yes, the original post was about the T60 and its support of 4GB. it then evolved to 32bits support 4gb of addressable memory, which it doesnt. That is all i was trying to say, that according to your link, there IS an inherent limitation and it is worked around by using 36 address pins (except for on Banias chips, at least thats what i picked up). That particular argument kinda branched off and I was not talking about the 945 anymore, just theoreticals. After all, arent those the most fun to debate?

anyhow, yeah, we wont know until someone tries it. In any case, It does not look optimistic to me, but it is not hopeless. now we just need to find someone with $1200 to flush down the toilet on 4GB of ram

[donking!]

Ah, theoreticals. Yes they always make for a good debate. And you're right on that point as you strickly define it.

But I think it is important to be clear that actual systems, generally referred to as 32-bit systems (as distinguished from your use of the term "32-bit" to refer to something that is an abstraction), are not limited in the way you're talking about. Because this question comes up a lot when people are talking about getting RAM for their systems. And the theoretical limit that you invoke is often cited, to convince people that they can't use 4Gb of RAM. When the proper answer is really: It depends on your motherboard.

Now, indeed, who's going to risk the big bucks? Although, with Crucial's return policy, I suppose someone could order the memory, try it out, and return the modules. Hmm.

[lithium726]

Yup, theyre just limited by DMA and possibly the chipset.

Probably where i got that 4GB was unaddressable on intel 32-bit boards, it was likely not addressable/DMA-able on previous gen chipsets/boards, therefore worthless. Who knows how its gonna turn out that 4GB is actually being considered in systems?

I feel old. I remeber when 64MB was awesome, i didnt even know what was in my 486, was too young. Ill feel really old when we're arguing about whether or not the last 25% of a 64 bit's system memory will be addressable...

[donking!]

I feel old. I remeber when 64MB was awesome, i didnt even know what was in my 486, was too young. Ill feel really old when we're arguing about whether or not the last 25% of a 64 bit's system memory will be addressable...

Hey, I remember being in school and using TRS80s with tape cassette drives. How old is that?